Src inhibition potentiates MCL-1 antagonist activity in acute myeloid leukemia

Abstract

The importance of MCL-1 in leukemogenesis has prompted development of MCL-1 antagonists e.g., S63845, MIK665. However, their effectiveness in acute myeloid leukemia (AML) is limited by compensatory MCL-1 accumulation via the ubiquitin proteasome system. Here, we investigated mechanisms by which kinase inhibitors with Src inhibitory activity e.g., bosutinib (SKI-606) might circumvent this phenomenon. MCL-1 antagonist/SKI-606 co-administration synergistically induced apoptosis in diverse AML cell lines. Consistently, Src or MCL-1 knockdown with shRNA markedly sensitized cells to MCL-1 inhibitors or SKI-606 respectively, while ectopic MCL-1 expression significantly diminished apoptosis. Mechanistically, MCL-1 antagonist exposure induced MCL-1 up-regulation, an event blocked by Src inhibitors or Src shRNA knock-down. MCL-1 down-regulation was associated with diminished transcription and increased K48-linked degradative ubiquitination. Enhanced cell death depended functionally upon down-regulation of phosphorylated STAT3 (Tyr705/Ser727) and cytoprotective downstream targets c-Myc and BCL-xL, as well as BAX/BAK activation, and NOXA induction. Importantly, the Src/MCL-1 inhibitor regimen robustly killed primary AML cells, including primitive progenitors, but spared normal hematopoietic CD34⁺ cells and human cardiomyocytes. Notably, the regimen significantly improved survival in an MV4-11 cell xenograft model, while reducing tumor burden in two patient-derived xenograft (PDX) AML models and increased survival in a third. These findings argue that Src inhibitors such as SKI-606 potentiate MCL-1 antagonist anti-leukemic activity in vitro and in vivo by blocking MCL-1 antagonist-mediated cytoprotective MCL-1 accumulation by promoting degradative ubiquitination, disrupting STAT-3-mediated transcription, and inducing NOXA-mediated MCL-1 degradation. They also suggest that this strategy may improve MCL-1 antagonist efficacy in AML and potentially other malignancies.

Fig. 1

MCL-1 inhibitors interact synergistically with SKI-606 to induce apoptosis in AML cells. a U937 cells were exposed to the indicated concentrations of S63845 ± SKI-606 for 24 h, followed by flow cytometric analysis of cell death after staining with 7-AAD (n = 13 in each group). Values represent the mean % ± standard deviations (SD). b Cells were incubated with S63845 ± SKI-606 for 24 h, after which PARP, cleaved-Caspase-3 and γH2A.X were monitored by immunoblotting analysis. β-actin was used as a loading control to ensure equal protein loading and transfer. c Cells were exposed (24 h) to S63845 ± SKI-606, followed by Annexin V and DAPI staining using fluorescence microscopy. Scale bar = 50 µm. d Cells were exposed (24 h) to varying concentrations of S63845 ± SKI-606 at a fixed ratio (1:100), after which the percentage of 7-AAD⁺ cells was determined. Median dose-effect analysis was then employed to characterize the nature of the interaction between these agents. Combination index values < 1.0 denote a synergistic interaction. e–h U937 cells were exposed to the indicated concentrations of MIK665 ± SKI-606 for 24 h. Assays were performed as in (a–d). e n = 7 in each group. i Synergistic interaction between S63845 or MIK665 and SKI-606 were tested in two additional AML cell lines. CF, cleavage fragment; DAPI, 4′,6-diamidino-2-phenylindole. ****P < 0.0001

