Mitochondrially targeted ceramides preferentially promote autophagy, retard cell growth, and induce apoptosis.

C6-pyridinium (d-erythro-2-N-[6′-(1′′-pyridinium)-hexanoyl]sphingosine bromide [LCL29]) is a cationic mitochondrion-targeting ceramide analog that promotes mitochondrial permeabilization and cancer cell death. In this study, we compared the biological effects of that compound with those of d-erythro-C6-ceramide, its non-mitochondrion-targeting analog. In MCF7 cells it was found that C6-pyridinium ceramide preferentially promoted autophagosome formation and retarded cell growth more extensively than its uncharged analog. This preferential inhibition of cell growth was also observed in breast epithelial cells and other breast cancer cells. In addition, this compound could promote Bax translocation to mitochondria. This redistribution of Bax in MCF7 cells could be blocked by the pan-caspase inhibitor zVAD-fmk but via a Bid-independent signaling pathway. Moreover, C6-pyridinium ceramide-induced translocation of Bax to mitochondria led to mitochondrial permeabilization and cell death. Overall, we show that mitochondrial targeting of C6-pyridinium ceramide significantly enhances cellular response to this compound.

Dharmacon. A mitochondria isolation kit was from Pierce. All other chemicals were from either Sigma or Fisher Scientifi c. C 6pyridinium ceramide (LCL29) and C 6 -ceramide were synthesized by the Lipidomics Core facility at Medical University of South Carolina as previously described and were dissolved in dimethyl sulfoxide ( 24 ).

Cell culturing
MCF7, MDA-MB-157, and MDA-MB-231 cells were cultured in DMEM supplemented with 10% FBS. MCF10A cells were cultured by using a MEGM kit supplemented with 10% FBS. All cells were maintained in 5% C O 2 at 37°C.

Development of MCF7 cells stably expressing green fl uorescent protein-LC3
Human LC3 cDNA was amplifi ed by PCR and cloned into Hind III and Eco RI sites of the LC3 -EGFP vector (GFP-LC3). To develop stable clones of MCF7 cells expressing GFP-LC3, MCF7 cells were seeded onto 10 cm plates and cultured in DMEM supplemented with 10% FBS. Cells were transfected with 12 g of C3-GFP-LC3 plasmid with FuGENE transfection reagent. The day after transfection, cells were selected with 0.2 mg/ml G418. After a week of selection, GFP-positive cells were sorted by fl ow cytometry and maintained in DMEM culture medium in the presence of 0.2 mg/ml G418.

MTT and Trypan blue exclusion assays
Cell growth assay was determined by using an in vitro 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)based assay kit according to the manufacturer's instructions. Briefl y, MCF7, MCF10A, MDA-MB-157, and MDA-MB-231 cells were seeded onto 6-well plates at a density of 6 × 10 5 cells/ml. After 24 h, cells were incubated with different concentrations of C 6 -pyridinium ceramide or C 6 -ceramide for specifi ed times. The IC 50 of each compound was determined from cell growth plots ( 26 ). For Trypan blue exclusion assay, MCF7 cells were plated onto 6-well plates (1 × 10 5 cells/plate) and then treated with 10 M C 6 -pyridinium ceramide or C 6 -ceramide for specifi ed times. Cells were then collected and mixed 1:1 (w/v) with Trypan blue dye. Cells that excluded the dye were counted.
Induction of apoptosis by TNF-␣ , staurosporine, and C 6 -pyridinium ceramide MCF7 or GFP-Bax stable MCF7 cells were plated onto 6-well plates. Cells were then treated with TNF-␣ , staurosporine, cetylpyridinium bromide, C 6 -ceramide, or C 6 -pyridinium ceramide. After specifi ed times, cells were either visualized with fl uorescence microscopy or subjected to immunofl uorescence labeling with anti-Bid or anti-cytochrome c antibody. Subcellular fractionation to prepare cytosolic and heavy membrane fractions from cells treated with these agents was carried out using a mitochondria isolation kit according to the manufacturer's protocol. Resulting protein samples were analyzed by Western blotting with an anti-Bax antibody. To visualize colocalization of Bax with mitochondria, MCF7 cells stably expressing GFP-Bax and mitochondrially-targeted red fl uorescent protein (DS-red-mito; 27) were treated with the ceramides for 18 h and then visualized with fl uorescence microscopy.

