The FDA-Approved and Pharmacopeial Drug Library (HY-L066) was utilized in this study and obtained from MedChemExpress (NJ, United States). Additionally, Visomitin (HY-100474) was purchased from the same supplier. N-acetylcysteine amide (NAC) was purchased from Sigma (Sigma, St. Louis, MO, United States; A9165). The following primary antibodies were used for western blots: Cyclin D1 (Santa Cruz Biotechnology, Dallas, TX, United States; sc-8396), CDK4 (Santa Cruz Biotechnology; sc-23896), p16 (Cell Signaling Technology, Danvers, MA, United States; 80772S), p21 (Cell Signaling Technology; 2947S), Rb (Santa Cruz Biotechnology; sc-102), Mcl-1 (Santa Cruz Biotechnology; sc-12756), Bcl-xl (eBioscience, Vienna, Austria; 14-6994-81), Bak (Santa Cruz Biotechnology; sc-832), Bax (Santa Cruz Biotechnology; sc-7480), P-ATM (Cell Signaling Technology; 5883S), P-Histone H2A.X (Cell Signaling Technology; 9718S), SYK (Santa Cruz Biotechnology; sc-1240), phospho-SYK (Cell Signaling Technology; 2710S), and β-actin (Santa Cruz Biotechnology; sc-4778). HRP-conjugated anti-rabbit and anti-mouse secondary antibodies were obtained from Bethyl Laboratories, Inc. (A120-101p and A90-116p-33; Montgomery, TX, United States).

The human AML cell lines HL60, U937 and mouse AML cell line M1 were purchased from Korean Cell Line Bank (Seoul, Korea). Murine bone marrow cells collected from BALB/c Slu-nu/u mice were purchased from Central Lab Animal Inc. In Seoul, Korea. These cells were cultivated in RPMI-1640 medium (Hyclone, Logan, UT, United States) supplemented with 10% fetal bovine serum (FBS; Hyclone, Melbourne, VIC, Australia), 1% L-glutamine, 1% N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) buffer, and 1% penicillin/streptomycin. Cells were cultured in an incubator at 37 °C under 5% CO2.

Cells were seeded at a density of 3.0 × 10⁵ cells/well in 12-well plates and treated with 500 nM Visomitin (MedChemExpress, NJ, United States; HY-100474) and/or 10 mM N-acetylcysteine (NAC; Sigma-Aldrich) for 72 h. The SYK inhibitor BAY 61-3606 (Sigma) was applied at 10 µM for 72 h. After that, cells were stained with PerCP-Cy™5.5 Mouse Anti-Human CD14 antibody (BD Biosciences) for 1 h at 4 °C. Stained samples were analyzed by using flow cytometry (BD FACSAria™ Fusion Flow Cytometer, BD Biosciences, Becton Drive Franklin Lakes, NJ, United States). The result was then analyzed with the use of FlowJo 7.6 software (TreeStar).

Total cellular RNA was extracted using TRIzol reagent (FATRR 001; Favorgen, Wien, Austria) and transcribed into cDNA using a Prime-Script RT Reagent Kit following the manufacturer’s instructions (RR047A; Takara, Kusatsu-shi, Japan). The differentiation related genes were measured by qRT-PCR using TOPreal qPCR PreMIX SYBR Green with low ROX (RT500M; En-zynomics). PCR was performed as follows: 95 °C for 15 min, 40 cycles at 95 °C for 15 s, 59 °C for 30 s, and 72 °C for 30 s. The primer sequences used in the analysis are provided in Table 1.

Cells were seeded at a density of 2.0 × 10⁵ cells/well in 12-well plates and treated with Artemotil for 72 h. Then, cells were harvested and fixed with methanol and dried. The slides were stained with Giemsa stain (Sigma, 48,900-500ML-F, St. Louis, MO, United States). Then, cells were rinsed with deionized water, air-dried, and observed under a light microscope (Olympus Corporation, Tokyo, Japan) at 400X magnification. The stained cells were assessed for size, regularity of the cell margin, and morphological characteristics of the nuclei. Representative images were captured using Images Plus 2.0 software (Motic Co., Ltd., Xiamen, China).