Fig. 2

MCL-1 down-regulation and activation of BAX and BAK are required for S63845/SKI-606-mediated apoptosis. a U937 cells were exposed to the indicated concentrations of S63845 or MIK665 ± SKI-606 for 24 h, after which the level of MCL-1 was determined by western blot analysis. Numerals under the blots correspond to densitometric readings normalized to untreated controls (1.0). b, c Ectopic expression of MCL-1 in U937 cells. b Cells were treated (24 h) with 20 nM S63845 ± 2 µM SKI-606 and cell death was determined by 7-AAD staining and flow cytometric analysis (n = 4 in each group). Inset, levels of MCL-1 by western blot after overexpression. c Western blot analysis of PARP, cleaved- PARP, as well as cleaved-Caspase-3 was performed. β-actin was assayed to ensure equivalent loading and transfer. d U937 cells were exposed to S63845 (20 nM) and SKI-606 (2 μM) alone or in combination for 24 h, after which subcellular fractions were obtained and subjected to western blot analysis to monitor the release of cytochrome c, BAK, and BAX into the cytosol. S-100, cytosol; Cyto c = cytochrome c. Numerals under the blots correspond to densitometric readings normalized to untreated controls (1.0). e U937 cells were exposed to S63845 ± SKI-606 for 24 h, after which cells were lysed in buffer containing 1% CHAPS; conformationally changed BAX and BAK proteins were immunoprecipitated using anti-BAX 6A7 and anti-BAK Ab1 antibodies, respectively, and subjected to western blot analysis using polyclonal BAX or BAK antibodies. f Following 24 h treatment, U937 cells were lysed in buffer and immunoprecipitated (IP) using anti-BAK antibody, followed by western blot analysis using anti-BAK, anti-BAX, or anti-MCL-1 antibodies as indicated. For all IP assay, IPs without cell lysate (-lysate) and/or with IgG (instead of primary antibodies) were carried out as controls; Input lysates were also subjected to western blot analysis to monitor relative protein levels. IgG levels are shown to ensure equal loading of IP antibodies. g BAK and/or BAX CRISPR knockout U937 cells were incubated with S63845 ± SKI-606 for 24 h. Cell death was determined using 7-AAD staining and flow cytometry. Values represent the mean % ± SD for three separate experiments. h BAK, BAX, PARP, cleaved-PARP and cleaved-Caspase-3 were detected by western blot. β-actin was assayed to ensure equivalent loading and transfer. CF, cleavage fragment; EV, empty vector; OE, overexpression. ****P < 0.0001; ^####P < 0.0001

Fig. 3

The S63845/SKI-606 regimen blocks phosphorylation of STAT3 (p-Y705/ p-S727) and diminishes MCL-1 transcription by reducing p-STAT3 (Y705) nuclear translocation. a Western blot analysis of p-STAT3 (Y705), p-STAT3 (S727), MCL-1, c-MYC, BCL-_XL in U937 cells treated with 20 nM S63845 ± 2 µM SKI-606 for 16 h. b Cells were stained with p-STAT3 (Y705), p-STAT3 (Y727) and DAPI, and then visualized by fluorescence microscopy and by ImageStream analysis. Representative cells are shown (BF = brightfield). Histograms of p-STAT3 (Y705) and p-STAT3 (Y727) intensity and fold-change are shown. Representative data for at least three replicates. c U937 cells were treated with S63845 ± SKI-606 for 16 h. Nuclear and cytoplasmic extracts were collected for performance of p-STAT3 (Y705) immunoblotting assays. P84 and β-actin were used as loading controls for nuclear and cytoplasmic protein, respectively. d Relative mRNA expressions of MCL-1 was determined by real-time RT-PCR analysis (n = 3 in each group). GAPDH served as an internal control. e–i U937 cells were infected with a lentivirus harboring constitutively-active STAT3 (FLAG fusion), and STAT3-CA clone #1 and #2 were selected for the following experiments. e Western blot analysis was performed to test Flag and p-STAT3 (Y705) levels in STAT3-CA cells. f STAT3 DNA-binding ELISA was used to evaluate STAT3 activity in STAT3-CA cell (n = 3 in each group). g Cells were exposed (24 h) to indicated concentrations of S63845 ± SKI-606, followed by CellTiter-Glo® Luminescent Cell Viability Assay to monitor cell viability. Values represent the mean % ± SD for four separate experiments performed in triplicate. h Western blot analysis of FLAG, cleaved-PARP and cleaved-Caspase-3, as well as γH2A.X was performed. β-actin were assayed to ensure equivalent loading and transfer. i Western blot analysis of FLAG, p-STAT3 (Y705), MCL-1, C-MYC, as well as BCL-xL was performed. β-actin were assayed to ensure equivalent loading and transfer. **P < 0.01, ***P < 0.001, ****P < 0.0001; ^####P < 0.0001

