MI-238 is a novel and potent Mcl-1 inhibitor

As an important pro-survival protein, Mcl-1 is over-expressed in various types of cancer. However, no specific Mcl-1 inhibitor is currently available for clinical use. Through cytotoxicity screening using the paired H1299 parental and Mcl-1 knockout (KO) cells, we discovered a small molecule that selectively inhibited the viability of H1299 parental cells, but not Mcl-1 KO cells (Additional file 1: Fig. S1), and we named this compound MI-238 (Fig. 1A). MI-238 was docked into Mcl-1 BH3 binding groove and proximal to the BH1 domain of Mcl-1 (Fig. 1A). To directly measure MI-238/Mcl-1 binding, we prepared GST tagged Mcl-1 recombinant protein (Fig. 1B). MI-238 exhibited a Ki of 0.45 ± 0.05 μM to human Mcl-1 protein in FP (fluorescence polarization) assay (Fig. 1C). Mcl-1 exerts its pro-survival function through binding with BH3-only pro-apoptotic proteins, such as Bak. To test whether MI-238 binding to Mcl-1 could disrupt Mcl-1/Bak association, purified Mcl-1 and Bak proteins were incubated in presence of increasing concentrations of MI-238. As shown in Fig. 1D, MI-238 disrupted Bak/Mcl-1 interaction in a dose-dependent manner in vitro. Collectively, these results demonstrated that MI-238 directly binds to Mcl-1 and inhibits Mcl-1 anti-apoptotic function.

Fig. 1

MI-238 targets Mcl-1 and disrupts Mcl-1/Bak association. A Chemical information and structure modeling of MI-238 in the BH3-binding pocket of Mcl-1 (PDB ID: 4HW3). B Purification of recombinant GST or GST-Mcl-1 proteins. C Fluorescence polarization (FP) assay was performed to measure the binding of MI-238 to recombinant GST-Mcl-1 protein. Data are represented as mean ± SD, n = 3. D GST pull down assay was conducted to analyze the association between recombinant GST-Mcl-1 (rGST-Mcl-1) and recombinant Bak (rBak) in presence of increasing concentrations of MI-238

MI-238 selectively induces apoptosis in Mcl-1 proficient cells

To measure whether MI-238 induced apoptosis depends on Mcl-1, we generated H1299 Mcl-1 KO cells (Fig. 2A, C) (Chen et al. 2018). Knockout of Mcl-1 in H1299 and MEF cells did not significantly affect the expression Bcl-2 protein. However, BH3-only proteins including Bak, Bim and Bax were shifted to Bcl-2 and Bcl-xL in Mcl-1 deficient cells, which indicated Mcl-1 KO cells relied more on Bcl-2 and Bcl-xL for survival (Additional file 1: Fig. S2). The annexin V apoptosis assay revealed that 20 μM of MI-238 treatment induced apoptosis in 50.1 ± 1.1% of parental H1299 cells, but intriguingly, 20 μM MI-238 failed to induce apoptosis in Mcl-1 deficient H1299 cells (Fig. 2B). Cleavage of caspase 3 initiates apoptotic DNA fragmentation and is recognized as an apoptosis hallmark (Carneiro and El-Deiry 2020). Consistent with the annexin V assay, MI-238 treatment caused caspase 3 cleavage in H1299 parental cells, but not in Mcl-1 KO cells (Fig. 2E). Similarly, we found that MI-238 treatment also induced apoptosis in wild type (WT) mouse embryonic fibroblast (MEF) cells, but not in Mcl-1 KO MEF cells (Fig. 2D, F). These results demonstrated that MI-238 induced apoptosis is dependent on proficient Mcl-1.

