|Home | About | Journals | Submit | Contact Us | Français|
The Bcl-2 antagonist ABT-737 kills transformed cells in association with displacement of Bim from Bcl-2. The histone deactetylase (HDAC) inhibitor suberoyl bis-hydroxamic acid (SBHA) was employed to determine whether and by what mechanism ABT-737 might interact with agents that upregulate Bim. Expression profiling of BH3-only proteins indicated that SBHA increased Bim, Puma, and Noxa expression, while SBHA concentrations that upregulated Bim significantly potentiated ABT-737 lethality. Concordance between SBHA-mediated Bim upregulation and interactions with ABT-737 was observed in various human leukemia and myeloma cells. SBHA-induced Bim was largely sequestered by Bcl-2 and Bcl-xL, rather than Mcl-1; ABT-737 attenuated these interactions, thereby triggering Bak/Bax activation and mitochondrial outer membrane permeabilization. Knockdown of Bim (but not Puma or Noxa) by shRNA or ectopic overexpression of Bcl-2, Bcl-xL, or Mcl-1 diminished Bax/Bak activation and apoptosis. Notably, ectopic expression of these antiapoptotic proteins disabled death signaling by sequestering different proapoptotic proteins, i.e., Bim by Bcl-2, both Bim and Bak by Bcl-xL, and Bak by Mcl-1. Together, these findings indicate that HDAC inhibitor-inducible Bim is primarily neutralized by Bcl-2 and Bcl-xL, thus providing a mechanistic framework by which Bcl-2 antagonists potentiate the lethality of agents, such as HDAC inhibitors, which upregulate Bim.
Cell death is regulated by complex interactions between members of the Bcl-2 family. The multidomain proapoptotic proteins Bax and Bak, when engaged, trigger mitochondrial outer membrane permeabilization (MOMP), which results in release of proapoptotic proteins (e.g., cytochrome c) from the mitochondria to the cytosol, thereby initiating the caspase cascade, which culminates in cellular demise (58). BH3-only proapoptotic family members include Bid, Bim, Noxa, Puma, Bad, Bik, Bmf, and Hrk and are responsible for conversion of various cellular insults into death signals (31) through a process that exhibits an absolute requirement for the multidomain proapoptotic proteins Bax and Bak (27, 74). Among BH3-only proteins, Bim and Bid have been classified as “activators” in view of their purported ability to engage directly and activate Bax and Bak (43). In contrast, other BH3-only proteins do not directly activate Bax and Bak; instead, they act indirectly by neutralizing antiapoptotic proteins, i.e., Bcl-2 and Bcl-xL (e.g., by Bad) and Mcl-1 (e.g., by Noxa) (8), and are classified as “sensitizers” or “derepressors” (43). One possible exception to the classification of “sensitizers” is Puma, which may act, at least in certain settings, as an “activator” (5, 36). Multidomain antiapoptotic members of the Bcl-2 family include Bcl-2, Bcl-xL, Bcl-w, Mcl-1, and A1/BFL1. Such family members govern apoptotic signaling through interactions with proapoptotic proteins, including Bax/Bak and/or BH3-only “activators” (27, 43). There is currently considerable debate regarding whether Bax and Bak must first be activated to initiate MOMP or whether they are constitutively activated, but under repressive control by antiapoptotic proteins that must be neutralized (i.e., by BH3-only proteins) for cell death to occur (47). To add to the complexity, it has recently been reported that activation of Bax and apoptosis can occur even in the absence of the “activators” Bid and Bim, suggesting the existence of other unknown cell death mechanisms operating independently of Bid and Bim (69). Despite this uncertainty, it is clear that Bim plays a critical role in the apoptotic regulatory machinery engaged by diverse environmental insults, particularly those involving anticancer agents (47). Thus far, three Bim isoforms have been identified (e.g., BimEL, BimL, and BimS), which vary functionally as well as in their tissue-specific expression (52).
ABT-737 is a small-molecule BH3-only mimetic that recapitulates the capacity of BH3-only proteins (e.g., Bad) to bind to the hydrophobic clefts of Bcl-2, Bcl-xL, and Bcl-w, thereby disrupting their antiapoptotic functions (53). It shows in vitro and in vivo activities against various transformed cells while exhibiting minimal toxicity toward normal cells (7, 29, 53). ABT-737 effectively antagonizes the actions of Bcl-2 and Bcl-xL but minimally affects Mcl-1 function (13). Recent studies suggested that the relative expression levels of Bcl-2/Bcl-xL versus Mcl-1 largely determine the susceptibility of transformed cells to ABT-737 (9, 37, 49, 62, 65). In addition, several groups have demonstrated that in various tumor cell types, interventions that downregulate Mcl-1 expression dramatically enhance ABT-737 lethality (9, 35, 37, 49, 65). Notably, ABT-737 displaces Bim from the BH3-binding pocket of Bcl-2, allowing Bim to activate Bax and induce MOMP (18). Thus, the extent of Bcl-2 bound to Bim, rather than total Bcl-2 expression levels, may determine cellular sensitivity to ABT-737 (18, 20). In this regard, ABT-737 has been shown to interact with certain anticancer agents capable of upregulating Bim, (21, 38, 39, 72). However, whether and how Bim upregulation plays a functional role in interactions between such agents has not yet been determined with certainty.
Histone deacetylase (HDAC) inhibitors represent a class of epigenetically acting agents known to upregulate Bim. Histone acetylation is regulated by the reciprocal actions of histone acetyltransferases and histone deacetylases. Such posttranslational histone modifications comprise a component of the histone code, an important regulator of gene transcription (3). Exposure to HDAC inhibitors results in acetylation of histone tails, leading to a more open chromatin structure conducive to the transcription of genes involved in cellular differentiation and cell death (3). However, it has been reported that HDAC inhibitors kill malignant cells through diverse mechanisms (17), including induction of oxidative injury, disruption of cell cycle checkpoints, and acetylation of nonhistone proteins, among others. Notably, it has recently been reported that exposure to HDAC inhibitors induces Bim upregulation via an E2F1-dependent mechanism (73). This phenomenon has been postulated to contribute to the lethality of HDAC inhibitors, administered either alone (33, 50, 73) or in combination with other agents (14, 15, 24).
The ability of ABT-737 to displace Bim from Bcl-2 raised the possibility that ABT-737 might enhance the activity of anticancer agents such as HDAC inhibitors which are capable of increasing Bim expression. To test this hypothesis, interactions between ABT-737 and the hydroxamate pan-HDAC inhibitor suberoyl bis-hydroxamic acid (SBHA) were examined in human leukemia and myeloma cells. The present results indicate that SBHA markedly induces Bim expression in these cells and that Bim upregulation plays a critical functional role in synergistic interactions between SBHA and ABT-737. Interestingly, it was observed that upregulated Bim was primarily bound to/sequestered by Bcl-2 and Bcl-xL rather than Mcl-1 and that coadministration of ABT-737 substantially diminished the association of Bim with both Bcl-2 and Bcl-xL but not with Mcl-1. Together, these findings provide a potential mechanism accounting for interactions between Bcl-2 antagonists like ABT-737 and anticancer agents such as HDAC inhibitors which act, at least in part, through Bim upregulation.
Human leukemia U937, HL-60, and Jurkat cells and human multiple myeloma U266 and RPMI 8226 cells were from ATCC and maintained in RPMI 1640 medium containing 10% fetal calf serum as previously reported (16). U937/Bcl-2 and U937/Bcl-xL were obtained by stable transfection of cells with full-length Bcl-2 and Bcl-xL cDNA, respectively (16). U937 cells stably overexpressing Mcl-1 were kindly provided by Ruth Craig (Dartmouth Medical School, Hanover, NH) (9). Wild-type and Bax/Bak knockout mouse embryonic fibroblasts (MEFs) were kindly provided by the laboratory of Stanley Korsmeyer (Dana-Farber Cancer Institute, Boston, MA) (67). All experiments utilized logarithmically growing cells (3 × 105 to 5 × 105 cells/ml). Peripheral blood samples were obtained with informed consent according to the Declaration of Helsinki from four patients with acute myeloid leukemia (AML; FAB subtype M2) undergoing routine diagnostic aspirations, with approval from the Virginia Commonwealth University Institutional Review Board. Primary leukemic cells were isolated as previously described (15).