To determine cell viability, CellTiter 96 AQueous One Solution [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2] cell proliferation assays (MTS) were performed according to the manufacturer’s instructions (Promega, Madi-son, WI, United States). Cells were seeded in 96-well plate at density of 3.0 × 10⁴ cells/well and treated with indicated drugs, and cell viability was measured using MTS assay. The MTS reagent (30 μL) was added to each well and incubated at 37 °C for 4 h. Absorbance was measured at 450 nm using a GloMaxTM Microplate multi-mode reader (Promega).

The cell proliferation rates were determined using trypan blue exclusion assays. Cells were seeded at 2.0 × 10⁵ cells/well in a 24-well cell culture plate on day 0 and counted every 24 h using a hemocytometer (Marienfeld, LaudaKönigshofen, Germany) under a phase-contrast microscope (Olympus CKX41; Olympus Corporation, Tokyo, Japan) after Trypan blue staining (Gibco, 15250061).

To analyze apoptotic rates, human Acute myeloid leukemia cell lines (U937, HL60) were plated in 12-well plates at a density of 1 × 10⁶ cells per well and treated with Visomitin (500 nM) for 72 h or treated with SYK inhibitor (BAY 61-3606, 10 μM) for 24 h. The murine bone marrow cells or AML cell lines M1 were plated in 12-well plates at a density of 1 × 10⁶ cells per well and treated with Visomitin (500 nM) for 24 h. The apoptosis rate was analyzed using flow cytometry (FACSVerse, BD biosciences) after staining with FITC Annexin V Apoptosis Detection Kit I following the manufacturer’s instructions (556,547; BD biosciences, Franklin Lakes, NJ, United States). Cells were first gated based on forward and side scatter (FSC and SSC) to include only live, single cells, thereby excluding apoptotic cells from the gating process. A total of 10,000 events per singlet gate were analyzed. DMSO was used as the vehicle control for Visomitin, and the corresponding solvent was used for NAC or other drug treatments. Compensation was performed using unstained controls and single-stained samples according to the manufacturer’s instructions in the BD FACSAria™ Fusion Flow Cytometer software. The gating strategy is illustrated in Supplementary Figure S1.

U937 cells were seeded in a 12-well plate at a density of 3 × 10⁵ cells per well and treated with Visomitin at a concentration of 500 nM. After 72 h, the cells were incubated with 1 μM of FluoSpheres (F8823; Invitrogen) for 1 h at 37 °C, following the manufacturer’s instructions. The GFP-positive cell population was analyzed using flow cytometry and visualized under a light microscope at ×400 magnification.

Intracellular mitochondrial ROS levels were measured using Dihydrorhodamine 123 (MCE, HY101894). U937 cells were cultured at 3.0 × 10⁵ cells/mL and treated with 500 nM Visomitin for 72 h. Cells were then incubated with Dihydrorhodamine 123 at 37 °C for 30 min, and oxidative burst was assessed by flow cytometry (FACS).

A colony formation assay was performed as previously described. Initially, cells were seeded at a density of 1 × 10⁴ cells per well in 12-well plates and treated with Visomitin. Then, colony-forming cells were evaluated microscopically 8 days after plating. For this microscopic evaluation, a ×400 magnification was achieved using an Olympus CX31 microscope (Olympus Corporation, Tokyo, Japan). Representative images were captured using Images Plus 2.0 software (Motic Co., Ltd., Xiamen, China).

