BAK activation is necessary and sufficient to drive ceramide synthase-dependent ceramide accumulation following inhibition of BCL2-like proteins

Abstract

Determining mechanistic details about how drugs kill cancer cells is critical for predicting which cancers will respond to given therapeutic regimens and for identifying effective combinations of drugs that more potently kill cancer cells while sparing normal cells. The BCL2 family of proteins and bioactive sphingolipids are intricately linked during apoptotic cell death. In fact, many chemotherapeutic drugs are known to cause accumulation of the pro-apoptotic sphingolipid ceramide; however, the mechanism by which this occurs is not completely understood. In the present study we demonstrate that direct inhibition of anti-apoptotic BCL2 proteins with ABT-263 is sufficient to induce C(16)-ceramide synthesis in multiple cell lines, including human leukaemia and myeloma cells. ABT-263 activates CerS (ceramide synthase) activity only in cells expressing BAK or in cells capable of activating BAK. Importantly, recombinant BAK is sufficient to increase in vitro CerS activity in microsomes purified from Bak-KO (knockout) cells and activated BAK more potently activates CerS than inactive BAK. Likewise, ABT-263 addition to wild-type, but not Bak-deficient, microsomes increases CerS in vitro activity. Furthermore, we present a feed-forward model by which BAK activation of CerS by chemotherapeutic drugs leads to elevated ceramide levels that result in synergistic channel formation by ceramide (or one of its metabolites) and BAX/BAK.

Figure 1. Treatment of human leukemia cell lines with the BCL2/BCLx_L /BCLw inhibitor, ABT-263, causes accumulation of C16-ceramide

A, U937, K562 and MV411 cells were incubated with increasing doses of ABT-263 (0.7 nM to 45 μM) for 48 hours and the relative numbers of viable cells were determined by reduction of alamar blue. All data points represent the compilation of at least biological triplicates. B, U937 cells were treated with 1.6 μM of ABT-263, or vehicle (DMSO), for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. C, K562 cells were treated with 150 nM of ABT-263, or vehicle, for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. D, MV411 cells were treated with 50 nM of ABT-263, or vehicle (DMSO), for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. D, ***, _P=<0.0005; **, _P=<0.005; *,p=<0.05.

Figure 2. ABT-263 induced ceramide accumulation and CerS activity can be inhibited by MCL1 in U937 cells

A, U937 cells stably expressing GFP and a portion of an irrelevant protein, a portion of ER (U937-ER) were treated with 1.6 μM of ABT-263, or vehicle (DMSO), for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. B, U937 cells stably expressing GFP and BCLx_L (U937-BCLx_L) were treated with 1.6 μM of ABT-263, or vehicle (DMSO), for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. C, U937 cells stably expressing GFP and MCL1 (U937-MCL1) were treated with 1.6 μM of ABT-263, or vehicle (DMSO), for 8 and 24 hours and the total amount of cellular ceramides were determined by HPLC-MS/MS. D. U937-MIG ER cells (U937 cells stably infected with MIG-ER virus) were incubated with 1.6 μM of ABT-263, or vehicle (DMSO), for 1 hour and 45 minutes. 1μM C₁₇-sphingosine was then added to the media for an additional 15 minutes. Cells were immediately lysed and the amount of C₁₇-ceramide in the cells were quantitated by HPLC-MS/MS. E, U937-BCLx_L cells were incubated with 1.6 μM of ABT-263, or vehicle (DMSO), for 1 hour and 45 minutes. 1μM of C₁₇-sphingosine was then added to the media for an additional 15 minutes. Cells were immediately lysed and the amount of C₁₇-ceramide in the cells were quantitated by HPLC-MS/MS. F, U937-MCL1 cells were incubated with 1.6 μM of ABT-263, or vehicle (DMSO), for 1 hour and 45 minutes. 1μM of C₁₇-sphingosine was then added to the media for an additional 15 minutes. Cells were immediately lysed and the amount of C₁₇-ceramide in the cells were quantitated by HPLC-MS/MS. Lipids were normalized to total lipid phosphate and data expressed as a fold-change of the untreated control. All data points represent the compilation of at least biological triplicates. ***, p=<0.0005; **, p=<0.005; *, p=<0.05.