The Bcl-2/Bcl-xL/Bcl-w antagonist ABT-737 was kindly provided by Gary Gordon (Abbott Laboratories, Abbott Park, IL) (53). It was dissolved in dimethyl sulfoxide (DMSO), aliquoted, and stored at −80°C. The pan-HDAC inhibitors SBHA (4) and oxamflatin were purchased from Calbiochem (San Diego, CA) and dissolved in sterile DMSO, aliquoted, and stored at −20°C. In all experiments, the final concentration of DMSO did not exceed 0.1%.
The extent of apoptosis was evaluated by flow cytometric analysis utilizing annexin V-fluorescein isothiocyanate (FITC)-propidium iodide or 3,3-dihexyloxacarbocyanine (DiOC6)-7-amino actinomycin D (7AAD) staining as described previously (15). Briefly, 1 × 106 cells were stained with annexin V-FITC (BD PharMingen) and 5 μg/ml propidium iodide (Sigma) in 1× binding buffer for 15 min at room temperature in the dark. Samples were then analyzed by flow cytometry within 1 h to determine the percentage of cells displaying annexin V positivity. In some cases, mitochondrial injury and cell death were assessed by double staining with 40 nM DiOC6 (Molecular Probes Inc., Eugene, OR) and 0.5 μg/ml 7AAD (Sigma) in phosphate-buffered saline at 37°C for 20 min and then analyzed using a Becton-Dickinson FACScan apparatus (Becton-Dickinson, San Jose, CA).
Samples for immunoblotting were prepared from whole-cell pellets as described previously (15). Total protein was quantified using Coomassie protein assay reagent (Pierce, Rockford, IL). An equal amount of protein (30 μg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransferred onto nitrocellulose membrane. Where indicated, the blots were reprobed with antibodies against β-actin (Sigma, St. Louis, MO) or α-tubulin (Oncogene, La Jolla, CA) to ensure equal loading and transfer of proteins. The following antibodies were used as primary antibodies: BH3-only protein detection set, anti-Bim (rabbit), anti-Noxa (rabbit), and anti-Puma (rabbit) (ProSci Inc., Poway, CA); anti-Bim (rat; Calbiochem); anti-Mcl-1, anti-caspase 9, and anti-caspase 3 (BD PharMingen, San Diego, CA); anti-Noxa (goat), anti-Pumaα (goat), anti-Pumaβ/δ (goat), anti-Bak, and anti-Bax (Santa Cruz Biotechnology, Santa Cruz, CA); anti-cleaved caspase 3 (Asp175), anti-cleaved caspase 9 (Asp315), anti-cleaved poly(ADP-ribose) polymerase (PARP; Asp214), and anti-Bcl-xL (Cell Signaling, Beverly, MA); anti-human Bcl-2 oncoprotein (Dako, Carpinteria, CA); anti-PARP (Biomol, Plymouth Meeting, PA). For expression profiling of BH3-only proteins, the densities of blots were quantified using a FluoChem 8800 imaging system and AlphaEaseFC software (Alpha Innotech, San Leandro, CA).
Interactions between BH3-only proteins and Bcl-2, Bcl-xL, or Mcl-1 were evaluated by coimmunoprecipitation analysis. For these studies, 3-[(3-cholamidopropyl)-diethylammonio]-1-propanesulfonate (CHAPS) buffer (150 mmol/liter NaCl, 10 mmol/liter HEPES, pH 7.4, protease inhibitors, and 1% CHAPS) was employed to avoid artifactual associations reported with buffers containing other detergents (e.g., NP-40 or Triton X-100) (30). Briefly, cells were lysed in CHAPS buffer and 200 μg of protein per condition was incubated with 1 μg anti-Bim (Santa Cruz Biotechnology), anti-Bcl-2 (Dako), anti-Bcl-xL (Cell Signaling), or anit-Mcl-1 (BD PharMingen) overnight at 4°C. Twenty microliters per reaction mixture per condition of Dynabeads (Dynal, Oslo, Norway) was then added and incubated for an additional 4 h. After washing, the bead-bound protein was eluted by vortexing and boiling in 20 μl 1× sample buffer. The samples were separated by SDS-PAGE and subjected to immunoblot analysis as described above. Anti-Bim (rat; Calbiochem), anti-Bcl-2 (Dako), anti-Mcl-1 (BD PharMingen), anti-Noxa (ProSci), and anti-Puma (ProSci) were used as primary antibodies.
A total of 2 × 106 cells were lysed in digitonin lysis buffer (54). Lysates were centrifuged, and the supernatant (S-100 cytosolic fraction) was collected and added to an equal volume of 2× sample buffer. The pellets (organelle/membrane fractions) were washed once in cold phosphate-buffered saline and lysed in 1× sample buffer. The S-100 fraction and pellet samples were quantified, separated by SDS-PAGE, and subjected to immunoblot analysis. For analysis of release of mitochondrial proapoptotic factors, anti-cytochrome c (BD PharMingen) and anti-apoptosis-inducing factor (anti-AIF; Santa Cruz Biotechnology) were used as primary antibodies. Anti-Bax antibody (Santa Cruz Biotechnology) was employed to evaluate translocation of Bax.
Cells were lysed in 1% CHAPS buffer, and 200 μg of protein was immunoprecipitated using anti-Bax (6A7; Sigma) or anti-Bak (Ab-1; Calbiochem), which only recognizes Bax or Bak that has undergone a conformation change, and Dynal Beads as described above. Immunoprecipitated protein was then subjected to immunoblot analysis by using anti-Bax and anti-Bak (Santa Cruz Biotechnology) as primary antibodies. Alternatively, cells were fixed and permeabilized using the FIX and PERM cell permeabilization reagents (Caltag Lab, Burlingame, CA) as per the manufacturer's instructions. Fixed cells were incubated with either anti-Bak (Ab-1; Calbiochem) or anti-Bax (clone 3; BD Transduction Lab) (68) on ice for 30 min and then with FITC-conjugated goat-anti-mouse immunoglobulin G (IgG; Southern Biotech, Birmingham, AL) for 30 min in the dark. After washing, the samples were analyzed by flow cytometry. For comparison, cells were stained with antibodies recognizing total Bax or Bak. The results for each condition were calibrated relative to values for cells stained with mouse IgG (Southern Biotech) to replace the primary antibody.
The pSUPER.retro.puro vector containing the human H1 RNA promoter for expressing small hairpin RNA (shRNA) was obtained from Oligoengine (Seattle, WA). pSR-Bim and pSR-con constructs, encoding shRNA for Bim (shBim) or scrambled shRNA as a negative control (shNC), were prepared by inserting the target sequence for human Bim (GenBank accession number AF032457, nucleotides 37 to 56; GACCGAGAAGGTAGACAATT) or a scrambled sequence (AATTCTCCGAACGTGTCACGT) into pSUPER.retro.puro (54). SureSilencing shRNA plasmids (neomycin resistance) were purchased from SABioscience (Frederick, MD), which included shBim (human BCL2L11; GAGACGAGTTTAACGCTTACT), shNoxa (human PMAIP1, NM021127; CTCAGCACATTGTATATGATT), shPuma (human BBC3, NM014417; ACCATCTCAGGAAAGGCTGTT), and shNC (GGAATCTCATTCGATGCATAC). U937, Jurkat, and U266 cells were stably transfected with these constructs by using the Amaxa Nucleofector device with cell line-specific Nucleofector kits (Amaxa GmbH, Cologne, Germany) as per the manufacturer's instructions, and clones with downregulated Bim, Noxa, or Puma expression were selected with puromycin for pSUPER.retro.puro vectors (U266; 2 μg/ml) or with G418 for SureSilencing shRNA vectors (U937 and U266, 400 μg/ml; Jurkat, 800 μg/ml).
The reported values represent the means ± standard deviations for at least three independent experiments performed in triplicate. The significance of differences between experimental variables was determined using Student's t test. To characterize the nature of interactions between ABT-737 and SBHA, median dose-effect analysis using Calcusyn software (Biosoft, Ferguson, MO) was performed to determine whether additive, synergistic, or antagonistic interactions occurred over a range of concentrations of the two agents administered at a fixed concentration ratio (15).