To perform Western blot, U937 cells were seeded in a 12-well plate at a density of 1 × 10⁶ per well, and treated with Visomitin and/or NAC (500 nM, 10 mM, each). Doxorubicin was treated at 100 nM. After 24 h, cells were harvested and lysed in RIPA buffer (ELPIS Biotechnology; Daejeon, Korea) with 1 mM Na-vanadate, 50 mM β-glycerophosphate disodium salt, β-mercaptoethanol (142 mM; BioWORLD, Visalia, CA, United States), ProteaseArrestTM (G-Bioscience; Maryland Heights, MO, United States), and EDTA (5 mM; G-Bioscience). Protein samples were boiled in a 5 X sample buffer at 100 °C for 10 min and loaded onto the poly-acrylamide gels. After gel electrophoresis, proteins were transferred onto Immobilon-P transfer membranes, followed by blocking in 1% bovine serum albumin (BSA; MP Bio-medicals, Santa Ana, CA, United States) dissolved in Tris-buffered saline containing 0.1% Tween-20 (TBST). Membranes were incubated with primary antibodies for 16 h at 4 °C on a rotor and washed thrice for 5 min each with TBST. The membranes were incubated with anti-mouse secondary antibodies or anti-rabbit secondary antibodies at room temperature for 1 h and rinsed thrice for 10 min each with TBST. Protein bands on membranes were treated with a chemilu-minescent substrate (EzWestLumi plus (ATTO, Osaka, Japan)) and visualized using a Luminograph II system (ATTO, Osaka, Japan) (Kim et al., 2019).

To evaluate caspase-3/7 enzymatic activity in AML cells, HL60 cells were treated with 500 nM Visomitin for 24 h. After treatment, Caspase-Glo 3/7 assay reagent (G8090; Promega) was added to the cells, followed by a 1-h incubation at room temperature. Luminescence was then measured using a GloMax™ Microplate multimode reader (Promega).

Bone marrow samples from AML patients were provided by Pusan National University Hospital after obtaining informed consent and approval from the institutional review board (IRB 2403-010-137). The mutation profiles for each patient sample, patient clinical characteristics, including subtype, blast percentage, and treatment status, are summarized in Table 2, and the list of genes screened by targeted NGS is provided in Supplementary Table S2. The AML patient samples were treated with Visomitin for 24 h. Subsequently, the apoptotic rate was evaluated by trypan blue staining.

Human AML cells were seeded in a 12-well plate at a density of 3.0 × 10⁵ cells/well, while mouse normal bone marrow cells or mouse AML cells were seeded in a 12-well plate at a density of 1.0 × 10⁶ cells/well. The cells were treated with 500 nM Visomitin for 24 h. Measurement of mitochondrial superoxide using flowcytometry (BD FACSAria™ Fusion Flow Cytometer; BD Biosciences) was carried out after MitoSOX Red (MedChemExpress; HY-D1055) staining at a final concentration of 5 μM for 30 min, at 37 °C, in dark, according to the manufacturer’s recommendations. The result was then analyzed with the use of FlowJo 7.6 software (TreeStar).

To assess complex I activity, U937 cells (1.0 × 10⁷) were treated with 500 nM Visomitin for 24 h. Complex I activity was then measured using the Complex I Enzyme Activity Microplate Assay Kit (Colorimetric; Abcam, ab109721) according to the manufacturer’s instructions. Luminescence was measured at 450 nm using a GloMax™ Microplate Multimode Reader (Promega).

For transient expression of SYK, U937 cells were cultured at 3.0 × 10⁵ cells/well and transfected with 100 ng of SYK overexpression vector or control vector using Metafectene PRO (Biontex, Germany) according to the manufacturer’s instructions. Cells were maintained in RPMI-1640 medium without penicillin/streptomycin. The DNA construct was kindly provided by Prof. Jae Youl Cho, Department of Bioinformatics and Life Science, Soongsil University, Korea.

All animal care and experimental procedures were approved by the Pusan National University-Institutional Animal Care and Use Committee (Approval No. PNU-2021-3056) and carried out in compliance with the university’s scientific research guidelines and regulations.

Male BALB/c Slu-nu/u mice, aged 4 weeks, were procured from Central Lab. Animal Inc. In Seoul, Korea. The mice were allocated randomly to different experimental groups.