Fig. 4

Co-treatment with S63845 and SKI-606 promotes the NOXA and ubiquitination-related degradation of MCL-1. a Effect of MG132 on MCL-1 expression in U937 cells treated with S63845 ± SKI-606. The cells were treated with 20 nM S63845 ± 2 µM SKI-606 for 16 h and then incubated with 10 µM MG132 for 2 h. b U937 cells were exposed to the indicated concentrations of S63845 ± SKI-606 for 24 h, after which cells were lysed in buffer and immunoprecipitated (IP) using anti-MCL-1 antibody, followed by western blot analysis using anti-Ubiquitin K48 antibody as indicated; Input lysates were also subjected to western blot analysis to monitor relative protein levels. IgG levels are shown to ensure equal loading of IP antibodies. c Western blot analysis showing expression of NOXA in U937 cells treated with S63845 or MIK665 ± SKI-606. d–g U937 cells were infected with a lentivirus harboring NOXA shRNA or EV, clones #1 and #2 were selected for the following experiments. d Cells were exposed to 20 nM S63845 ± 2 μM SKI-606 for 24 h, after which cell viability was determined by CellTiter-Glo® Luminescent Cell Viability Assay. Values represent the mean % ± SD for four separate experiments performed in triplicate. Inset: expression of NOXA by WB after infection with a lentivirus harboring EV or shNOXA. e Western blot analysis of PARP, cleaved-PARP, cleaved-Caspase-3, and γH2A.X was performed. β-actin controls were assayed to ensure equivalent loading and transfer. f Effect of NOXA depletion on MCL-1 expression. g Ubiquitination levels of MCL-1 in U937 NOXA knock-down cells treated with S63/SKI. IP with IgG (instead of primary antibodies) were carried out as controls. EV, empty vector, CF, cleavage fragment. ****P < 0.0001; ^####P < 0.0001

Fig. 5

The MCL-1 inhibitor/SKI-606 regimen kills primary AML blasts but not normal CD34⁺ cells, blocks STAT3 phosphorylation (p-Y705), and down-regulates MCL-1. a Representative primary bone marrow cells from patients with AML were exposed to 50 nM S63845 or MIK665 ± 2 µM SKI-606 for 24 h, after which the cells were stained with PE-CD34, FITC-Annexin V and DAPI. Scale bar = 20 μm. b Primary AML patient samples were exposed (16–20 h) to varying concentrations of S63845 or MIK665 ± SKI-606 at a fixed ratio (1:40), after which the percentage of apoptotic (annexin V⁺) cells was determined. Median dose-effect analysis was then employed to characterize the nature of the interaction between these agents. Combination index values < 1.0 denote a synergistic interaction. c Assessment of cell viability in primary CD34⁺ AML cells following a 16–20 h exposure to 50 nM S63845 or MIK665 ± 2 µM SKI-606 by multi-color flow cytometric determination of Annexin V-FITC/7-AAD uptake. Lines indicate mean and SD. d Parallel experiments were conducted in primary AML specimens exhibiting a more primitive phenotype (CD34⁺CD123⁺CD38^-). e Analogous experiments were carried out with primary cord blood (CB) CD34⁺ samples. f Western blot analysis of MCL-1, p- STAT3 (Y705) and cleavage of Caspase-3 in representative primary patient samples following exposure to S63845 or MIK665 50 nM ± SKI-606 2 µM for 16–20 h. β-actin controls were assayed to ensure equivalent loading and transfer. ****P < 0.0001; ns, not significant