BH3-only proteins are functionally divided two groups, (i) “activators” Bid and Bim (including BimEL, BimL, and BimS isoforms), and (ii) “sensitizers/derepressors” Bad, Bik, Noxa, Puma, Hrk, and Bmf (47). In this context, the expression profile of BH3-only proteins in U937 cells exposed to the HDAC inhibitor SBHA was first examined. To this end, U937 cells were untreated or exposed to the indicated concentrations (5 to 30 μM) of SBHA for 24 h and then subjected to immunoblot analysis using rabbit polyclonal antibodies of the BH3-only protein detection set (ProSci). When compared to untreated controls, exposure to SBHA concentrations of ≥15 μM resulted in marked increases in the expression of Bim, particularly BimEL, although upregulation of BimL and BimS was also apparent after longer exposure of blots (Fig. (Fig.1A).1A). However, no change was noted in the expression of Bid, which is primarily involved in the death receptor-initiated extrinsic pathway (48). In addition, SBHA concentrations of ≥5 μM discernibly increased the expression of Noxa and Puma but had little or no effect on levels of Bad, Bik, Bmf, or Hrk (Fig. (Fig.1B).1B). Relative increases in levels of each BH3-only protein were then quantified in relation to SBHA concentration and expressed as the increase versus untreated controls. As shown in Fig. 1C and D, quantified results of BH3-only expression profiles from three separate experiments revealed distinctly different patterns of Bim, Noxa, and Puma expression in SBHA-treated U937 cells, i.e., (i) a dose-dependent induction of BimEL, BimL, and BimS expression occurred at SBHA concentrations of ≥15 μM; (ii) increased expression of Noxa occurred at lower SBHA concentrations (i.e., ≥5 μM) and remained at plateau levels until SBHA concentrations reached 30 μM; and (iii) upregulation of Puma also occurred at SBHA concentrations of ≥5 μM, reaching plateau levels at SBHA concentrations of ≥10 μM. These findings indicate that exposure to SBHA results in increased expression of Bim, Noxa, and Puma, but the dose-dependent nature of these responses differs distinctly between the three proteins.
To determine whether upregulation of BH3-only proteins (i.e., Bim, Noxa, and Puma) by SBHA might be associated with enhanced susceptibility of human leukemia cells to ABT-737, U937 cells were exposed for 24 h to a minimally toxic concentration of ABT-737 (e.g., 300 nM) in the presence or absence of increasing concentrations of SBHA. As shown in Fig. Fig.1E,1E, cotreatment with ≥15 μM SBHA resulted in a marked, dose-dependent increase in ABT-737-mediated cell killing, consistent with the pattern of SBHA-induced increase in Bim expression (Fig. (Fig.1C).1C). In contrast, lower SBHA concentrations (e.g., 5 to 10 μM), which failed to increase Bim expression but significantly upregulated Puma and Noxa levels (Fig. (Fig.1D),1D), did not potentiate ABT-737 lethality. Median dose-effect analysis of cell death induction in U937 cells in which SBHA (20 to 40 μM) was administered at a fixed concentration ratio (100:1) with ABT-737 (200 to 400 nM) yielded combination index (CI) values significantly less than 1.0 (e.g., 0.288 to 0.460), indicating synergistic interactions (Fig. (Fig.1E,1E, inset). In addition, coadministration of another HDAC inhibitor, oxamflatin, also enhanced ABT-737 lethality in U937 cells (Fig. (Fig.1F).1F). Moreover, immunoblot analysis using antibodies from the indicated sources (e.g., rat anti-Bim [Calbiochem] and goat anti-Noxa and anti-Puma [Santa Cruz Biotechnology]) confirmed a marked increase in expression of BimEL, BimL, and BimS in cells exposed to SBHA with or without ABT-737, as well as discernible increases in Puma and Noxa expression (Fig. (Fig.1G).1G). Notably, ABT-737 by itself failed to modify either basal Bim levels or SBHA-induced Bim upregulation. In conjunction with the profiling data for BH3-only protein expression, these findings indicate that the capacity of SBHA to potentiate ABT-737 lethality in human leukemia cells correlates most closely with upregulation of Bim.
Previous studies showed that the sensitivity of leukemia cells to ABT-737 is inversely related to basal levels of Mcl-1 expression (9, 37, 62). Therefore, the question of whether basal expression of Bim might also contribute to ABT-737 sensitivity, and importantly, influence interactions between SBHA and ABT-737 was then examined. To this end, comparisons were made between the ability of SBHA to enhance ABT-737 lethality in multiple human leukemia cell lines expressing disparate basal levels of Mcl-1 and Bim. Interestingly, basal levels of Bim expression (e.g., U937 ≥ Jurkat HL-60) (Fig. (Fig.2A,2A, upper panel) by themselves did not predict the sensitivities of leukemia cells to ABT-737 (HL-60 > Jurkat > U937), which instead were largely determined by basal expression of Mcl-1 (data not shown), consistent with the results described in previous reports (9, 37). Notably, despite the disparate expression of Mcl-1 and Bim in these leukemia cell types, SBHA effectively potentiated ABT-737 lethality in all three cell lines (Fig. (Fig.1E1E and and2B),2B), although the concentration of ABT-737 employed in these studies differed (e.g., U937, ≥200 nM; Jurkat, ≥100 nM; HL-60, ≥50 nM), reflecting the differential ABT-737 sensitivities of these cells, which varied reciprocally with Mcl-1 expression. As observed in U937 cells (Fig. (Fig.1E,1E, inset), a highly synergistic interaction between SBHA (Jurkat, 8 to 24 μM; HL-60, 7.5 to 22.5 μM) and ABT-737 (Jurkat, 100 to 300 nM; HL-60, 50 to 150 nM) was also observed in both Jurkat (CI, 0.080 to 0.252) and HL-60 cells (CI, 0.278 to 0.280) (Fig. (Fig.2B,2B, inset). Moreover, potentiation of ABT-737 lethality by SBHA was also associated with clear evidence of Bim upregulation in both Jurkat and HL-60 cells, whereas no change occurred in the expression of Mcl-1 with any treatment (Fig. (Fig.2C2C).
Analogous studies were then performed in primary blast samples from four patients with AML (FAB M2). As might be anticipated, levels of both Mcl-1 and Bim varied between primary AML specimens obtained from different patients, e.g., in the case of patients 1 and 3 (Fig. (Fig.2A,2A, bottom panels). Although responses to ABT-737 (300 nM) and SBHA (15 to 30 μM) individually also varied between the samples, cotreatment with these agents resulted in a marked increase in lethality in each instance (Fig. (Fig.2D).2D). Notably, immunoblot analysis demonstrated that treatment with SBHA in the presence or absence of ABT-737 resulted in a marked increase in the expression of Bim (specifically, BimEL), accompanied by a pronounced increase in PARP cleavage in primary leukemia blasts coexposed to these agents (Fig. (Fig.2E).2E). Major changes in the expression of Mcl-1 or Bcl-xL were not regularly observed, although modest downregulation of these proteins was occasionally noted in some samples, possibly representing caspase-mediated cleavage, reflected by the appearance of a Bcl-xL cleavage fragment (Fig. (Fig.2E2E).
Finally, to determine whether interactions between SBHA and ABT-737 were restricted to leukemia cells, parallel studies were performed in human myeloma cells. As shown in Fig. Fig.2A2A (upper panel), human myeloma RPMI 8226 and U266 cells exhibited relatively higher levels of Mcl-1, a critical survival factor for this cell type (22), compared with human leukemia U937, Jurkat, and HL-60 cells. Nevertheless, treatment with minimally toxic concentrations of ABT-737 (e.g., 200 to 500 nM) in conjunction with SBHA resulted in a pronounced increase in lethality in both U266 and RPMI 8226 cells (Fig. (Fig.2F),2F), analogous to results obtained in leukemia cells. Median dose-effect analysis of cell death induced by ABT-737 (200 to 400 nM) in conjunction with SBHA (10 to 20 μM) at a fixed concentration ratio (1:50) also demonstrated synergistic interactions in myeloma cells (U266, CI of 0.607 to 0.684; 8226, CI of 0.488 to 0.593) (Fig. (Fig.2F,2F, inset). Moreover, these events were also associated with the clear upregulation of Bim by SBHA, accompanied by increased cleavage of caspase 9 and PARP following coexposure to SBHA and ABT-737 (Fig. (Fig.2G).2G). Whereas no changes in the total levels of Bcl-2, Bcl-xL, or Mcl-1 expression were observed with any treatment, a clear increase in Bcl-2 cleavage occurred in myeloma cells coexposed to both agents (Fig. (Fig.2G).2G). Together, these findings indicate that potentiation of ABT-737 lethality by SBHA appears closely related to Bim upregulation in various human leukemia cell types exhibiting diverse basal levels of Bim and Mcl-1 expression, as well as in human myeloma cells exhibiting high levels of expression of Mcl-1.