Immunofl uorescence labeling and fl uorescence microscopy
For immunofl uorescence labeling, MCF7 cells were plated onto 6-well plates. Cells were subjected to apoptosis induction with the above-stated agents, in the presence or absence of 50 M zVAD-fmk, for specifi ed time periods. Both treated and untreated of mitochondrial membrane potential (8)(9)(10)(11) and the release of cytochrome c from mitochondrial intermembrane spaces (12)(13)(14)(15). It has been reported that the addition of exogenous C 2 -and C 6 -ceramides would result in the redistribution of the proapoptotic protein Bax to mitochondria in HL-60 cells and in Bax-transfected DU-145 cells ( 16 ). In addition, it has been reported that ultraviolet irradiation could activate sphingomyelinases, resulting in ceramide upregulation and Bax conformational change in HeLa cells ( 17 ). Moreover, it has been shown that expression of mitochondrion-targeting bacterial sphingomyelinase could trigger Bax translocation to mitochondria ( 18 ).
Currently, two major apoptotic pathways that signal Bax translocation to mitochondria have been identifi ed ( 19 ). In the extrinsic pathway, the binding of death ligands such as FAS and tumor necrosis factor-␣ (TNF-␣ ) to their respective receptors results in the activation of caspase-8. Caspase-8 then cleaves the BH3-only Bid protein, and truncated Bid (tBid) leads to Bax activation and translocation to mitochondria (20)(21)(22). In the intrinsic pathway, apoptotic stimuli such as staurosporine and ultraviolet irradiation trigger Bax translocation to mitochondria via mechanisms that are independent of caspase-8 and Bid.
Ceramides can be chemically modifi ed to direct their sequestration to a particular cellular compartment. D -erythro -2-N -[6 ′ -(1 ′ ′ -pyridinium)-hexanoyl]sphingosine bromide (LCL29) (C 6 -pyridinium) ceramide is a positively charged analog of the C 6 -ceramide that targets mitochondria and causes mitochondrial permeabilization ( 23,24 ). Its L -threo isomer (LCL124) has displayed a potent antitumor effect in human head and neck squamous cell carcinomas ( 25 ). As C 6 -pyridinium ceramide elicits a number of the abovementioned biological effects, it is unclear how the mitochondrial targeting of this compound will affect its potency. Here, we show that this positively charged ceramidoid, compared with its neutral counterpart, elicits enhanced profi ciency in regulating various cellular processes in MCF7 cells, including cellular growth, autophagy, and apoptosis. In addition, we show that C 6 -pyridinium ceramideinduced cell death involves Bax redistribution, independently of Bid.

Materials
MCF7, MCF10A, MDA-MB-157, and MDA-MB-231 cells were obtained from American Type Culture Collection. LysoTracker Red, fetal bovine serum (FBS), Dulbecco's modifi ed Eagle's medium (DMEM), and tissue culture supplements were from Invitrogen, Inc. FuGENE 6 transfection reagent was from Roche, and Maxiprep kits were from Qiagen. Horseradish peroxidase-and Cy3-conjugated secondary immunoglobulins and enhanced chemiluminescence (ECL) Western blotting detection kit were obtained from Amersham. Anti-Bid and cytochrome c antibodies were from BD Pharmingen. Anti-LC3 antibody was from Santa Cruz Biotechnology. Human LC3 TruClone cDNA was purchased from Origene. Pan-caspase inhibitor zVAD-fmk and a mammary epithelial growth medium (MEGM) kit were purchased from Lonza Corp. Oligonucleotides and transfection reagents used in short-interfering RNA (siRNA) studies were purchased from gosome vesicles. Similarly, cells treated with C 6 -pyridinium ceramide had numerous autophagosome vesicles, albeit fewer than the tamoxifen-treated cells. On the other hand, in untreated and C 6 -ceramide-treated cells, GFP-LC3 remained primarily in the cytoplasm.