Fig. 2

MI-238 selectively induces apoptosis in Mcl-1 proficient cells. A Western blotting analysis of indicated protein expressions in cell lysate derived from H1299 parental or Mcl-1 knockout (KO) cells. B H1299 parental or Mcl-1 KO cells were treated with or without 20 μM MI-238 for 48 h, and the cell apoptosis were analyzed by annexin V staining. C Western blot analysis as above were performed in MEF wild-type (WT) or Mcl-1 KO cells. D Apoptosis analysis as above were performed in MEF wild-type (WT) or Mcl-1 KO cells. E, F Cell apoptosis were analyzed by caspase 3 cleavage in H1299 (E) or MEF (F). Data are represented as mean ± SD from three independent replicates, ***P < 0.001 by two-tailed t-test

MI-238 effectively kills AML cells

Induction of apoptosis through targeting Bcl-2 anti-apoptotic proteins is an effective therapeutic strategy for hematologic malignancies, and Bcl-2 inhibitor venetoclax is widely used to treat AML and CLL (Ramsey et al. 2018). To test the therapeutic efficacy of MI-238 in AML cells, we treated a variety of AML cells with increasing concentrations of MI-238 and found that the IC50 of MI-238 against AML is around 5–30 μM (Fig. 3A). In addition, the IC50 of MI-238 was inversely proportional to the Mcl-1 expression levels among AML cells (Fig. 3B, C, Additional file 1: Fig. S3). We then examined the PARP1 and caspase 3 cleavage, the well-known apoptosis markers in Molm13 and MV-4–11 cells after MI-238 treatment (Carneiro and El-Deiry 2020). As shown in Fig. 3D, MI-238 induced PARP1 and caspase 3 cleavage in a dose-dependent manner, which indicated MI-238 potently induces apoptosis in AML cells. We further employed annexin V staining to measure the apoptosis frequency in Molm13 and MV-4-11 cells after MI-238 treatment, and found that 40 μM of MI-238 caused 60 ± 0.4% and 35 ± 1.5% apoptotic cell death in Molm13 and MV-4-11 cells respectively (Fig. 3E, F). Taken together, these results demonstrated that MI-238 effectively induces apoptosis in AML cells.

Fig. 3

MI-238 effectively induces apoptosis in AML cells. A The cell viability of indicated 8 AML cell lines in presence of increasing concentrations of MI-238 were analyzed by cell counting kit-8 (CCK8) assay and the half-maximal inhibitory concentration (IC50) IC50 were determined. B Western blot analysis of the expression of Mcl-1, Bcl-2 and Bcl-xL in AML cell lines. C Correlation of Mcl-1 protein level and the IC50 of MI-238 were determined. D, E, F Molm13 and MV-4-11 cells were treated with increasing concentrations of MI-238 for 48 h, and the cell apoptosis was analyzed by western blot analysis of PARP1 and caspase 3 cleavage (D) and flow cytometry of annexin V staining (E, F). Data are represented as mean ± SD from three independent replicates, *P < 0.05, **P < 0.01,  ***P < 0.001, ****P < 0.0001 by two-tailed t-test

MI-238 treatment induces activation of BH3-only proteins

Mcl-1 inhibits apoptosis by sequestering pro-apoptotic BH3-only proteins, such as Bax, Bak, Bim and Puma (Kotschy et al. 2016). To check whether MI-238 treatment could release BH3-only proteins from Mcl-1, we performed an immunoprecipitation (IP) assay using antibodies against anti-apoptotic proteins including Mcl-1, Bcl-2 and Bcl-xL, and the result showed that Mcl-1 mainly binds to Bak, Bim, and Puma, but not Bax in Molm13 cells (Fig. 4A). Meanwhile, treatment of MI-238 could disrupt Mcl-1 association with BH3-only pro-apoptotic proteins including Bak, Bim and Puma (Fig. 4A). Whereas, MI-238 failed to interrupt the binding of BH3-only proteins to Bcl-2 and Bcl-xL (Fig. 4A), suggesting MI-238 treatment specifically inhibits Mcl-1, but not Bcl-2 and Bcl-xL. Bak release from Mcl-1 causes its conformation change and homo-oligomerization to initiate apoptosis. We then employed flow cytometry analysis of Bak activation after MI-238 treatment by staining with activation-specific antibody. Consistent with the IP assay, MI-238 treatment induced Bak activation in Molm13 cells in a dose-dependent manner (Fig. 4B).