The preceding data indicated that while SBHA-mediated Bim upregulation was not modified by ABT-737, pronounced lethality was only observed in cells cotreated with both agents, raising the possibility that SBHA-induced Bim might be sequestered/inactivated by antiapoptotic proteins. In this context, previous studies demonstrated that Bim binds to all antiapoptotic proteins in in vitro assays, with dissociation constants of <10 nM for Bcl-2, Bcl-xL, and Mcl-1 (6, 8). To investigate this possibility, coimmunoprecipitation approaches were employed using CHAPS buffer to avoid artifactual associations caused by other detergents (30). In untreated U937 cells, Bim was predominantly coimmunoprecipitated by Bcl-2 and Bcl-xL and to a lesser extent by Mcl-1 (Fig. (Fig.3A).3A). Notably, exposure of U937 cells to SBHA not only induced Bim upregulation (Fig. 1A and G) but also led to a marked increase in the amount of Bim bound to both Bcl-2 (Fig. (Fig.3B)3B) and Bcl-xL (Fig. (Fig.3C),3C), but not Mcl-1 (Fig. (Fig.3D).3D). This indicates that upregulated Bim was primarily sequestered by Bcl-2 and Bcl-xL, rather than by Mcl-1. None of the treatments appreciably modified total expression of these proteins, although a Bcl-2 cleavage fragment was observed in cells cotreated with SBHA and ABT-737. Notably, exposure to ABT-737 (e.g., 300 or 500 nM) resulted in a striking reduction in basal Bim/Bcl-2 and Bim/Bcl-xL associations (Fig. 3B and C), findings consistent with previous reports (18, 32). Importantly, coadministration of ABT-737 substantially diminished the association of upregulated Bim with both Bcl-2 and Bcl-xL in SBHA-treated cells (Fig. 3B and C).
Coimmunoprecipitation was also performed to determine whether ABT-737-mediated release of Bim from binding by Bcl-2 and Bcl-xL might contribute to synergistic interactions between this agent and SBHA. To this end, U937 cells were exposed to a series of concentrations (e.g., 10 to 500 nM) of ABT-737 in the absence or presence of SBHA (30 μM). In cells exposed to SBHA, ABT-737-mediated Bim/Bcl-xL dissociation was discernible at 100 nM and pronounced at ≥300 nM, whereas ABT-737 concentrations of ≥50 nM substantially diminished Bim/Bcl-2 binding (Fig. (Fig.3E).3E). In parallel, flow cytometric analysis demonstrated that ABT-737 concentrations of ≥100 nM interacted with SBHA to induce a significant increase in cell death (P < 0.05 for 100 nM and P < 0.01 for 300 to 500 nM, compared with SBHA alone). These results were confirmed by immunoblot analysis monitoring PARP cleavage (Fig. (Fig.3E).3E). Median dose analysis (Fig. (Fig.1E,1E, inset) revealed synergistic interactions between ABT-737 and SBHA over a range of ABT-737 concentrations (e.g., 200 to 400 nM) capable of disrupting binding of Bim by both Bcl-2 and Bcl-xL.
Parallel studies were performed in other human leukemia cells (e.g., Jurkat and HL-60) as well as myeloma cells (e.g., U266 and RPMI 8226). As noted in U937 cells, exposure to SBHA resulted in a clear increase in binding of Bim to both Bcl-2 and Bcl-xL rather than to Mcl-1 in leukemia cells (Jurkat and HL-60 [Fig. [Fig.4A])4A]) and myeloma cells (U266 and RPMI 8226 [Fig. [Fig.4B]).4B]). In addition to the BimEL isoform, increased binding of BimL to Bcl-2 was also noted in certain cell types, such as HL-60 (Fig. (Fig.4A,4A, upper left panel) and U266 (data not shown). Interestingly, exposure to ABT-737 alone modestly increased Mcl-1/Bim complex formation in HL-60 cells (Fig. (Fig.4A,4A, left bottom panel) while slightly decreasing or exerting no discernible effect on Mcl-1/Bim binding in U937 cells (Fig. (Fig.3D)3D) or Jurkat cells (Fig. (Fig.4A,4A, right bottom panel), respectively. It is possible that the former phenomenon may reflect a cell-type-dependent compensatory response to displacement of Bim from Bcl-2/Bcl-xL by ABT-737. In addition, coadministration of SBHA diminished Bim/Mcl-1 binding in HL-60 cells (Fig. (Fig.4A)4A) through a yet-to-be-determined mechanism. Nevertheless, coadministration of ABT-737, administered at various concentrations depending upon the cell type, dramatically disrupted associations between Bim (including both BimEL and BimL isoforms) and Bcl-2 or Bcl-xL (Fig. (Fig.4).4). Together, these findings suggest that in human leukemia and myeloma cells, SBHA-induced Bim is primarily sequestered by Bcl-2 and Bcl-xL rather than by Mcl-1 and that both of these associations are disrupted by ABT-737. They also raise the possibility that ABT-737 may cooperate with SBHA to trigger cell death by freeing upregulated Bim from its inactivating associations with Bcl-2 and Bcl-xL.
Efforts were then undertaken to determine whether release of Bim from binding to Bcl-2 and Bcl-xL by ABT-737 might be involved in engagement of the apoptotic signaling cascade. To this end, immunoprecipitation using antibodies specifically recognizing conformationally changed/active forms of Bax (6A7) or Bak (Ab-1) followed by immunoblotting with antibodies directed against total Bax or Bak was employed to detect conformational changes in Bak and Bax. As shown in Fig. Fig.5A,5A, exposure to ABT-737 resulted in a modest increase in conformational changes of Bax but not Bak as previously described (9), while cotreatment with SBHA led to a pronounced increase in conformational changes of both Bax and Bak. Moreover, cotreatment with SBHA and ABT-737 resulted in the marked translocation of Bax from the cytosol to the pellet (i.e., the organelle/membrane fraction containing mitochondria), without modifying total Bax levels (Fig. (Fig.5B).5B). Total Bak protein levels remained unchanged with all treatments. Parallel blots for Bak and α-tubulin documented equivalent loading of samples and the absence of contamination between the two fractions (Fig. (Fig.5B).5B). Furthermore, coexposure to SBHA and ABT-737 resulted in a dramatic increase in MOMP, manifested by both loss of mitochondrial membrane potential (reflected by “low” uptake of DiOC6 [data not shown]) and release of the mitochondrial proapoptotic proteins cytochrome c and AIF (Fig. (Fig.5C,5C, upper panels). These events were accompanied by the pronounced cleavage/activation of caspases 3 and 9, as well as PARP degradation (Fig. (Fig.5C,5C, lower panels). In addition, Bax/Bak double knockout (Bax−/− Bak−/−) MEFs were fully resistant to cell death induced by cotreatment with SBHA and ABT-737, while either Bax−/− or Bak−/− MEFs displayed only partial resistance (data not shown). Taken in conjunction with previous findings, these results support the notion that ABT-737 displaces upregulated Bim from Bcl-2 and Bcl-xL in SBHA-treated cells, thereby triggering Bak and Bax activation, leading in turn to engagement of the mitochondrial/intrinsic apoptotic cascade.