Fig. 6

The S63845/SKI-606 regimen exhibits significant in vivo activity. a–e NOD/SCID-γ NSG mice (5 mice/group) were inoculated in the right flank with 1 × 10⁶ U937 cells. Treatment was initiated after 8 days. S63845 (25 mg/kg, twice a week, I.P.) ± SKI-606 (150 mg/kg, 5 days weekly, p.o.) were administrated weekly for 2 weeks. a Tumor size was monitored every other day. Mean tumor volume was calculated using the formula (1/2 × [length × width²]). b, c At day 21, tumors were harvested and weighed. d Mouse body weights were monitored every other day throughout the treatment period. P > 0.05. e Immunoblotting analysis was then performed to monitor expression of PARP, cleaved caspase-3 and γH2A.X. β-actin was assayed to ensure equivalent loading and transfer. f–i Mice (10 mice/group) were inoculated via tail vein with 5 × 10⁶ MV4-11 cells stably expressing luciferase. After signals were visible (e.g, 13 days after injection of tumor cells), S63845 (15 mg/kg, twice a week for two weeks, then adjusted to one time weekly for three weeks, I.P.) ± SKI-606 (150 mg/kg, 5 days weekly, p.o.) were administrated for 5 weeks. Control animals were administered equal volumes of vehicle. f Tumor burden was monitored every week after subcutaneous (sub-Q) injection with 150 mg/kg luciferin using the IVIS 200 imaging system. g Quantification of the luminescent signal. Data represents the means ± SD performed on all mice for each group. h Kaplan–Meier survival plot (P = 0.0452 and 0.0038 for the combination vs S63845 or SKI-606 alone, respectively, by log-rank test). i Mouse body weights during treatment (P > 0.05). CF, cleavage fragment. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 7

Combined MCL-1 inhibitor/SKI-606 exposure inhibits leukemic cell expansion in an AML Patient-Derived Xenograft (PDX) model. a–h NOD/SCID-gamma SCF/GM-CSF/IL3 (NSG-SGM3) mice (4 mice/group) were inoculated via tail vein with 0.5 × 10⁶ primary patient-derived AML cells. Treatment was initiated after 6 days. S63845 (20 mg/kg, twice a week, I.P.) ± SKI-606 (150 mg/kg, 5 days weekly, p.o.) were administrated each week for 5 weeks. Control animals were administered equal volumes of vehicle. Peripheral blood human CD45⁺ (hCD45⁺) cells were monitored every two weeks. a hCD45⁺ cells in the peripheral blood were quantified by flow cytometry. Data are shown as the mean % ± SD. b, c After 5-weeks of treatment, mice were sacrificed and hCD45⁺ cells in the bone marrow and spleen were quantified by flow cytometry. d The percentage of hCD45⁺ cells in the peripheral blood, bone marrow and spleen in different groups is illustrated in the histogram. e IHC of bone marrow (femur) and spleen, stained with monoclonal antibodies for hCD45 in experimental mice. Scale bar = 50 μm. f The percentage of cell death in hCD45⁺ cell population in different groups is reflected in the dot plots. g Cell death was assessed selectively in the hCD45⁺ cell population by Annexin V/7-AAD staining. h Body weights during treatment were monitored twice/week (P > 0.05). i Kaplan–Meier survival plot. Mice (10 mice/group) were inoculated with 5 × 10⁶ primary blasts #03 via tail vein. *P < 0.05, **P < 0.01

Fig. 8

Schematic model of MCL-1/SRC inhibitor interactions in AML cells. MCL-1 antagonists, through a yet-to-be-determined mechanism, disrupt degradative MCL-1 ubiquitination, leading to MCL-1 stabilization, accumulation, and prevention of apoptosis. This phenomenon is antagonized by Src inhibitors, resulting in down-regulation of this anti-apoptotic protein (1). In addition, Src inhibitors counteract the activating phosphorylation of the STAT3 trasnscription factor, thereby inhibiting the transcription of multiple genes implicated in AML cell survival and proliferation e.g., BCL-xL, c-MYC, and MCL-1 (2). Moreover, combining Src and MCL-1 inhibitors leads to up-regulation of NOXA, a well-established promoter of MCL-1 degradation (3). Collectively, these events lead to dissociation of MCL-1 from BAX and BAK, thereby activating these proteins and triggering mitochondrial injury and apoptosis (4)