Xenograft tumors were established by injecting 1.0 × 10⁷ U937 cells subcutaneously into the flank of each mouse, with daily monitoring of tumor growth. The mice were allocated randomly into a vehicle group and a treatment group. An Visomitin solution was prepared in DMSO and delivered via an intraperitoneal injection. Mice bearing xenograft tumors were treated with 4 mg kg^-1 Visomitin for 1 week and 2 mg kg^-1 Visomitin for the remaining days. The initial dose was chosen based on previously reported toxicity studies of Visomitin (Petrov et al., 2016) and in vitro activity, and the reduction was made after a slight body weight decrease to maintain tolerability while preserving anti-tumor efficacy. The tumor volumes (mm³) were calculated using the following formula: (A) × (B2)/2, where A was the largest diameter (mm) and B the smallest (mm). Cyclophosphamide and cyclosporine were administered intraperitoneally to mice for immunosuppression 7 days before tumor cell injection, at doses of 60 mg/kg and 10 mg/kg, respectively. Xenograft tumor tissues were disrupted between the frosted ends of microscope slides, and a single tumor cell suspension was produced using a cell strainer (SPL; 93,100) to remove the aggregates and debris. The collected tumor cells were washed with PBS, centrifuged for 5 minutes at 2000 rpm at room temperature, and stained with MitoSOX Red of 5 μM for 30 min, at 37 °C, in dark, according to the manufacturer’s recommendations. The stained samples were analyzed by flow cytometry, and the results were processed using FlowJo 7.6 software (TreeStar).

IHC staining were performed as previously described (Kim et al., 2018). For IHC staining of Ki-67 in tissues, a polyclonal anti-Ki-67 antibody and a polymer-horseradish peroxidase anti-rabbit antibody were used as a primary and a secondary antibody, respectively, followed by 3,3-diaminobenzidine treatment to visualize the proteins.

The tissues from mice were fixed with a 10% neutral buffered formalin solution (Sigma; St Louis, MO, United States) and embedded in paraffin. Paraffin sample blocks were sectioned into 4 M, deparaffinized, hydrated, and stained with H&E. H&E-stained samples were observed with an Olympus CX31 microscope (Olympus Corporation, Tokyo, Japan) at 100 magnification. Representative images were captured using Images Plus 2.0 software (Motic Co. Ltd., Xiamen, China) (Kim et al., 2018).

All procedures were replicated independently a minimum of three times to ensure the reliability of the experimental results. The data are represented as the mean ± standard deviation (SD). Either a Mann–Whitney U test or a one-way ANOVA, followed by Tukey’s post hoc analysis, were applied to identify the significant differences. These statistical analyses were performed using Microsoft Office Excel and GraphPad Prism five software (GraphPad Software, Inc., San Diego, CA, United States).

To identify potential differentiation-inducing agents for acute myeloid leukemia (AML), we randomly selected and screened 1,140 compounds from the FDA-Approved and Pharmacopeial Drug Library using the human AML cell line U937 (Supplementary Table S1). CD11b expression—a cell surface marker upregulated during myeloid cell differentiation and commonly used as a primary indicator—was measured by flow cytometry (FACS) analysis. Hits were defined based on a chemical/ATRA activity ratio greater than 0.5. Among the compounds, Visomitin exhibited a ratio of 0.572 and was the only compound meeting the predefined cutoff, demonstrating AML-selective activity. Visomitin also notably increased CD11b expression and was therefore prioritized as a potential differentiating agent for AML treatment (Figure 1). Other compounds did not surpass the threshold, highlighting the specificity of Visomitin as a candidate.