The functional significance of Bim upregulation in interactions between SBHA and ABT-737 was investigated further. To this end, stable cell lines in which Bim was knocked down by shRNA were established in human leukemia (U937 and Jurkat) and myeloma (U266) cells. Cells transfected with shBim displayed a pronounced reduction in expression of Bim (EL, L, and S isoforms) compared to cells transfected with shNC (Fig. (Fig.6A)6A) as well as their parental counterparts (data not shown). The effects of Bim knockdown by shRNA on SBHA-induced Bim expression and the lethality of SBHA in the presence or absence ABT-737 were then examined. Notably, shRNA knockdown of Bim not only conferred significant resistance to the apoptosis induction by SBHA administered alone but also almost completely abrogated the interaction between SBHA and ABT-737, manifested by strikingly diminished PARP cleavage (U937 [Fig. [Fig.6B],6B], Jurkat [Fig. [Fig.6C],6C], and U266 [Fig. [Fig.6D]),6D]), as well as cell death (7AAD+) (Fig. (Fig.7A,7A, B, and C) and loss of Δψm (data not shown). Consistent with these findings, Bim shRNA dramatically blocked BimEL upregulation by SBHA in all three cell lines (U937 [Fig. [Fig.6B],6B], Jurkat [Fig. [Fig.6C],6C], and U266 [Fig. [Fig.6D]).6D]). In contrast, expression of Mcl-1, Bcl-2, and Bcl-xL remained unchanged in cells expressing either control shRNA or Bim shRNA under all treatment conditions (data not shown), as observed in the case of their parental counterparts (Fig. (Fig.22 and and33).
Finally, to elucidate how Bim shRNA might prevent cell death, activation levels of Bax and Bak were examined by monitoring conformational change of these proteins by both immunoprecipitation and flow cytometry using antibodies (Ab-1 in the case of Bak and clone 3 in the case of Bax) that only recognize the conformationally active forms (68). As shown in Fig. Fig.7D,7D, exposure to SBHA (most clearly at 30 μM) or ABT-737 individually resulted in a modest increase in conformationally changed Bak and Bax in U937 cells expressing control shRNA, consistent with results obtained in parental U937 cells (Fig. (Fig.5A).5A). Significantly, Bim knockdown by shRNA essentially abrogated Bak or Bax conformational changes induced by either SBHA alone or in combination with ABT-737 (Fig. (Fig.7D).7D). Virtually identical results for Bak conformational change were obtained in Jurkat cells transfected with control and Bim shRNA (Fig. (Fig.7E,7E, upper panel). In contrast, neither Bax in whole-cell lysates nor conformationally active Bax, as determined by immunoprecipitation, could be detected in these cells (Fig. (Fig.7E,7E, lower panel), consistent with previously published results involving Jurkat cells (60). Lastly, similar phenomena were also observed in U266 cells transfected with Bim shRNA in which flow cytometry was employed to monitor conformational changes of Bax and Bak (Fig. (Fig.7F),7F), whereas no change was observed when antibodies against total Bax or Bak were employed as primary antibodies to replace clone 3 or Ab-1, respectively (data not shown). Together, these findings argue strongly that Bim upregulation by SBHA plays a critical functional role in potentiating ABT-737 lethality through activation of Bak and Bax.
In addition to Bim, the expression profile of BH3-only proteins demonstrated that Noxa and Puma were also clearly upregulated in U937 cells exposed to SBHA (Fig. (Fig.1B).1B). Consequently, studies were then performed to determine whether treatment with SBHA and ABT-737 alone or in combination might influence the associations between Mcl-1 and Noxa or Puma. Such associations are known to play important roles in regulating Mcl-1 expression and function in the case of Noxa (8), as well as the ability of Puma to induce apoptosis (51). Unexpectedly, while in vitro binding studies and coimmunoprecipitation analyses have demonstrated that Noxa and Puma are able to bind to Mcl-1 in 293T cells transfected with wild-type Noxa (8) and colorectal cancer cell line Puma+/+ HCT116 (51), respectively, no detectable Noxa and Puma coimmunoprecipitated with Mcl-1 in U937 cells (Fig. (Fig.8A).8A). Although the concentrations of SBHA that induced expression of Noxa and Puma did not correlate with potentiation of ABT-737 lethality in these cells (Fig. 1D and E), the possibility remained that upregulation of these BH3-only proteins might still contribute to SBHA/ABT-737-induced apoptosis. To test this possibility, U937 and U266 cells were stably transfected with constructs encoding shRNAs targeting Noxa or Puma (Fig. (Fig.8B).8B). As reported previously (55), inhibition of Noxa upregulation by shRNA dramatically reduced the lethality of the proteasome inhibitor bortezomib in U937 cells, manifested by markedly diminished PARP cleavage (Fig. (Fig.8C)8C) and cell death (data not shown). It has also been reported that Puma-deficient cells are resistant to apoptosis induced by proteasome inhibitors (11). Blockade of Puma upregulation by shRNA partially but significantly prevented bortezomib-mediated PARP degradation (Fig. (Fig.8D)8D) and cell death (data not shown) in U937 cells. Notably, while shRNA markedly attenuated SBHA-mediated upregulation of Noxa (Fig. (Fig.8C)8C) and Puma (Fig. (Fig.8D),8D), these approaches, in striking contrast to Bim knockdown, failed to prevent the potentiation of ABT-737 lethality by SBHA. Similar phenomena were observed in U266 cells transfected with shRNA directed against Noxa (Fig. (Fig.8E)8E) or Puma (Fig. (Fig.8F).8F). Together, these results argue against the possibility that SBHA-mediated upregulation of Noxa or Puma plays a significant functional role in interactions with ABT-737 in human leukemia or myeloma cells.
To determine the functional roles of Bcl-2 and Bcl-xL in regulation of Bim function, U937 cells stably transfected with either Bcl-2 or Bcl-xL were employed. As shown in Fig. 9A and B, SBHA induced Bim upregulation in cells overexpressing Bcl-2 or Bcl-xL, as well as in their empty vector counterparts (pCEP or pcDNA3.1, respectively), although basal levels of Bim varied to some extent between these cell lines. In addition, cells ectopically expressing Bcl-2 (Fig. (Fig.9A),9A), Bcl-xL (Fig. (Fig.9B),9B), or Mcl-1 (see Fig. 10A, below) exhibited slightly lower basal levels of Bcl-xL, Mcl-1, or Bcl-2, respectively, possibly representing a compensatory response to altered expression of these antiapoptotic proteins. Nevertheless, levels of each of these antiapoptotic proteins remained essentially unchanged following drug treatment. Significantly, overexpression of both Bcl-2 and Bcl-xL dramatically blocked cell killing mediated by cotreatment with ABT-737 and SBHA (Fig. 9C and D), as documented by substantially diminished PARP cleavage (Fig. 9A and B). Efforts were then undertaken to determine whether this phenomenon might reflect altered associations between Bim and Bcl-2 or Bcl-xL. As shown in Fig. Fig.9E,9E, overexpression of Bcl-2 or Bcl-xL led to increased binding of Bim in untreated cells and to an even greater extent in SBHA-treated cells. Analogous to results in parental U937 cells (Fig. (Fig.3),3), ABT-737 essentially abrogated binding of Bim to Bcl-2 or Bcl-xL in empty vector-transfected cells exposed to SBHA. Notably, Bcl-2 overexpression largely prevented ABT-737 (300 nM) from attenuating Bim/Bcl-2 binding (Fig. (Fig.9E,9E, upper panels). However, Bcl-xL overexpression partially restored Bim/Bcl-xL binding after treatment with ABT-737 in the presence or absence of SBHA (Fig. (Fig.9E,9E, lower panels). Significantly, ectopic expression of Bcl-2 or Bcl-xL both largely diminished conformational changes of Bax and Bak induced by the SBHA/ABT-737 regimen (Fig. (Fig.9F)9F) and strikingly attenuated cell death (Fig. 9C and D). Together, these findings suggest that the protective effects of Bcl-2 overexpression primarily stems from restoration of Bim/Bcl-2 binding in ABT-737/SBHA-treated cells, whereas the antiapoptotic actions of ectopically expressed Bcl-xL may involve other factors in addition to increased sequestration of Bim.