Figure 3. MCL1 inhibits ABT-263 induced C₁₆-ceramide generation and increased in situ CerS activation in RPMI8226 human plasma cell leukemia

A, RPMI8226 cells stably expressing GFP and a portion of an irrelevant protein, ER. (RPMI-ER) were treated with 0.5 μM of ABT-263, or vehicle (DMSO), for 2 and 8 hours and C₁₆-ceramide quantified by HPLC-MS/MS. B, RPMI8226 stably expressing GFP and MCL1 (RPMI-MCL1) were treated with 5 μM of ABT-263, or vehicle (DMSO), for 2 and 8 hours and C₁₆-ceramide quantified by HPLC-MS/MS. C. RPMI-ER cells were incubated with 0.5 μM of ABT-263, or vehicle (DMSO), for 1 hour and 45 minutes. 1μM C₁₇-sphingosine was then added to the media for an additional 15 minutes. Cells were immediately lysed and the amount of C₁₇-ceramide in the cells were quantitated by HPLC-MS/MS. D, RPMI-MCL1 cells were incubated with 5 μM of ABT-263, or vehicle (DMSO), for 1 hour and 45 minutes. 1μM of C₁₇-sphingosine was then added to the media for an additional 15 minutes. Cells were immediately lysed and the amount of C₁₇-ceramide in the cells were quantitated by HPLC-MS/MS. Lipids were normalized to total lipid phosphate and data expressed as a fold-change of the untreated control. All data points represent the compilation of at least biological triplicates. ***, p=<0.0005; **, p=<0.005; *, p=<0.05.

Figure 4. Bak is necessary and sufficient for ceramide synthase activity, in vitro.

A, Bak deficient mitochondria are severely deficient in CerS activity. Increasing amounts of mitochondria isolated from wild-type (wt) or Bak deficient (Bak KO) baby mouse kidney (BMK) cells were incubated with C₁₇-sphingosine for 15 minutes. Lipids were extracted and the conversion of C₁₇-sphingosine into C₁₇-ceramide was quantitated by HPLC-MS/MS. Data shown are the combination of biological triplicates done in duplicate. B, Soluble Bak, but not Bax, activates CerS on Bak KO microsomes. Microsomes isolated from Bak KO BMK cells (10 μg) were incubated with increasing concentrations of cleaved. soluble Bak (cBak, closed circles) or soluble Bax (open circles) for 15 mm and the activity of CerS was determined. C, Activated Bak is a more potent activator of CerS, in vitro. Microsomes from Bak KO BMK cells (10 μg) were incubated with recombinant Bak (5 ng), cleaved, active Bid (tBid) (10 ng), or Bak and tBid (5 ng Bak; 10 ng tBid). Reactions were incubated with C₁₇-sphingosine for 15 minutes and the conversion to C₁₇-ceramide was quantitated by HPLC-MS/MS. D. Baby mouse kidney cells established from wild-type mice (WT BMK) were treated with 2.67 μM of ABT-263 (ABT) or vehicle (DMSO) for 2 and 8 hours. Lipids were extracted and the total cellular levels of individual ceramide species were quantitated by HPLC-MS/MS. E, Baby mouse kidney cells established from Bak KO mice (BAK KO BMK) were treated with 2.67 μM of ABT-263 (ABT) or vehicle (DMSO) for 2 and 8 hours. Lipids were extracted and the total cellular levels of individual ceramide species were quantitated by HPLC-MS/MS. C, Heavy membranes were isolated from WT (closed circles) or Bak KO (open circles) BMK cells and incubated with increasing concentrations of ABT-263. The relative increase in the conversion of C₁₇-sphingosine to C₁₇-ceramide is indicated, as compared to vehicle treated heavy membranes from the respective genotype of BMK cell. Dashed line indicates no fold change. Each point represents biological triplicates done in duplicate. ***, p=<0.0005; **, p=<0.005; *, p=<0.05.

Figure 5

Feed-forward model for how the generation of ceramide, through a BAK-dependent mechanism, can contribute to MOMP-induced apoptosis. See text for discussion for detailed description.