In addition, Western blotting analysis of cell lysates was carried out to determine the LC3-II-to-LC3-I ratio, which serves as an indicator of autophagosome formation. As shown in Fig. 2A ,, upon treatment with C 6 -pyridinium ceramide or tamoxifen, there was an obvious increase in the LC3-II-to-LC3-I ratio, compared with that in untreated and C 6 -ceramide-treated cells. We also examined the lysates from MCF10A breast epithelial cells, MDA-MB-137 medullary carcinoma, and MDA-MB-231 breast adenocarcinoma cells. Interestingly, both of the treatments with C 6 -ceramide and C 6 -pyridinium ceramide resulted in an increase in the LC3-II-to-LC3-I ratio in MDA-MB-231 cells. No LC3 was detected in MCF10A and MDA-MB-137 cell lysates. Moreover, the level of Beclin-1 in MCF7 cells following ceramide treatment was also examined. We found that treatment with C 6 -pyridinium ceramide did not appreciably increase the level of Beclin-1 ( Fig. 2B, C ).

C 6 -pyridinium ceramide inhibits cell growth and promotes cell death
To assess the effects of C 6 -pyridinium ceramide and C 6ceramide on cell growth, MCF7 cells were treated with different concentrations of these compounds. At 24 h posttreatment, cells were then washed three times in phosphate-buffered saline (PBS), fi xed in 6% paraformaldehyde for 30 min, and permeabilized with a 0.06% saponin solution in PBS for 15 min. Cells were subsequently blocked for 30 min in the blocking buffer containing PBS, 5% FBS, and a 0.06% saponin mixture. Cells were then incubated with 5 g/ml anti-Bid, or 5 g/ml anti-cytochrome c antibody diluted in blocking buffer containing a 0.04% saponin mixture for 2 h. Cells were then washed and incubated with 6 g/ml Cy3-labeled goat anti-rabbit or goat anti-mouse immunoglobulin in the blocking buffer containing a 0.04% saponin solution for an additional 2 h. Cells were then washed with PBS and treated with the antifade reagent. Cells were visualized with an Olympus IX-70 fl uorescence microscope using an LCPlanFI 20× objective lens at 1.5× intermediate magnifi cation. Images were acquired with an Optronics DEI-750D digital imaging camera.

Induction of autophagy by ceramide
For LysoTracker staining studies, GFP-Bax stable MCF7 cells were plated onto 6-well plates. Cells were then incubated with 10 M C 6 -pyridinium ceramide or C 6 -ceramide for 16 h. Cells were then stained with 0.75 M LysoTracker Red for 30 min and visualized with an Olympus IX-70 fl uorescence microscope. To visualize autophagosome formation, GFP-LC3 stable MCF7 cells were plated onto 6-well plates, treated with C 6 -pyridinium ceramide or C 6 -ceramide for 16 h, and visualized by fl uorescence microscopy. For assessment of the LC3-II-to-LC3-I ratio by immunoblotting, cells were treated with 10 M C 6 -pyridinium ceramide or C 6ceramide for 18 h. Cell lysates were then analyzed by Western blotting with an anti-LC3 antibody. To determine the level of Beclin-1 in the cell lysates, Western blotting analysis with an anti-Beclin-1 antibody was carried out. The expression level of Beclin-1 was normalized to that of ␤ -actin, using a Scion image program.

C 6 -pyridinium ceramide preferentially mediates autophagy in MCF7 cells
Elevation in ceramide levels has been attributed to an elevation in cellular autophagy ( 5 ). To determine the efficacy of short-chained ceramides in promoting autophagosome formation, we treated MCF7 cells with either C 6 -ceramide or C 6 -pyridinium ceramide for 16 h. Cells were then stained with LysoTracker Red and visualized by fl uorescence microscopy. As shown in Fig. 1A , a number of cells treated with C 6 -pyridinium ceramide displayed LysoTracker Red-stained vesicles. On the other hand, untreated and C 6ceramide-treated cells showed far fewer stained vesicles. In addition, MCF7 cells stably expressing GFP-LC3 were generated. Cells were then treated with the above-stated compounds for 16 h and visualized by fl uorescence microscopy. As shown in Fig. 1B , the cells treated with the positive control tamoxifen displayed numerous GFP-labeled autopha- tions of C 6 -pyridinium ceramide signifi cantly enhanced the inhibition of cell growth. The IC 50 s of these compounds were found to be 52 M, 4 M, and 48 M, respectively, for the above-listed cells. On the other hand, the growth of these cells remained static following C 6 -ceramide treatment. We also monitored cell growth inhibition over time in the presence of 10 µM of these compounds. As shown in Fig. 4B , cell growth inhibition and most likely a decrease in cell viability were evident over time for all three cell lines upon C 6 -pyridinium ceramide treatment, compared with that with C 6 -ceramide treatment. cells were analyzed using the MTT assay. As shown in Fig. 3A , increasing concentrations of C 6 -pyridinium ceramide significantly enhanced the inhibition of cell growth. The IC 50 of this compound was found to be 20 µM compared with an IC 50 of 80 M for C 6 -ceramide ( Fig. 3A ). We also monitored cell growth inhibition over time in the presence of 10 µM of these compounds. As shown in Fig. 3B , cell growth inhibition was signifi cantly higher after 24 h treatment with C 6pyridinium ceramide than with C 6 -ceramide. To corroborate the MTT assay results of growth inhibition, Trypan blue exclusion analyses of untreated cells and MCF7 cells treated with 10 M C 6 -ceramide or C 6 -pyridinium ceramide were carried out. As shown in Fig. 3C , the number of live cells following C 6 -ceramide treatment remained relatively constant over time, whereas the number live cells decreased over time following C 6 -pyridinium ceramide treatment. This suggests that in addition to inhibiting cell growth, C 6 -pyridinium ceramide may also promote cell death.
