Obatoclax Potently Inhibits Constitutive Interactions Between MCL-1 and BAK in Intact Mitochondria.

Because the hydrophobic character of obatoclax is compatible with membranes, the ability of the compound to interfere with BCL protein–protein interactions was determined in isolated, intact mitochondria. We took advantage of the fact that MCL-1 constitutively interacts with BAK in the mitochondrial outer membrane (18, 19). Chemical cross-linking was used to measure this interaction because it can be conducted while maintaining the structural integrity of the membrane. A screen of potential cross-linking agents was performed and revealed LC-SMCC [3-aminopropyltriethoxysilane (APTES), succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxy-(6-amidocaproate)] (a membrane-permeant hetero-bifunctional compound that covalently links a primary amine and a thiol) as the optimal agent. LC-SMCC is not reactive with obatoclax. Mitochondria from human SK-Mel5 melanoma cells were incubated with either vehicle (DMSO) or vehicle containing 0.1 mM LC-SMCC, and the mitochondria subsequently solubilized with detergent. MCL-1 was recovered by immunoprecipitation with anti-MCL-1 antibody, and the presence in the precipitate of MCL-1 cross-linked to BAK (MCL-1/BAK dimer), as well as monomeric MCL-1, determined by immunoblot (Fig. 1b). Pretreatment of the mitochondria with 10 nM of either obatoclax or a synthetic 25-aa peptide corresponding to the BH3 domain of human NOXA (which interacts selectively with MCL-1; ref. 24) before the addition of cross-linking agent strongly inhibited the formation of MCL-1/BAK dimers (Fig. 1b). Of note, however, addition of concentrations as high as 2.0 μM obatoclax to isolated mitochondria did not stimulate release of cytochrome c in vitro, whereas a corresponding concentration of BID BH3 peptide, which can activate BAX and BAK (25), depleted the mitochondrial pool of cytochrome c (SI Fig. 9).

Thus, obatoclax seems to disrupt MCL-1/BAK interactions in intact mitochondria but BAK at this site is either not optimal for activation in isolated mitochondria under the conditions of this assay or requires a second stimulus to become activated (e.g., provided by tonic signaling pathways in cancer cells), causing release of cytochrome c. In lymphoma cells expressing MCL-1 and treated with obatoclax as a single agent, activation of BAX and BAK and release of cytochrome c from mitochondria was observed, upstream of caspase activation (26). However, to confirm that obatoclax can inhibit endogenous MCL-1/BAK interactions within the cell, SK-Mel5 cells were treated with compound, the cells recovered and washed, the cell pellet dissolved in excess of solution containing the detergent Nonidet P-40, and MCL-1 recovered by overnight immunoprecipitation. The amount of MCL-1 and BAK in the immunoprecipitate was determined by immunoblotting. Despite significant dilution of the cell content of obatoclax during immunoprecipitation, exposure of cells to 1 μM obatoclax for 5 h caused a significant reduction in the amount of BAK recovered with MCL-1, with near complete inhibition recorded at 5 μM (Fig. 1c). Coimmunoprecipitations have also revealed disruption of BCL-2 interactions with BIM (data not shown).

Obatoclax Antagonizes the BCL Pathway, Dependent on BAX or BAK.

Examination of the ability of obatoclax to directly antagonize MCL-1 in mammalian cells is complicated by the fact that multiple indirect pathways, for example proapoptotic stress pathways, can also influence the activity of this BCL member. The yeast, Saccharomyces cerevisiae, however, does not contain any known BCL-2 homolog, but it has been noted that overexpression of either BAX or BAK inhibits its cell growth (27). This inhibition can be countered by expression of MCL-1 (and BCL-2, BCL-XL, or BCL-w) (SI Fig. 10a). At 50 μM, obatoclax did not affect the growth of control yeast cells; however, it significantly overcame the ability of the pro-survival members to antagonize BAX- or BAK-mediated inhibition of cell growth. Although 50 μM is a relatively high external concentration of the compound, the proteins were expressed by using high-copy number plasmids; moreover, uptake and retention of drugs by yeast differs from that of mammalian cells and is consistent with the finding that yeast cells typically require at least 100- to 1,000-fold higher concentrations of anti-cancer compounds to inhibit cell growth compared with the concentrations that are effective in mammalian cells (28). Examples pertaining to camptothecin and doxorubicin relative to several human cell lines are provided in SI Fig. 10b. Importantly, expression of each of the pro-survival BCL members in the absence of BAX or BAK did not confer sensitivity of the yeast cells to obatoclax, indicating that the effect of the compound on yeast cell growth depended on the presence of BAX or BAK.