Next, to further evaluate the therapeutic potential of Visomitin in AML, we assessed its ability to induce differentiation in two human AML cell lines, U937 and HL60, using CD14 as a surface marker. CD14 is known to be upregulated during the differentiation of myeloid cells into macrophages. FACS analysis revealed that Visomitin treatment significantly upregulated CD14 expression in both U937 and HL60 cells (Figure 2A). Morphological changes indicative of myeloid differentiation, such as the emergence of granulocyte-like structures with multilobed nuclei and cytoplasmic granules, were observed via Giemsa staining (Figure 2B). The phagocytic activity of cells with granulocyte-like structures induced by Visomitin treatment was assessed using FluoSphere beads and was found to be increased compared to the control group (Figure 2C). In addition, oxidative burst activity—an established functional feature of mature myeloid cells detectable by DHR-123 oxidation (Trayner et al., 1995; Smoljo et al., 2023)—was significantly increased in Visomitin-treated cells (Figure 2D), providing further evidence of functional differentiation. Together, these morphological changes, increased phagocytosis, and enhanced oxidative burst suggest that Visomitin efficiently induces differentiation in AML cells. Additionally, Visomitin treatment upregulated nine myeloid differentiation markers, including MMP9, c-Jun, and ITGAM. Conversely, it significantly reduced the expression of c-Myc, Myb, and SOX4, which are known to inhibit the differentiation of normal hematopoietic cells and to promote the onset and progression of cancer (Figure 2E). Furthermore, to assess macrophage gene signatures as an additional measure of functional differentiation, we examined mRNA expression of M1-associated markers (CD86, TNF-α, IL-1β) and M2-associated markers (CD163, IL-10) (Yuan et al., 2015; Murray, 2017; Wu et al., 2022; Strizova et al., 2023). All markers were significantly upregulated upon Visomitin treatment in U937 cells, indicating differentiation toward a macrophage-like phenotype (Figure 2F). These results collectively suggest that Visomitin induces differentiation of AML cells.

The differentiation of AML has been reported to be closely associated with cell cycle arrest, which plays a critical role in regulating cell proliferation and differentiation (Ketley et al., 2000; Schnerch et al., 2012). In order to better understand the impact of differentiation induced by Visomitin on the cell cycle dynamics, we specifically examined the alterations in cell cycle progression and arrest in AML cell lines treated with Visomitin. Notably, a significant cell cycle arrest was observed, with an increase in the proportion of cells in the G0/G1 phase when compared to the control group following Visomitin treatment, suggesting an inhibitory effect on cell cycle progression (Figure 3A). To further investigate the underlying molecular mechanisms, Western blot analysis was performed, revealing that Visomitin treatment led to a marked downregulation in the expression levels of cyclin D1 and CDK4, both of which are essential regulators of cell cycle progression during the G0/G1 phase (Figure 3B). In addition, the expression of key cyclin dependent kinase inhibitors (CKIs), including p21 and p16, which act by binding to CDK complexes and inhibiting the activation of cyclin/CDK complexes to promote cell cycle arrest, was significantly increased following Visomitin treatment, further corroborating the observed cell cycle arrest (Figure 3C). Taken together, these findings suggest that Visomitin may have potential anti-leukemia activity by disrupting critical cell cycle regulatory pathways.

Above data show that Visomitin significantly induced differentiation and cell cycle arrest in AML. Our next aim was to investigate whether Visomitin exhibited anti-tumor effects, particularly by inhibiting cell viability. To evaluate this, we treated U937 and HL60 cells with varying concentrations of Visomitin and assessed cell viability. The results clearly demonstrated that Visomitin treatment significantly inhibited cell viability in both cell lines in a dose-dependent manner, as shown in Figure 4A. In addition, Trypan blue staining was performed to further investigate the effect of Visomitin on cell proliferation. The staining revealed that Visomitin not only reduced cell proliferation but also exerted significant effects on cell survival (Figure 4B). Further analysis using PI and Annexin V staining followed by FACS analysis provided more insight, showing that Visomitin effectively induced apoptosis in AML cells, which is a critical mechanism contributing to its cytotoxic effects (Figure 4C). In addition, colony-forming assays were performed to assess the clonogenic potential of AML cells. These assays demonstrated that Visomitin impaired the ability of AML cells to form colonies, further confirming its inhibitory effects on cell survival (Figure 4D). Taken together, these data suggest that Visomitin exerts potent cytotoxic effects in AML.

Next, we examined the expression of Bcl-2 family proteins to better understand how Visomitin induces apoptosis in AML cells. Western blot analysis revealed changes in the expression of pro- and anti-apoptotic proteins from the Bcl-2 family after Visomitin treatment. Visomitin decreased the expression of the anti-apoptotic proteins Mcl-1 and Bcl-xL, while increasing the expression of the pro-apoptotic proteins Bax and Bak (Figure 4E). This implies that an imbalance between pro- and anti-apoptotic Bcl-2 family members may contribute to the Visomitin-induced cell death.