In addition to examining effects on MCF7 cells, we also compared the effects of these two compounds on the growth of MCF10A, MDA-MB-157, and MD-MB-231 cells by using the MTT assay. Figure 4A shows that increasing concentra-   3. C 6 -pyridinium ceramide (C 6 -Pyr-Cer) inhibited MCF7 cell growth in a dose-and time-dependent manner. MCF7 cells were plated onto 6-well plates and treated with different concentrations (0-80 M) of C 6 -pyridinium ceramide or C 6 -ceramide (C 6 -Cer) for 24 h (A), or with 10 µM of these compounds for various time periods (B). At specifi ed time points, MTT assays were performed. Time-course results were normalized to values from the control cells at the initial time point. C: Control MCF7 cells and cells treated with 10 µM C 6 -ceramide or C 6 -pyridinium ceramide were collected and stained with Trypan blue to determine the number of live cells at different time points. All results were averaged from three independent studies. Values are means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 versus the control. permeabilization and cell death ( 23 ). Because Bax can induce mitochondrial permeabilization, we therefore set out to determine the role of this protein in C 6 -pyridinium ceramide-mediated cell killing. A key step by which Bax promotes cell death involves its translocation from the cytoplasm to mitochondria ( 10,11 ). In order to enable a rapid assessment of Bax subcellular localization, we used MCF7 breast carcinoma cell line stably expressing GFP-Bax ( 27 ).
Fluorescence microscopy analysis indicated that under nonapoptotic conditions, in greater than 98% of the cells Bax was localized to the cytoplasm and that in fewer than 2% of the cells Bax was localized to mitochondria. We found that when these cells were subjected to apoptosis induction with a variety of agents including the extrinsic pathway inducer TNF-␣ (10 ng/ml for 16 h), the intrinsic pathway inducer staurosporine (0.2 M for 6 h), or C 6pyridinium ceramide (10 M for 16 h), there was a shift in the fl uorescence pattern of GFP-Bax from a diffuse cytoplasmic state to a punctate membrane-bound state ( Fig. 6A ). On the other hand, GFP-Bax remained cytoplasmic in cells treated with C 6 -ceramide or cetylpyridinium bromide (both, 10 M for 16 h). To verify that punctate GFP-Bax localized with mitochondria, GFP-Bax-stable and Ds-red-mito-stable MCF7 cells were treated with C 6 -ceramide or C 6 -pyridinium ceramide and then visualized by fl uorescence microscopy. As shown in Fig. 6B , punctate GFP-Bax colocalizes with DS-red-mito-labeled mitochondria. These results indicate that C 6 -pyridinium ceramide, which localizes to mitochondria, induces Bax translocation to mitochondria.