The findings in yeast cells suggest that mammalian cells lacking the expression of BAX and BAK should exhibit resistance to apoptosis in response to obatoclax. This prediction was confirmed in E1A/p53DN-transformed kidney epithelial cells derived from the Bax,Bak double knock out mouse (29). Compared with transformed WT cells, the DKO cells resisted the activation of caspase-3 (Fig. 2a) and the cleavage of DNA into oligo-nucleosomal fragments in response to obatoclax (Fig. 2b), both of which are hallmarks of apoptosis. Moreover, the ability of obatoclax to induce apoptosis in the Bax,Bak-expressing WT cells also extended to its ability to inhibit colony formation of malignant cancer cells. Such activity was examined in the murine Eμ-Myc lymphoma cell model (30, 31), in which protection against apoptotic stress mediated by Myc expression was provided by overexpression of MCL-1, BCL-2, or A1 (Fig. 2c). In all three cell lines, colony formation and survival was strongly inhibited by obatoclax.

Open in a separate window

Fig. 2.

Induction of apoptosis and inhibition of colony formation by obatoclax. (a) Deletion of Bax and Bak causes resistance to obatoclax. Baby kidney epithelial cell lines derived from the WT and Bax,Bak double knockout (DKO) mouse were treated with vehicle (DMSO) or vehicle containing 2.0 μM obatoclax or 0.5 μM staurosporin (STS), and 20 h later the processed large subunit of caspase-3 (arrow) was detected by SDS/PAGE and immunoblot analysis. (b) As in a, except that oligonucleosomal fragments were detected by agarose gel electrophoresis and staining with ethidium bromide after treatments with 0.05 μM (lanes 3 and 10), 0.2 μM (lanes 4 and 11) and 1 μM (lanes 5 and 12) STS for 16 h; obatoclax treatment was 0.5 μM (lanes 6 and 14), 1 μM (lanes 7 and 15) and 2 μM (lanes 8 and 16) for 24 h. (c) Sensitivity of murine Eμ Myc B cell lymphoma cell lines overexpressing Bcl-2, Mcl-1, or A1 to obatoclax. Cells were exposed to 1 μM obatoclax or ABT737 for 24 h, the media (containing compound) diluted 30-fold, and colony formation on soft agar recorded after 7 days, as described in ref. 45.

Obatoclax Overcomes Resistance to Apoptosis Conferred by MCL-1.

In contrast to obatoclax, the BH3 mimetic ABT-737 inhibited colony formation of Eμ-Myc lymphoma cells that overexpress BCL-2, but not those overexpressing MCL-1 or A1 (Fig. 2c). This result is consistent with both the finding that ABT-737 does not antagonize A1 or MCL-1 (16, 21) and the observation that MCL-1 can confer cellular resistance to ABT-737-induced apoptosis (32, 33). Such a resistance mechanism to ABT-737 can thus be exploited to determine whether obatoclax can overcome MCL-1-mediated resistance to apoptosis. To that end, a KB oral carcinoma cell line was generated that overexpresses ectopic BCL-2 (to restrict cell death) but also expresses sufficient endogenous MCL-1 to restrict cytotoxicity conferred by ABT-737 (Fig. 3 a–c). The KB/BCL-2 model for MCL-1-mediated resistance to ABT-737 was validated by siRNA-mediated knockdown of MCL-1 (Fig. 3 b and c). As judged by annexin V reactivity and uptake of propidium iodide, neither knock down of MCL-1 alone nor treatment of the cells with 0.5 μM ABT-737 alone induced cell death; however, combining the two treatments was strongly cytotoxic (Fig. 3c).

Open in a separate window

Fig. 3.

BCL-2-overexpressing KB cells whose endogenous MCL-1 confers resistance to ABT-737. (a) Stable overexpression of BCL-2 in KB cells. (b) Reduction of MCL-1 expression in KB/BCL-2 cells by MCL-1 siRNA. (c) MCL-1 knockdown confers sensitivity to ABT-737 in KB/BCL-2 cells. KB/BCL-2 cells were transfected with human MCL-1 siRNA or luciferase (Luc) siRNA and plated into media containing 0.5 μM ABT-737. After 72, the percentage of cells incorporating annexin V FITC and propidium iodide was determined by flow cytometry.