Fig. 4

MI-238 disrupts the association of Mcl-1 with BH3-only proteins. A Molm13 cells were treated with indicated concentrations with MI-238 for 24 h, followed by Immunoprecipitation (IP) assay using anti-Mcl-1, anti-Bcl2 and anti-Bcl-xL antibodies and western blot analysis with indicated BH3-only proteins. B Molm13 cells were treated with indicated concentrations of MI-238, followed by flow cytometry analysis of Bak activation using the antibody (clone: G317-2) specifically recognized activated Bak. Data are presented as mean ± SD from three independent replicates, ****P < 0.0001 by two-tailed t-test. C Molm13 cells were treated with or without 20 μM MI-238 for 24 h. Then, subcellular fractionation was subsequently performed and the cytochrome C level in mitochondria and cytosol were analyzed by western blot

Activation of BH3-only proteins results in the release of cytochrome C from the mitochondria into the cytosol, which in turn triggers apoptosis (Kluck et al. 1997). We then performed cell fractionation analysis to examine cytochrome C levels in mitochondrial and cytosol after MI-238 treatment. As shown in Fig. 4C, we observed a significant decrease of mitochondrial cytochrome C level and increase cytosol cytochrome C after MI-238 treatment, indicating MI-238 could induce cytochrome C translocation from mitochondrial to cytosol.

MI-238 synergizes with venetoclax to induce apoptosis in AML cells

Given that Mcl-1 is the primary venetoclax-resistant protein (Ramsey et al. 2018), we then tested whether MI-238 could sensitize AML cells to venetoclax treatment. As shown in Fig. 5A, B, 10 μM of MI-238 induced 34.8 ± 1.2% apoptosis, 0.02 μM of venetoclax caused 26.1 ± 1.3% apoptosis in Molm13 cells, while, their combination induced 87.4 ± 0.3% apoptosis. Besides, synergistic effects of MI-238 and venetoclax on apoptosis induction in Molm13 cells have also been detected in different combinations (10 μM + 0.1 μM, 5 μM + 0.02 μM, and 5 μM + 0.1 μM). (Fig. 5A–F, and Additional file 1: Fig. S4). Consistent with the annexin V staining assay, the combination treatment of MI-238 and venetoclax induced greater caspase 3 cleavage compared with the treatment of MI-238 only or venetoclax only (Fig. 5G). In addition, MI-238 and venetoclax combination induced significantly greater activation of Bak compared with MI-238 or venetoclax treatment alone (Additional file 1: Fig. S5). These results indicate that MI-238 could sensitize AML cells to venetoclax treatment and combination of MI-238 and venetoclax induces synergistic anti-tumor effects.

Fig. 5

MI-238 synergizes with venetoclax to induce apoptosis in AML cells. A–E Molm13 cells were treated with indicated concentrations of MI-238, venetoclax, or their combination for 48 h, followed by apoptosis assay by annexin V staining. Data are represented as mean ± SD from three independent replicates, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by two-tailed t-test. F The combination index (CI) was calculated by CompuSyn software. G Molm13 and MV-4-11 cells were treated as indicated, and the caspase 3 cleavage was analyzed by western blot

MI-238 and venetoclax have a synergistic effect in AML xenografts

To evaluate the therapeutic efficacy of MI-238 and venetoclax in vivo, mice were intravenously (i.v) injected with Molm13 cells stably expressing luciferase (Molm13-Luc) to generate Molm13 AML xenograft model. We started treatment at 10 days after cell implantation, and monitored the cancer progression once a week by the bioluminescence imaging (Fig. 6A). At the beginning of the treatment (10 days after Molm13 implantation), we clearly detected luciferase signal in all mice, that is proportional to amounts of leukemic cells (Fig. 6B). Bioluminescence images obtained after drug treatment (17 and 24 days) showed a significant reduction of leukemia burden in response to MI-238 alone, while, the greater suppression of leukemia progression was seen in the combination treatment compared with MI-238 or venetoclax alone (Fig. 6B).

Fig. 6

The combination of MI-238 and venetoclax potently inhibited the development of AML in murine model. A Schematic diagram of the experimental design showing the timeline for the treatment and imaging. B Representative bioluminescent images of the Molm13 tumor burden in mice treated with vehicle, MI-238 (70 mg/kg), venetoclax (50 mg/kg) or their combination. C, D The percentage of human CD45 (hCD45) and human CD33 (hCD33) positive cells were analyzed by flow cytometry to measure the Molm13 tumor burden. The representative flow cytometry plots (C) and quantification of hCD45 and hCD33 double positive (hCD45+/hCD33+) cells (D) were shown. Data are represented as mean ± SD from three independent replicates. ***P < 0.001 and ****P < 0.0001 by two-tailed t-test. E, F Immunochemistry (IHC) analysis of hCD45 expression in bone marrow from experimental mice. Representative staining (E) and quantification (F) were shown. Data represent mean ± SD from three independent replicates, *P < 0.05 and ****P < 0.0001 by two-tailed t-test. G Kaplan–Meier analysis showed MI-238 in combination with venetoclax resulted in a survival benefit in Molm13 AML xenograft mice. **P < 0.01 by log-rank (Mantel–Cox) test (n = 5)