To determine the time course of C 6 -pyridinium ceramide-mediated Bax translocation to mitochondria, GFP-Bax-stable MCF7 cells were treated with 10 M of this compound for various periods of time ( Fig. 7A ). We also treated cells with different concentrations of either C 6ceramide or C 6 -pyridinium ceramide for 18 h and quantitated the extent of GFP-Bax localization to mitochondria by fl uorescence microscopy ( Fig. 7B ). Results showed that increasing concentrations of C 6 -pyridinium ceramide As C 6 -pyridinium ceramide treatment induces cell death as well as autophagy, we then examined whether inhibition or promotion of autophagy enhances cell survival. MCF7 cells were treated with the autophagy inhibitor 3-methyl adenine or the autophagy promoter tamoxifen, in the presence or absence of C 6 -pyridinium ceramide for 24 h. Cells were then stained with Hoechst dye to detect apoptotic nuclei. As shown in Fig. 5 , the presence of tamoxifen reduced the number of apoptotic nuclei following C 6pyridinium ceramide treatment. This result suggests that in MCF7 cells, autophagy may serve as a protective mechanism against C 6 -pyridinium ceramide-mediated mitochondrial damage. C 6 -pyridinium ceramide, but not C 6 -ceramide, induces Bax translocation from the cytoplasm to mitochondria in MCF7 cells C 6 -pyridinium ceramide has been previously shown to be a potent anticancer agent that mediates mitochondrial  10% in the presence of zVAD-fmk, indicative of a direct role of caspase in TNF-␣ -mediated Bax translocation to mitochondria. On the other hand, the addition of zVADfmk had no effect on staurosporine-mediated Bax translocation. As for C 6 -pyridinium ceramide, the addition of zVAD-fmk signifi cantly decreased Bax translocation to mitochondria ( Fig. 8A, B ). In addition to using fl uorescence microscopy, we carried out subcellular fractionation of control and treated cells. As shown in Fig. 8C , TNF-␣ , staurosporine, and C 6 -pyridinium ceramide resulted in a shift in Bax (both GFP-tagged and endogenous) distribution from the cytosolic to mitochondrial fraction. Incubation of cells with zVAD-fmk signifi cantly inhibited Bax redistribution to mitochondria in TNF-␣ -and C 6 -pyridinium ceramide-treated cells. Therefore, caspases appear to be involved in the action of both agents on Bax.
To determine the role of Bid in C 6 -pyridinium ceramidemediated Bax translocation to mitochondria, GFP-Bax-stable MCF7 cells were treated with this compound either in the presence or absence of zVAD-fmk, and the treated cells were then subjected to immunofl uorescence labeling with an anti-Bid antibody. For comparison, we also treated cells with 10 ng/ml TNF-␣ or 0.2 M staurosporine. As shown in increased the percentage of cells in which Bax was localized to mitochondria. On the other hand, we found that non-mitochondrion-targeting C 6 -ceramide had little effect on promoting Bax translocation to mitochondria.

C 6 -pyridinium ceramide induces Bax translocation from the cytoplasm to mitochondria via a caspase-dependent and Bid-independent manner
In order to examine the signaling pathway leading to Bax translocation to mitochondria in GFP-Bax-stable MCF7 cells treated with C 6 -pyridinium ceramide, we fi rst examined whether this translocation process was caspase dependent by the addition of the pan-caspase inhibitor zVAD-fmk. For verifi cation of effi cacy and specifi city, cells were also treated with TNF-␣ or staurosporine in the presence or absence of zVAD-fmk. As shown in Fig. 8A and B , in the absence of zVAD-fmk, treatment with TNF-␣ resulted in approximately 90% of cells had Bax localized to mitochondria. In contrast, the percentage of GFP-Bax punctate cells in response to TNF-␣ was greatly reduced to  GFP-Bax-stable MCF7 cells were treated with various concentrations of C 6 -ceramide (C 6 -Cer) or C 6 -pyridinium ceramide for 16 h. Percentages of GFP-Bax punctate cells were quantitated from four separate visual fi elds. Results were averaged from three independent studies. Error bars represent SDs. or C 6 -pyridinium ceramide-mediated Bax redistribution. Overall, these results indicate that C 6 -pyridinium ceramidemediated Bax translocation to mitochondria in MCF7 cells is dependent on a caspase but independent of Bid.

C 6 -pyridinium ceramide-induced Bax translocation leads to cytochrome c release and cell death
Previously, it has been shown that Bax translocation to mitochondria leads to the release of cytochrome c ( 14,31,32 ) and cell death. By using immunofl uorescence labeling with an anti-cytochrome c antibody, we found that in healthy cells, where GFP-Bax was cytoplasmic, cytochrome c was retained in mitochondria ( Fig. 10A ). On the other hand, cytoplasmic cytochrome c became evident in TNF-␣ -, staurosporine-, and C 6 -pyridinium ceramide-treated cells that had GFP-Bax localized to mitochondria. In addition, by staining both the untreated and treated cells with Hoechst nuclear stain, we found that there was signifi cant cell death, as evidenced by nuclear condensation in cells that had GFP-Bax localized to mitochondria ( Fig. 10B ). These results indicate that C 6 -pyridinium ceramide-induced mitochondrial permeabilization and cell death involve Bax. DISCUSSION C 6 -pyridinium ceramide was developed as a positively charged ceramide that preferentially accumulates in mi- Addition of zVAD-fmk, which blocked Bax translocation to mitochondria, also resulted in a cytoplasmic Bid localization pattern. Treatment with either staurosporine or C 6 -pyridinium ceramide resulted in GFP-Bax localization to mitochondria, while Bid remained in the cytoplasm, and this local ization pattern was independent of zVAD-fmk addition.