The misproportion disrupts the mitochondrial membrane potential (MMP), leading to the release of cytochrome c into the cytoplasm. To investigate the mitochondrial dynamics associated with the increase in Bax and Bak and the decrease in Mcl-1 and Bcl-xL, we treated AML cell lines with Visomitin and measured MMP using the mitochondrial ΔΨm-sensitive dye JC-1. Control cells with high ΔΨm displayed red-fluorescent J-aggregates formed by JC-1. However, a decrease in ΔΨm resulted in the disassembly of these aggregates into monomers, leading to a fluorescence shift from red to green. Treatment with Visomitin increased green fluorescence and significantly reduced the red/green fluorescence ratio, indicating a disruption of the MMP (Figure 4F). Additionally, we further examined caspase activation. Apoptosis is regulated by a family of cysteine-aspartic acid-specific proteases known as caspases. Upon initiation of apoptotic signaling, caspases are classified into two groups: initiator caspases (e.g., caspase-9 and -10) and effector caspases (e.g., caspase-3, -6, and -7). Visomitin was found to induce caspase-3/7 activity (Figure 4G). These results suggest that Visomitin primarily regulates Bcl-2 family proteins, thereby activating the intrinsic apoptosis pathway, destabilizing mitochondrial functions, and inducing apoptosis. Collectively, Visomitin inhibits AML cell survival by inducing differentiation, cell cycle arrest, inhibiting cell proliferation, and promoting apoptosis, ultimately leading to the anti-cancer effects in AML.

To further validate the anti-tumor effects of Visomitin in a clinical context, we treated five primary AML patient samples with Visomitin. Remarkably, treatment with Visomitin led to a reduction in cell proliferation in all patient samples, further supporting its potent cytotoxic effects in AML (Figure 4H). Although the number of samples is small and robust correlations between individual genetic alterations and drug response cannot be determined, treatment with Visomitin reduced cell viability by 80% in patient 1, 33.3% in patient 2, 58.8% in patient 3, 35.7% in patient 4, and 80% in patient 5, suggesting that its anti-leukemic activity is consistent across diverse genetic backgrounds. These results suggest that Visomitin’s ability to inhibit cell proliferation is not limited to cell lines but extends to primary AML patient cells as well, highlighting its potential therapeutic relevance in clinical settings.

Unlike the reported potent antioxidant activity of Visomitin, as shown in Figure 5A, we found that Visomitin-treated AML cells exhibited higher mitochondrial ROS (mtROS) levels compared to the control group. We hypothesized that Visomitin contributes to apoptosis and differentiation in AML cells by accumulating mtROS, and to test this hypothesis, we assessed cell counting and CD14 expression after Visomitin treatment, both in the presence and absence of N-acetyl-L-cysteine (NAC). The results showed that NAC significantly restored cell differentiation and survival by inhibiting Visomitin-induced mtROS generation (Figures 5B,C). Next, we aimed to explore the mechanism by which Visomitin-induced mtROS accumulation leads to apoptosis and differentiation in AML cells. To investigate whether the accumulation of mtROS is linked to mitochondrial dysfunction, we performed a mitochondrial complex I activity assay. Visomitin treatment significantly reduced complex I activity, as indicated by a decreased rate of OD increase (mOD/min) compared to the control (Figures 5D,E). These results suggest that inhibition of complex I may contribute to the selective accumulation of ROS in AML cells through impaired electron transfer and increased electron leakage (Li et al., 2003; Hirst, 2013; Jang and Javadov, 2018), providing a potential mechanistic explanation for the pro-oxidant effect of Visomitin in malignant mitochondria. High levels of ROS have been reported to induce oxidative stress and damage cellular components, including lipid membranes and nucleic acids. To determine whether ROS-induced DNA damage contributes to apoptosis, we evaluated the expression of p-Ataxia-Telangiectasia Mutated (ATM) protein kinase and Phosphorylated H2A histone family member X (γH2AX), markers of DNA double-strand breaks, using Western blot analysis. The expression of p-ATM and γH2AX was significantly increased by Visomitin treatment in U937 cells and was abolished by NAC treatment (Figure 5F). To investigate whether ROS accumulation-induced DNA damage leads to cell cycle arrest and apoptosis (Feng et al., 2019; Kwon et al., 2025), we evaluated the activation of the p21/p16/Rb axis. We observed that the increased expression of p21, p16, and Rb following Visomitin treatment was inhibited by NAC treatment (Figure 5G). Therefore, these results suggest that ROS accumulated by Visomitin induce DNA damage and activate the p21/p16/Rb axis, leading to differentiation and apoptosis in AML cells.