Parallel studies were performed in U937 cells ectopically expressing Mcl-1 (9). Analogous to results involving cells ectopically expressing Bcl-2 or Bcl-xL, both ectopic Mcl-1-overexpressing cells and their empty vector counterparts displayed upregulation of Bim following treatment with SBHA, but no changes were observed in the expression of Bcl-2 or Bcl-xL for any drug treatment (Fig. 10A). Moreover, ectopic Mcl-1 overexpression also largely abrogated PARP cleavage (Fig. 10A) and cell death (Fig. 10B) induced by cotreatment with SBHA and ABT-737. Consistent with these findings, ectopic expression of Mcl-1 prevented conformational changes of both Bax and Bak by this regimen, as determined by both immunoprecipitation (Fig. 10C) and flow cytometry (data not shown).
In striking contrast to results obtained in cells ectopically expressing either Bcl-2 or Bcl-xL, binding of Mcl-1/Bim was negligible after treatment with ABT-737 in the presence or absence of SBHA in both Mcl-1-overexpressing cells and their empty vector counterparts (Fig. 10A, bottom panels). ABT-737 was nevertheless capable of disrupting the association between Bim and endogenous Bcl-xL in U937 cells ectopically expressing Mcl-1 following exposure to this agent alone or in the presence of SBHA (Fig. 10D, upper panel), as observed in parental cells (Fig. (Fig.3).3). Notably, whereas treatment with SBHA resulted in a modest but discernible increase in binding of Bak to Mcl-1, ABT-737 failed to unleash Bak from binding to Mcl-1 (Fig. 10D, lower panels).
Finally, it is possible that attenuation of SBHA/ABT-737 lethality by ectopic expression of Mcl-1 might involve interactions between Mcl-1 and Noxa or Puma. To test this possibility, coimmunoprecipitation assays were performed. Similar to results obtained in parental U937 cells (Fig. (Fig.8A),8A), no detectable Noxa or Puma was coimmunoprecipitated with Mcl-1 in U937 cells ectopically overexpressing Mcl-1 (Fig. 10E). Together, these findings argue that ectopically expressed Mcl-1 potently blocks SBHA/ABT-737-mediated Bax/Bak activation and lethality through a mechanism that does not involve Bim neutralization or interactions with Noxa or Puma. Instead, they suggest that this phenomenon in all likelihood stems from increased sequestration of Bak by Mcl-1. In conjunction with the preceding results, such findings further support the notion that ABT-737-mediated release of Bim from Bcl-2 and Bcl-xL, rather than from Mcl-1, plays a key role in potentiation of SBHA lethality.
To gain further insights into the roles of Bcl-2, Bcl-xL, and Mcl-1 in lethality of the SBHA/ABT-737 regimen, parallel studies were performed employing a considerably higher concentration of ABT-737 (e.g., 10 μM) than used in the previous studies (e.g., 300 to 500 nM). Such high concentrations by themselves killed essentially all parental U937 cells as well as empty vector controls (data not shown). As shown in Fig. 11A, ABT-737 administered alone at this concentration moderately induced cell death in ectopic Bcl-2-overexpressing cells and to a slightly greater extent in Bcl-xL-overexpressing cells, but not in cells ectopically overexpressing Mcl-1. Importantly, ABT-737 lethality was significantly enhanced by coadministration of SBHA in Bcl-2- and Bcl-xL-overexpressing cells. In striking contrast, ectopic Mcl-1 overexpression essentially blocked the lethality of ABT-737 in the presence or absence of SBHA under these conditions (Fig. 11A). Moreover, ABT-737, administered at this high concentration, strikingly diminished both basal and SBHA-induced Bim/Bcl-2 binding in cells ectopically overexpressing Bcl-2 (Fig. 10B), presumably because the higher concentration of ABT-737 was able to neutralize the effects of overexpressed Bcl-2 in a stoichiometric manner.
Similar phenomena were observed in Bcl-xL-overexpressing cells (Fig. 11C). Interestingly, whereas ectopic Bcl-xL overexpression also resulted in a clear increase in the binding of Bak to Bcl-xL, high concentrations of ABT-737 (10 μM) markedly displaced Bak from overexpressed Bcl-xL (Fig. 11C), consistent with previous results demonstrating that ABT-737 disrupts the Bcl-xL/Bak association (9, 32). Such findings raise the possibility that ectopic overexpression of Bcl-xL opposes cell death by binding to and neutralizing both Bim and Bak and that the latter events are also reversible by increasing ABT-737 concentrations. They may also explain the discordance between the partial disassociation of Bcl-xL/Bim by ABT-737 (Fig. (Fig.9E)9E) and the virtually complete blockade of Bak activation (Fig. (Fig.9F)9F) in Bcl-xL-overexpressing cells coexposed to SBHA and lower concentrations (e.g., 500 nM) of ABT-737.
Finally, in striking contrast to these findings, a high concentration (10 μM) of ABT-737 failed to block binding of Mcl-1 to Bim in U937 cells ectopically overexpressing Mcl-1; in fact, Bim/Mcl-1 binding was if anything slightly increased (Fig. 11D). Notably, ectopic overexpression of Mcl-1 resulted in a clear increase in binding of Mcl-1 to Bak, which was not affected by ABT-737, presumably because this agent does not target Mcl-1 (9). Consistent with these results, the high concentration (10 μM) of ABT-737 induced activation and Bak and Bax by itself, and this event was significantly enhanced by coadministration of SBHA in cells overexpressing Bcl-2 or Bcl-xL, but not in those ectopically overexpressing Mcl-1 (data not shown). Together, these findings are consistent with the notion that ectopic overexpression of these antiapoptotic proteins acts to prevent cell death induced by the SBHA/ABT-737 regimen via neutralization of Bim (by Bcl-2), neutralization of both Bim and Bak (by Bcl-xL), or neutralization of Bak (by Mcl-1), respectively. They also argue that interference with only the first two of these events is involved in the interaction between SBHA and ABT-737.
Cell fate is determined by the balance between prosurvival and prodeath signals, which are controlled precisely by a complex network of proteins. The Bcl-2 family represents a critical group of molecules involved directly in the regulation of cell death. In contrast to normal cells, overexpression of anti-apoptotic Bcl-2 family proteins (e.g., Bcl-2, Bcl-xL, or Mcl-1) is characteristic of the malignant cell phenotype (12, 44, 59). This phenomenon contributes to both loss of growth control due to disabling of the cell death arm and to resistance to diverse anticancer agents (46, 57). However, transformed cells also express proapoptotic protein members of the Bcl-2 family, including multidomain proteins (e.g., Bax and Bak) and some, if not all, BH3-only proteins, including Bim (47). These proapoptotic proteins are rendered inactive through associations with their highly expressed antiapoptotic counterparts (13, 47). Such findings prompted the development of anticancer strategies designed to recapitulate the physiologic death process, i.e., induction/upregulation of proapoptotic Bcl-2 family proteins or downregulation/antagonism of antiapoptotic proteins. In the case of the former, attempts have focused on BH3-only proteins, because levels of the multidomain proteins Bax and Bak are relatively stable (1) and they must be either activated directly by BH3-only “activators” or unleashed from their antiapoptotic counterparts by BH3-only proteins to induce MOMP (27). In this context, several novel agents have been found to induce the expression or prevent the degradation of BH3-only proteins, such as Noxa, Puma, and particularly Bim (1). Specifically, HDAC inhibitors have been shown to induce Bim expression in transformed cells through an E2F-dependent process, an event functionally related to lethality (73). In the second case, strategies include downregulation of short-lived antiapoptotic proteins by agents that inhibit transcription (e.g., Mcl-1 downregulation by certain cyclin-dependent kinase inhibitors) (56) or by antisense oligonucleotides (e.g., Bcl-2 repression by oblimersen/G3139) (59). An alternative approach has been to develop specific inhibitors of antiapoptotic Bcl-2 family proteins. For example, the compound ABT-737 was developed through high-throughput nuclear magnetic resonance screening/structure-based design and mimics BH3-only proteins (e.g., Bad) by binding avidly to the hydrophobic BH3-binding groove of Bcl-xL, Bcl-2, and Bcl-w, but not Mcl-1 and A1 (53). An analog (ABT-263) of ABT-737 is currently in clinical development (64). Several previous studies have shown that the sensitivity of tumor cells to ABT-737 is inversely related to Mcl-1 expression and that interventions that diminish Mcl-1 expression potentiate ABT-737 lethality (37, 49, 62, 65), most likely by freeing Bak from both Bcl-xL and Mcl-1 (9). In addition, it has been shown that ABT-737 releases Bim from binding to Bcl-2, thus inducing tumor cell death or sensitization to other agents (18, 32). Such considerations raised the possibility that the two strategies (i.e., simultaneous upregulation of Bim and prevention of Bim/Bcl-2 binding) might cooperate to trigger apoptosis. Indeed, the present results indicate that the HDAC inhibitor SBHA interacts synergistically with ABT-737 to induce cell death in various malignant hematopoietic cells and that Bim upregulation by SBHA plays an essential role in this phenomenon. Collectively, these findings suggest that, in addition to approaches involving Mcl-1 downregulation, an alternative strategy to potentiate cell death induced by Bcl-2 antagonists (e.g., ABT-737) would be to upregulate Bim (e.g., by HDAC inhibitors or potentially other anticancer agents that upregulate Bim).