In order to corroborate immunofl uorescence labeling assays, Western blotting analysis of cell lysates was performed with an anti-Bid antibody to detect Bid cleavage. As shown in Fig. 9B , treatment with TNF-␣ resulted in Bid cleavage in a caspase-dependent manner, in agreement with previous reports ( 29,30 ). No Bid cleavage, however, was observed in staurosporine-or C 6 -pyridinium ceramide-treated cells.
To ascertain that Bid is not involved in C 6 -pyridinium ceramide-mediated Bax translocation to mitochondria, GFP-Bax-stable MCF7 cells were transfected with Bid siRNA oligonucleotides to knockdown the expression level of this protein. As shown in Fig. 9C , treatment with Bid siRNA knocked down Bid protein level without affecting the expression levels of other cellular proteins including Bcl-2 and ATP synthase ␤ -subunit. Both scrambled and Bid siRNA oligonucleotide-transfected cells were treated with TNF-␣ , staurosporine, or C 6 -pyridinium ceramide. Figure 9D shows that whereas Bid knockdown affected TNF-␣ -mediated Bax translocation to mitochondria, it had no effect on staurosporine- Fig. 8. C 6 -pyridinium ceramide (C 6 -Pyr-Cer)-induced Bax translocation to mitochondria is partially inhibited by zVAD-fmk. A: GFP-Bax-stable MCF7 cells were treated with 10 ng/ml TNF-␣ (16 h), 0.2 M staurosporine (6 h), or 20 M C 6 -pyridinium ceramide (16 h) either in the absence or presence of 50 M zVAD-fmk. Cells were visualized by fl uorescence microscopy. B: A quantitative analysis of the effect of zVAD-fmk on regulation of TNF-␣ , staurosporine (STS), and C 6 -pyridinium ceramide-mediated Bax translocation to mitochondria. Percentages of cells having GFP-Bax localized to mitochondria were averaged from three separate visual fi elds. The study was carried out three times. C: MCF7 cells stably expressing GFP-Bax were treated with the abovestated compounds. Treated cells were then separated into soluble protein (C) and heavy membrane (M) fractions. Protein samples were analyzed by Western blotting with an anti-Bax antibody.
preferentially promotes cellular autophagy, retards cell growth, and induces cellular apoptosis.
Short-chained ceramides have been shown to promote cellular autophagy ( 5 ). In the current study, we show that tochondria of living cells ( 23 ). In this report, we compared the biological effects of this mitochondrion-targeting C 6ceramidoid to its neutral non-mitochondrion-targeting analog. We found that this positively charged ceramide Fig. 10. Bax localization to mitochondria in GFP-Bax-stable MCF7 cells is associated with the release of cytochrome c (Cyt. c) and nuclear condensation. A: GFP-Bax-stable MCF7 cells were treated with 10 ng/ml TNF-␣ (18 h), 0.2 M staurosporine (STS) (6 h), or 10 M C 6 -pyridinium ceramide (C 6 -Pyr-Cer) (18 h). Untreated and treated cells were then subjected to immunofl uorescence labeling with an anti-cytochrome c antibody and visualized by fl uorescence microscopy. B: GFP-Bax-stable MCF7 cells were treated with the above-stated apoptotic stimuli. Cells were then stained with Hoechst nuclear stain and visualized by fl uorescence microscopy. Fig. 9. Bid is not involved in C 6 -pyridinium ceramide-(C 6 -Pyr-Cer) mediated Bax translocation to mitochondria. A: MCF7 cells stably expressing GFP-Bax were treated with 10 ng/ml TNF-␣ (18 h), 0.2 M staurosporine (STS) (6 h), or 10 M C 6 -pyridinium ceramide (18 h) either in the absence or presence of zVAD-fmk. Cells were then subjected to immunofl uorescence labeling with an anti-Bid antibody. Labeled cells were visualized by fl uorescence microscopy. B: GFP-Bax-stable MCF7 cells were treated with various apoptosis inducing agents (TNF-␣ , staurosporine, C 6 -pyridinium ceramide) in the presence or absence of zVAD-fmk. Cell lysates prepared from untreated and treated cells were analyzed by Western blotting with an anti-Bid antibody. Cells were subsequently visualized by fl uorescence microscopy and the percentages of GFP-Bax punctate cells were quantitated from four separate visual fi elds. The study was carried out three times. ated by C 6 -pyridinium ceramide in MCF7 cells that is dependent on a caspase but independent of Bid.