The KB/BCL-2 cells were also resistant to 0.1 μM obatoclax. However, like MCL-1 siRNA, 0.1 μM obatoclax overcame MCL-1-mediated resistance to ABT-737, leading to cell death in combination (Fig. 4a). The treatment with obatoclax had no effect on MCL-1 levels but among those BH3-only members that have been reported to antagonize MCL-1 (BID, NOXA, BIM, and PUMA; ref. 24), a modest increase in BIM was noted (Fig. 4b). However, E1A/p53DN transformed kidney epithelial cells derived from the Bim^−/− mouse were as sensitive to obatoclax-mediated killing as were WT cells (Fig. 4c), whereas Bim deletion conferred partial resistance to paclitaxel, as described in ref. 34. Although BIM is not required for obatoclax-mediated cytotoxicity, its increase in response to obatoclax could potentiate the inhibition of MCL-1 by the drug.

Open in a separate window

Fig. 4.

Obatoclax overcomes the resistance of KB/BCL-2 cells to ABT-737. (a) KB/BCL-2 cells were incubated with the indicated concentrations of ABT-737 in the absence or presence of 0.1 μM obatoclax. The percentage of cells that took up propidium iodide was determined by flow cytometry. (b) Influence of obatoclax on expression of MCL-1 and BH3-only antagonists of MCL-1 in KB/BCL-2 cells. After treatment with or without 0.1 μM obatoclax for 72 h, 25 μg of total cell lysate was analyzed by SDS/PAGE and immunoblot for the indicated proteins. (c) Bim deletion does not interfere with obatoclax cytotoxicity. Bim^−/− or Bim^+/+ transformed baby mouse kidney cells were treated with obatoclax or paclitaxel at the indicated concentrations for 72 h and cell viability determined. Standard error is shown (bars).

Obatoclax Overcomes MCL-1-Mediated Resistance to the Proteasome Inhibitor Bortezomib in Melanoma.

A unique property of MCL-1, relative to other pro-survival members of the BCL-2 family, is its rapid steady-state turnover by the 26S proteasome complex (18, 20), resulting from constitutive ubiquitination of MCL-1 by the E3 ligase MULE/LASU1 (35, 36). Interference with either the E3 ligase or the activity of the proteasome itself can result in significant elevation in the level of the MCL-1 protein. Moreover, such elevation of MCL-1 can interfere with the therapeutic response to the proteasome inhibitor bortezomib (Velcade) in cells derived from malignancies for which the drug is currently approved for treatment, multiple myeloma (37) and mantle cell lymphoma (26). In the case of both multiple myeloma (38) and mantle cell lymphoma (39), combining bortezomib with obatoclax partially overcame resistance to bortezomib and resulted in an enhanced cytotoxic response. The ability of obatoclax to overcome MCL-1-mediated resistance to bortezomib, therefore, could ameliorate the response barrier to the use of bortezomib as a single agent in other indications.

An example is illustrated in the melanoma cell line B16-F1 (Fig. 5). This cell line was selected because bortezomib was found not to influence Noxa (data not shown), which, like MCL-1, can also be elevated in response to proteasome inhibition (39, 40). This cell line was particularly resistant to bortezomib even at relatively high concentrations of the drug. Bortezomib caused elevation of MCL-1 to near maximum levels at 100 nM of drug (Fig. 5 Top) and the cells resisted treatment with the drug at concentrations as high as 500 nM for 24 h (Fig. 5 Middle and Bottom). MCL-1 siRNA prevented the induction of MCL-1 protein levels over a wide range of bortezomib concentrations (Fig. 5 Top) and sensitized the cells to induction of apoptosis by 100 nM bortezomib, as measured by annexin V staining, while having no effect on its own (Fig. 5 Middle). Because obatoclax can antagonize MCL-1, it would be expected to have an effect similar to that of MCL-1 siRNA. A concentration of obatoclax was selected that alone had only modest effects on annexin V staining (Fig. 5 Bottom, left side of histogram), but it was found to potentiate either 50 or 500 nM bortezomib. Thus, melanoma as well as potentially other cancer indications could benefit from the combination of a proteasome inhibitor and an MCL-1 antagonist.

Open in a separate window

Fig. 5.