Bim is universally expressed in human cancer cells but is sequestered by antiapoptotic proteins, thus preventing it from activating Bax and Bak (10, 36, 45). Unlike BH3-only “sensitizers” that selectively bind to certain antiapoptotic proteins, Bim binds stoichiometrically in a 1:1 ratio to all known antiapoptotic Bcl-2 family proteins, and particularly to Bcl-2, Bcl-xL, and Mcl-1, with high affinities (Kd < 10 nM for all three in in vitro assays) (6, 8). However, the specific role of each of these antiapoptotic proteins in neutralizing Bim function may vary depending upon their basal expression levels. For example, chronic lymphocytic leukemia cells exhibit high levels of Bcl-2 and Bim but low levels of Mcl-1. Consequently, these cells are primed for the lethality of ABT-737, which targets Bcl-2 but not Mcl-1 (6). Indeed, Bim is largely sequestered by Bcl-2 in chronic lymphocytic leukemia cells, and once displaced by ABT-737, results in Bax activation and MOMP (18). Similar phenomena have been described in acute lymphoblastic leukemia cells, although in some cases, these cells exhibit comparable levels of Mcl-1 and Bcl-2 (19). Notably, while exposure of U937 cells to SBHA (e.g., 15 to 30 μM) resulted in a marked increase in Bim expression, only a modest increase in cell death was observed at these concentrations. Coimmunoprecipitation analysis indicated that SBHA treatment also markedly increased the amount of Bim bound to both Bcl-2 and Bcl-xL but had little effect on Bim/Mcl-1 binding. These findings suggest that in SBHA-treated cells, upregulated Bim is primarily sequestered by Bcl-2 and Bcl-xL and is therefore prevented from inducing Bax and Bak activation. The finding that marginally toxic concentrations of SBHA substantially increased Bim expression argue that Bim induction (e.g., by HDAC inhibitors) is not lethal per se; instead, it must be released from its inhibitory associations with antiapoptotic proteins such as Bcl-2 and Bcl-xL (e.g., by ABT-737) for full engagement of the death signaling cascade (i.e., Bax/Bak activation, MOMP, and caspase activation). An alternative view is that SBHA treatment, by increasing Bim expression, primes cells for killing by agents such as ABT-737 that disrupt Bcl-2/Bcl-xL function.
Because ABT-737 does not target Mcl-1 (53), cells expressing high levels of this protein are relatively resistant to ABT-737 lethality (9, 37, 62), a phenomenon that can be overcome by interventions that diminish Mcl-1 expression (9, 37, 49). It is noteworthy that interactions between SBHA and ABT-737 were observed in various human leukemia and myeloma cell types exhibiting disparate basal levels of Mcl-1. Such findings, along with evidence that SBHA did not increase the amount of Bim bound to Mcl-1, suggest that the enhanced lethality of the SBHA/ABT-737 regimen stems from factors in addition to or unrelated to Mcl-1. On the other hand, SBHA-mediated potentiation of ABT-737 lethality was very closely associated with Bim upregulation in diverse cell types, including established human leukemia and myeloma cell lines, as well as primary AML blasts. Notably, abrogation of SBHA-induced Bim upregulation by an shRNA approach significantly attenuated the lethality of the SBHA/ABT-737 regimen, arguing strongly that Bim upregulation plays a critical functional role in interactions between these two agents. Exposure to SBHA also resulted in upregulation of other BH3-only proteins, including Noxa and Puma, both of which have been implicated in cellular responses to physiologic death signals as well as drug treatment (66). While induction of both Puma and Noxa are generally considered to be p53-dependent events (66), p53-independent mechanisms of PUMA and Noxa upregulation have also been described (25, 34). The finding that SBHA induced upregulation of Puma and Noxa in p53-null U937 cells suggests that HDAC inhibitors induce expression of these BH3-only proteins via a p53-independent mechanism. Puma has been shown to function as a BH3-only “activator” (i.e., directly activating Bax) with respect to the entire protein but not as an isolated BH3 domain (5, 36). In contrast, Noxa is a pure “sensitizer” BH3-only protein which selectively binds to Mcl-1, displacing Bak from Mcl-1, leading to ubiquitination/proteasomal degradation of Mcl-1 (68). However, it is noteworthy that in U937 cells, both Puma and Noxa were induced by lower SBHA concentrations (e.g., 5 to 10 μM) than those required for Bim induction (i.e., >15 μM). Significantly, these lower SBHA concentrations failed to enhance ABT-737 lethality, despite inducing significant increases in Puma and Noxa expression, whereas only higher SBHA concentrations capable of upregulating Bim markedly potentiated ABT-737-mediated apoptosis. Such findings argue, albeit indirectly, that SBHA-mediated upregulation of Bim, rather than Noxa or Puma, is primarily responsible for enhancing ABT-737-induced cell death. Furthermore, shRNA knockdown of Puma and Noxa, in marked contrast to knockdown of Bim, failed to attenuate SBHA-mediated potentiation of ABT-737 lethality. Lastly, although exposure to SBHA did not affect expression of other BH3-only proteins (e.g., Bid, Bad, Bik, Bmf, and Hrk), the possibility that total levels of these proapoptotic proteins may have an impact on cell death induced by the SBHA/ABT-737 regimen cannot be excluded.
ABT-737 targeted Bcl-2 and Bcl-xL by disrupting their association with Bim, either in the absence or presence of SBHA, in virtually all cell types employed in the present study. In contrast to these cell-type-independent events, exposure of different cells to ABT-737 alone resulted in divergent, albeit modest, effects on total Bim levels or the amount of Bim bound to Mcl-1. For example, exposure of HL-60 cells to subtoxic concentrations of ABT-737 alone resulted in a modest but discernible increase in the amount of Bim bound to Mcl-1 (Fig. (Fig.4A,4A, bottom panel) but did not clearly affect levels of Bim protein (Fig. (Fig.2C,2C, lower panels). One possible explanation for this phenomenon could be that ABT-737, which only targets Bcl-2 and Bcl-xL, but not Mcl-1, thereby releases Bim from complexes with Bcl-2 and Bcl-xL (Fig. (Fig.4A,4A, upper panels). This could lead in turn to increased availability of free Bim for Mcl-1 binding in such cell types. However, other explanations cannot be excluded, including the possibility that drug treatment could directly affect the binding capacity between Bim and different antiapoptotic proteins (e.g., Bcl-2, Bcl-xL, and Mcl-1) that exhibit significant differences in the structural properties responsible for binding of BH3-only proteins including Bim (42). On the other hand, treatment with ABT-737 alone led to a slight decrease in amount of Bim coimmunoprecipitated by Mcl-1 in U937 cells (Fig. (Fig.3D,3D, bottom panel), which may reflect the modest reduction in total Bim levels with this treatment, particularly at 500 nM ABT-737 (Fig. (Fig.1G).1G). In this context, it has recently been reported that binding to certain proteins (e.g., Hsp70 or pVHL ) increases the stability of Bim protein through prevention of ubiquitination and proteasomal degradation. It is therefore possible that ABT-737 frees Bim from binding to Bcl-2 and Bcl-xL and in so doing diminishes its stability. Finally, ABT-737 treatment did not affect total levels of Bim protein (Fig. (Fig.2C,2C, upper panels) or the amount of Bim bound to Mcl-1 in Jurkat cells (Fig. (Fig.4A).4A). Further studies will be needed to define the basis for these cell-type-specific phenomena.