Previously, it was reported that C 6 -pyridinium ceramide could induce mitochondrial permeabilization ( 23 ). Based on the current study, it is likely that this permeabilization is mediated through Bax translocation to mitochondria, causing the release of apoptogenic proteins such as cytochrome c. In addition, it has been shown that addition of C 2 -and C 6 -ceramide could cause a Bax conformational change and localization to mitochondria ( 16,(39)(40)(41). Fluorescently labeled short-chain ceramide species have been shown to be localized to the Golgi apparatus ( 23,42 ). Surprisingly, in our study, we found that C 6 -ceramide had little effect in triggering Bax translocation, even at high concentrations. This difference may be due to the different cell types used and the method of detecting Bax translocation. Alternatively, it may be possible that C 6 -ceramide also reaches mitochondria; however, the local concentrations are either lower or it is metabolized differently than C 6pyridinium ceramide. Interestingly, we found that addition of cetylpyridinium bromide did not promote Bax translocation, suggesting that the pyridinium moiety of the C 6 -pyridinium ceramide, although important for targeting to mitochondria, is not responsible for Bax activation.
It was recently reported that D -erythro -sphingosine could promote Bax activation in rhabdomyosarcoma cells ( 43 ). This effect appeared to be stereo-specifi c as its L -erythro and DL -threo analogs were ineffective at inducing cell death. We have tested L -erythro -and L -threo -C 6 -pyridinium ceramides (LCL187 and LCL124, respectively) and found them to be equally effective at triggering Bax translocation to mitochondria (Hou et al., unpublished data). This suggests that stereochemistry of C 6 -pyridinium ceramide is not critical in regulating Bax translocation.
Bax undergoes a conformational change as it redistributes to mitochondria ( 44 ). It has been previously shown that Bax conformational change could be induced by nonionic detergents ( 9,28 ). In our study, we found that incubation of soluble Bax with C 6 -pyridinium ceramide did not lead to a conformational change (Hou and Hsu, data not shown). This suggests that Bax activation is a consequence of the signaling event initiated by this mitochondriontargeting ceramide rather than a direct alteration of Bax conformation by this compound.
In the TNF-␣ -mediated apoptotic pathway, activation of caspase-8 causes Bid cleavage. tBid then translocates to mitochondria and activates Bax. In our GFP-Bax-stable MCF7 cells, we also observed this TNF-␣ -mediated Bid cleavage and the cotranslocation of Bid and Bax to mitochondria. This is in contrast with the staurosporine-induced apoptotic pathway in which we observed Bax translocation to mitochondria in the absence of Bid cleavage and mitochondrial translocation. On the other hand, we found that C 6 -pyridinium ceramide triggered Bax translocation to mitochondria via a caspase-dependent pathway but without Bid cleavage and without colocalization to mitochondria. Consistently, we found that knockdown of Bid by the siRNA approach attenuated Bax translocation in TNF-␣ -but not in staurosporine-or C 6 -pyridinium ceramide-treated cells. C 6 -pyridinium ceramide treatment of MCF7 cells significantly enhanced the production of acidic vesicular organelles and autophagosomes and elevated the level of LC3-II. On the other hand, no signifi cant formation of acidic vesicular organelles, autophagosomes, or LC3-II was observed in C 6 -ceramide-treated MCF7 cells. Interestingly, formation of LC3-II was observed in MDA-MB-231 cells treated with C 6 -ceramide. This suggests that the formation of autophagosomes induced by short-chained ceramide may be cell-type dependent. Moreover, upregulation of Beclin-1 has been suggested as a mechanism by which ceramide promotes autophagy ( 5 ). However, in our study, we did not observe a ceramide-induced elevation in Beclin-1. Alternatively, it is possible that ceramide may enhance JNK1-depedent Bcl-2 phosphorylation to enable its dissociation from Beclin-1 to promote autophagy ( 33 ).