Obatoclax overcomes MCL-1-mediated resistance to bortezomib in melanoma cells. (Top) Mouse melanoma B16-F1 cells were transfected with luciferase siRNA or mouse Mcl-1 siRNA for 24 h followed by a 24-h treatment with bortezomib at the indicated concentrations. Total cell lysates were analyzed by immunoblot for Mcl-1 expression. (Middle) As in Top, except that the percentage of cells that incorporated annexin V was determined by flow cytometry. (Bottom) B16-F1 cells were incubated with bortezomib alone or in combination with 50 nM obatoclax for 24 h and apoptosis (annexin V staining) determined as in Middle.

Obatoclax Exhibits Single-Agent Antitumor Activity.

Multiple cancer cell lines exhibit cytotoxicity in response to treatment with obatoclax. When formulated for i.v. administration, obatoclax was found to have single-agent antitumor effects in several standard mouse tumor models. Examples are provided in SI Fig. 11 a and b. Antitumor activity without animal weight loss was observed in mice bearing solid tumors, including those derived from mouse mammary and human colon, prostate and cervical carcinoma cell lines.

Coimmunoprecipitation from Cells.

SK-Mel5 cells were incubated at 37°C for 5 h with obatoclax or DMSO. Cell pellets were resuspended in ice-cold 10 mM Tris (pH 8.0), 60 mM KCl, 1 mM EDTA, 1 mM DTT, 0.5% Nonidet P-40 (IGEPAL), 10 mM sodium vanadate, 50 mM NaF, 0.1 mM PEFABLOC, 1 μg/ml pepstatin A and leupeptin, and 10 μg/ml aprotinin and gently pipetted up and down 10–15 times. The samples were incubated for 10 min on ice and then centrifuged at 10,000 × g for 15 min at 4°C. Cell extracts (250 μg of protein) were brought to a volume of 1 ml with immunoprecipitation (IP) buffer (50 mM Hepes, pH 7.7, 150 mM NaCl, 0.1% Nonidet P-40 (IGEPAL), 5 mM EDTA) and incubated with 2 μg of rabbit anti-MCL-1 antibody (SC-819; Santa Cruz Biotechnology) overnight at 4°C. Immune complexes were precipitated by incubation with protein A-coated Sepharose beads previously equilibrated with IP buffer, and washed five times with 1 ml of IP buffer. Immunoprecipitated proteins were separated on a 12.5% SDS/PAGE gel, and analyzed by Western blot using the following antibodies: mouse anti-MCL-1 antibody (PharMingen) (1/1,000 dilution); rabbit anti-BAK (Upstate) antibody (1/1,000 dilution); and HRP-conjugated goat anti-rabbit or anti-mouse antibody (1/5,000 dilution). Proteins were detected by using the ECL plus Western blot detection kit (Amersham).

siRNA Transfection.

Pools of four siRNAs against human MCL-1 or mouse Mcl-1 mRNAs were purchased from Dharmacon. KB/BCL-2 cells were transfected twice with the pool of human MCL-1 siRNA or luciferase siRNA using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Twenty-four hours after the second transfection, cells were plated into media containing the compounds. Seventy-two hours later, cells were analyzed by flow cytometry. For mouse B16-F1 cells, pools of mouse Mcl-1 siRNAs or luciferase siRNAs were transfected once by using Lipofectamine 2000. Twenty-four hours after transfection, cells were plated into medium containing the compounds; cells were collected the following day and analyzed by flow cytometry.

We thank Eileen White (Rutgers University, Piscataway, NJ) for kindly providing WT and Bax,Bak double knockout mouse epithelial cell lines. We thank the following for contributions: Salah Aouad, Angela Babineau, Sylvie Baillie, Mark Berger, Geneviève Brindle, Helen Chan, Gang Chen, Gaétan Gagnon, Michel-Olivier Gratton, Jeff Hincks, Anne Jang, Abdelkrim Khadir, Mark Mariani, Sandra Naranjo, Youssouf Soumounou, Brigitte St.-Denis, Nancy Steenaart, Émilie Turcotte, Zhiying Zhang, Giorgio Attardo, Anne Bélanger, Stéphane Branchaud, Sylvie Charron, Kenza Dairi, Lionel Dumas, Walid Elbast, Noura El-Djabri, Éric Fournier, Gerson Gonzalez, Andrea Hill, Jean-François Lavallée, Manon Lavoie, Daniel Rabouin, Élise Rioux, and Sasmita Tripathy. This work was supported in part by research grants from the Canadian Institutes of Health Research, the Leukemia Foundation of Australia, Cancer Council Victoria, the Canadian Cancer Society, and the National Health and Medical Research Council of Australia.