Meanwhile, the percentage of Molm13 cells in the murine peripheral blood was quantified by flow cytometry using anti-human CD45 (hCD45) and anti-hCD33 monoclonal antibodies, since hCD45/hCD33 double positive was recognized as the human AML marker (Ehninger et al. 2014). As shown in Fig. 6C and D, MI-238 alone treatment significantly reduced the hCD45+ /hCD33 + cells in the peripheral blood compared with vehicle-treated mice (11.7 ± 3.4% vs. 24.7 ± 3.9%). Although, venetoclax alone also decreased percentage of hCD45+ /hCD33+ cells (10.5 ± 2.1%), venetoclax in combination with MI-238 could decrease hCD45+/hCD33+ leukemia cells to 3.0 ± 1.7% (Fig. 6C, D). Similarly, immunohistochemical (IHC) analysis of hCD45+ cells in bone marrow also proved that MI-238 treatment alone could significantly decrease tumor burden, while MI-238 in the combination with venetoclax induced a greater reduction in the tumor burden (Fig. 6E, F). In addition, survival analysis revealed that MI-238 alone or in combination with venetoclax could significantly prolong the survival of tumor-bearing mice (vehicle treated mice = 24.6 days, vs MI-238 treated mice = 28 days, vs venetoclax treated mice = 29.8 days, vs combination treated mice = 35.8 days) (Fig. 6G).

MI-238 treatment alone or in combination with venetoclax is effective in AML patient samples

In order to further validate the therapeutic efficacy of MI-238 and its combination with venetoclax, primary patient AML cells were analyzed. Mononuclear bone marrow cells from 3 different AML patients were treated with increasing concentrations of MI-238 and the apoptosis was analyzed by annexin V staining. Consistent with AML cell line, MI-238 treatment induced apoptosis in AML patient samples in a dose-dependent manner and more than half of bone-marrow mononuclear cells underwent apoptosis in presence of 40 μM MI-238 treatment in all three patient samples (Fig. 7A–C, Additional file 1: Fig. S6). Similarly, we detected increasing cleavage of caspase 3 after treatment of MI-238 (Fig. 7D–F), which confirmed that MI-238 potently induced apoptotic cell death in tumor cells from AML patient samples. Meanwhile, 20 μM of MI-238 treatment failed to induce apoptosis in bone-marrow mononuclear cells from healthy donor (Additional file 1: Fig. S7). Then, we treated patient AML cells with MI-238, venetoclax or their combination to test whether MI-238 could sensitize AML patient samples to venetoclax. As shown in Fig. 7G–I, we detected significantly greater apoptosis in patient AML cells treated MI-238 plus venetoclax compared with cells treated MI-238 or venetoclax alone. Besides, greater cleavage of caspase 3 was detected in patients AML cells treated with MI-238 and venetoclax combination, which further demonstrated that MI-238 is effective in primary patient AML cells (Fig. 7J–L). Besides, a significantly synergistic effects of MI-238 and venetoclax on apoptosis induction in primary AML patient samples (Fig. 7M–O). Collectively, these data demonstrated that MI-238 alone or its combination with venetoclax efficiently induces apoptosis in the bone marrow samples of AML patient, further supporting its therapeutic efficacy to treat AML.

Fig. 7

MI-238 treatment alone or in combination with venetoclax effectively induces apoptosis in primary patient AML cells. Bone-marrow mononuclear cells from three different AML patients were treated as indicated for 48 h, and the apoptosis were measured by annexin V staining (A–C; G–I) or western blot analysis of caspase 3 cleavage (D–F; J–L). The combination index (CI) was calculated by CompuSyn software (M–O). Data represent mean ± SD. *P < 0.05,  **P < 0.01, ***P < 0.001, and ****P < 0.0001