Autophagy has been proposed to act either as a tumor suppressor to promote cancer cell death or as a cytoprotective mechanism to shield cancer cells against anticancer treatments ( 34 ). In our study, we found that treatment of MCF7 cells with the autophagy promoter tamoxifen could attenuate C 6 -pyridinium ceramide-mediated cell death, suggesting that autophagy may represent a cellular protective mechanism. In addition, we found that this compound induced autophagy within the same time frame as apoptotic cell death in MCF7 cells. In our study, we found that the pan-caspase inhibitor zVAD-fmk could attenuate C 6 -pyridinium ceramide-mediated Bax redistribution and cell death but not autophagosome formation (data not shown). Either these two events are independent of each other or autophagosome formation is upstream of the zVAD-fmk-sensitive portion of the Bax-activation signaling pathway. Interestingly, Akt activation has been shown to suppress apoptosis and autophagy ( 35,36 ), and a ceramide could suppress Akt activity though activation of protein phosphatase 2 ( 37,38 ). It is possible that C 6 -pyridinium ceramide could more effectively activate protein phosphatase 2 than C 6 -ceramide to promote both apoptosis and autophagy.
Previously, it has been reported that C 6 -pyridinium ceramide can attenuate the proliferation of resting T cells ( 4 ). In the current study, it was shown that C 6 -pyridinium ceramide could greatly reduce the rate of cell growth compared with its uncharged analog in breast epithelial and breast cancer cell lines, with higher IC 50 values in breast epithelial cells than in cancer cells. The accumulation of C 6 -pyridinium ceramide in mitochondria and its ability to promote mitochondrial permeabilization ( 23 ) likely reduced mitochondrial function and cell viability, thus contributing to the observed decrease in the rate of cell growth.
Currently, two main apoptotic signaling pathways have been proposed that regulate Bax translocation to mitochondria. In the extrinsic pathway of apoptosis, caspase-8cleaved Bid activates Bax and leads to Bax translocation to mitochondria. On the other hand, in the intrinsic pathway of apoptosis, Bax activation and mitochondrial translocation are independent of caspase-8 and Bid. In this study, we have uncovered a unique Bax-activating process medi-TNF-␣ has been previously shown to elevate ceramide production in HL-60 and A431 cells ( 45,46 ). It remains to be determined to what extent this elevated pool of ceramide contributes to Bax activation.
It has been reported that ceramides derived from acid sphingomyelinase in the lysosome could activate cathepsin D ( 47 ), which in turn, could lead to Bax activation by an as-yet-unknown mechanism ( 48 ). We found that pepstatin A, which has been reported to inhibit cathepsin D and Bax activation in staurosporine-treated T lymphocytes, had no effect in infl uencing Bax localization in C 6 -pyridinium ceramide-or staurosporine-treated MCF7 cells (Hou and Hsu, unpublished observation). This suggests that Bax translocation induced by C 6 -pyridinium ceramide in MCF7 cells likely does not involve cathepsin D and may be distinct from the lysosomal pathways mediated by ceramide.
In MCF7 cells, we found that C 6 -pyridinium ceramidemediated Bax redistribution appeared to be caspasedependent. Further analysis with various caspase inhibitors indicated that the caspase-5 inhibitor z-WEHD-fmk could also signifi cantly inhibit Bax redistribution (Hou and Hsu, unpublished data). Although this caspase-dependent signaling pathway appeared to be important for Bax redistribution, it is unique, however, only to MCF7 cells. We found that in DU-145, NT-2, LS-180, and DLD cells stably expressing GFP-Bax, addition of zVAD-fmk did not attenuate C 6pyridinium ceramide-mediated Bax redistribution (Hou and Hsu, unpublished data).
Overall, our study shows that selective compartmentalization of ceramide may affect the potency of this signaling molecule in regulating various cellular processes. This is evidenced by differential regulations of cellular autophagy, growth, and viability that are elicited by mitochondrion-and non-mitochondrion-targeting ceramide species. It will be interesting in the future to identify the direct molecular targets of C 6 -pyridinium ceramide that facilitate the regulation of the above-stated biological processes.