The observation that release of Bim from Bcl-2/Bcl-xL by ABT-737 in SBHA-treated cells induced a pronounced conformational change in Bak and Bax, as well as Bax translocation, and that these events were largely prevented by Bim shRNA, suggests that free Bim may act directly to activate multidomain proapoptotic proteins. While Bim as well as Bid have been classified as “activator” BH3-only proteins which directly activate Bax/Bak (40, 45), this view has been called into question by recent findings suggesting that (i) Bim does not physically interact with Bax (e.g., a Bim BH3 domain peptide failed to bind to Bax in vitro, and the Bim/Bax complex is difficult to detect in cells), and (ii) Bax can engage the apoptotic program in cells lacking Bim or Bid (69). It has therefore been proposed that Bim acts by binding to antiapoptotic proteins, neutralizing their constraining effect on Bax/Bak (1). However, very recent reports indicate that a modified Bim peptide which recapitulates the natural configuration of the Bim protein does bind tightly to Bax in vitro (26). In addition, Bax induction in the absence of Bim and Bid could reflect the presence of other, yet-to-be-identified “activators” (47). It is noteworthy that ABT-737 alone exhibited only modest lethality at concentrations (e.g., 300 to 500 nM) that diminished basal binding of Bim to Bcl-2/Bcl-xL. In this context, SBHA-mediated priming of cells (i.e., by upregulation of Bim) may be required for ABT-737 to induce Bak activation and Bax translocation, which together initiate MOMP and caspase activation. Interestingly, whereas ectopic overexpression of Mcl-1 prevented SBHA/ABT-737 lethality primarily by sequestering Bak, it is noteworthy that Mcl-1 overexpression also diminished Bax conformational change/activation. One possible explanation for this phenomenon involves cooperativity between Bak and Bax activation, as previous studies have suggested (9). This notion is supported by a recent report demonstrating that knockdown of Bak (e.g., by siRNA) abolishes Bax activation by cisplatin and that the failure of cisplatin to activate Bax can be reversed by ABT-737 in cells that have been depleted of the voltage-dependent anion channel 1, which acts downstream of Bak but upstream of Bax (63). Results obtained in Bax or Bak knockout MEFs demonstrating that the presence of both Bax and Bak is required for SBHA/ABT-737-mediated cell killing are consistent with such findings. An alternative possibility is that Mcl-1 may interfere with other yet-to-be-identified “activators” that can directly activate Bax (47). Nevertheless, the observation that upregulation of Bim (e.g., by SBHA) cooperates with its release from Bcl-xL/Bcl-2 (e.g., by ABT-737) to promote a pronounced increase in Bak activation, Bax conformational change, and Bax translocation, is compatible with the direct activation model of Bim action.
In SBHA-treated U937 cells, inducible Bim was largely sequestered by Bcl-2 and Bcl-xL, rather than Mcl-1, suggesting that these antiapoptotic proteins may play disparate roles in interactions between SBHA and ABT-737. It is noteworthy that in other cell types (e.g., fibroblasts and epithelial cells), newly expressed BimEL associates with both Bcl-xL and Mcl-1 following serum withdrawal (23), suggesting that mechanisms regulating Bim may vary between different cell types and/or death stimuli. In this context, selectivity in the binding of BH3-only “sensitizers” to specific multidomain proteins has been described. For example, Bad binds to both Bcl-2 and Bcl-xL, whereas Noxa primarily binds to Mcl-1 (8). In addition, Bak is sequestered by both Mcl-1 and Bcl-xL, but not by Bcl-2 (68), while Bax binds to Bcl-2, Bcl-xL, Bcl-W, and Bcl-B (70). While all of these antiapoptotic proteins have been shown to bind to Bim (6, 8), the present results suggest that they may act differentially with respect to Bim neutralization. This notion is supported by the disparate responses of cells ectopically expressing these proteins to regimens combining SBHA with low versus high concentrations of ABT-737. First, ectopic expression of either Bcl-2, Bcl-xL, or Mcl-1 all conferred marked resistance to cell death induced by SBHA in the presence of low concentrations (e.g., 500 nM) of ABT-737, a phenomenon associated with abrogation of Bax and Bak activation. On the other hand, ectopic Bcl-2 overexpression markedly increased Bim/Bcl-2 binding in untreated cells as well as in those exposed to SBHA. However, low concentrations of ABT-737 failed to abolish Bim/Bcl-2 binding, presumably because the abundance of Bcl-2 exceeded the capacity of this concentration of ABT-737, an agent that binds to Bcl-2 stoichiometrically (41, 53), to unleash Bim. The finding that Bim/Bcl-2 binding was largely reversed by increasing ABT-737 concentrations (e.g., to 10 μM) supports this concept. As in the case of Bcl-2, ectopic Bcl-xL overexpression also resulted in an increase in Bim/Bcl-xL binding in both untreated and SBHA-treated cells. However, unlike results obtained in cells overexpressing Bcl-2, low concentrations of ABT-737 partially but significantly diminished Bim/Bcl-xL binding in cells overexpressing Bcl-xL. This phenomenon most likely reflects the higher inhibitory potency of ABT-737 toward Bcl-xL versus Bcl-2 (i.e., in vitro median inhibitory concentrations of 35 to 64 nM and 103 to 120 nM for Bcl-XL and Bcl-2, respectively [53, 71]). On the other hand, the discordance between the virtual abrogation of Bax/Bak activation despite only partial disruption of Bim/Bcl-xL binding in cells coexposed to SBHA and ABT-737 suggests the involvement of an alternative mechanism of Bcl-xL antiapoptotic actions, e.g., direct binding to and neutralization of Bak (9, 61, 68). Indeed, ectopic Bcl-xL overexpression resulted in a marked increase in Bak/Bcl-xL binding. Significantly, a high concentration of ABT-737 not only dramatically diminished Bim/Bcl-xL binding but also markedly disrupted the association between Bak and Bcl-xL in Bcl-xL-overexpressing cells, accompanied by a pronounced increase in Bak/Bax activation and cell death. Finally, in striking contrast, ectopic Mcl-1 overexpression did not increase binding of Bim to Mcl-1, but instead substantially increased Bak/Mcl-1 binding. Notably, the latter phenomenon could not be reversed by increasing ABT-737 concentrations, presumably due to the low binding affinity of ABT-737 for Mcl-1 (53), thereby accounting for the failure of the regimen to trigger Bak/Bax activation and cell death in ectopic Mcl-1-overexpressing cells. Taken together, these findings argue strongly that ABT-737-mediated release of Bim from Bcl-2 and Bcl-xL, as well as Bak from Bcl-xL, but not from Mcl-1, exert critical roles in interaction between SBHA and ABT-737.
A model summarizing the current findings is presented in Fig. 11E. According to this model, HDAC inhibitors induce upregulation of proapoptotic BH3-only proteins (particularly Bim), while antiapoptotic molecules such as Bcl-2 and Bcl-xL act to neutralize Bim, and in so doing, prevent activation of Bax and Bak. In this model system, Mcl-1, in contrast, primarily acts by sequestering/inactivating Bak, rather than Bim. Induction of Bim by HDAC inhibitors in conjunction with Bim displacement from Bcl-2 and Bcl-xL (i.e., by ABT-737) cooperatively activates both Bak and Bax to initiate the cell death process. Finally, the present findings also suggest that the protective effect of Bcl-2, Bcl-xL, or Mcl-1, particularly in the case of cells expressing high levels of these proteins, may stem from different mechanisms, i.e., sequestration/neutralization of Bim (in the case of Bcl-2), both Bim and Bak (in the case of Bcl-xL), or primarily Bak (in the case of Mcl-1).
This work was supported by grants CA63753, CA93738, and CA100866 from the National Institutes of Health; award 61 81-10 from the Leukemia and Lymphoma Society of America; an award from the Multiple Myeloma Research Foundation; an award from the V Foundation; Lymphoma SPORE award 1P50 CA130805; and RC2CA148431-01 from the National Cancer Institute.
Published ahead of print on 5 October 2009.