Visomitin is known to accumulate in mitochondria due to its positive charge and effectively scavenge mitochondrial reactive oxygen species (mtROS) through the antioxidant action of its plastoquinyl group. Intriguingly, treatment with Visomitin resulted in a significant increase in ROS levels in AML cells (Figure 5A), which was contrary to its known ROS-inhibiting effects. To investigate whether this increase in ROS accumulation is specific to AML cells or if it reflects a previously unrecognized ROS-inducing activity of Visomitin, we examined the ROS levels in both normal mouse bone marrow cells and the mouse AML cell line M1 following Visomitin treatment. In total mouse bone marrow cells, there was no significant difference in ROS levels between the control and Visomitin treatment groups (Figure 6A). However, in the myeloid cell population (CD45+/CD11b+), ROS levels were notably decreased following Visomitin treatment compared to the control group (Figure 6B), suggesting that Visomitin may inhibit ROS accumulation in normal myeloid cells. In contrast, treatment with Visomitin significantly induced ROS accumulation in the mouse AML cell line M1 compared to the control, indicating that Visomitin specifically induces ROS accumulation in AML cells (Figure 6C). To further assess the effect of Visomitin on cell survival of normal mouse myeloid cells and mouse AML cells, we evaluated apoptosis by measuring the population of Annexin V-positive cells after exposure to Visomitin. Visomitin treatment resulted in a significant increase in the Annexin V-positive cells in M1 cells, not in normal bone marrow cells or myeloid cells (Figures 6A–C). Consistent with these findings, Visomitin treatment did not induce γH2AX expression in normal mouse bone marrow cells (Supplementary Figure S2), whereas γH2AX induction was observed in AML cells (Figure 5F), indicating a differential DNA damage response between normal and malignant cells. These findings support a model in which Visomitin exerts context-dependent redox effects: it remains non-pro-oxidant in normal hematopoietic cells, while selectively disrupting mitochondrial respiratory function in AML cells, thereby exacerbating intrinsic redox imbalance and promoting ROS-mediated apoptosis.

Next, we aimed to identify the signaling pathway responsible for mediating the activation of the p16/p21/Rb axis induced by ROS accumulation. We hypothesized that the SYK could be a plausible candidate (Cheng et al., 2011; Wang et al., 2014; Qiu et al., 2015; Ishikawa et al., 2018). Treatment with Visomitin significantly reduced the expression of phospho-SYK (Figure 7A). To determine whether the blockade of the SYK activation by Visomitin was induced by ROS, we observed the phosphor-SYK after hydrogen peroxide treatment. In the group where ROS production was induced by hydrogen peroxide, the p-SYK/SYK ratio significantly decreased compared to the control (Figure 7B). To clarify whether SYK inhibition leads to cell differentiation, we treated the AML cells with the SYK inhibitor BAY 61-3606 and observed the expression of CD14. SYK inhibition significantly increased the expression of CD14 (Figure 7C). Additionally, to investigate the effect of SYK inhibition on apoptosis, we performed Annexin V staining followed by FACS analysis. Treatment with the SYK inhibitor significantly increased the proportion of apoptotic cells compared to control (Figure 7D). Similarly, SYK inhibition in AML cells reduced cell viability, further supporting the role of SYK in regulating apoptosis in AML cells (Figure 7E). To investigate the causal relationship between ROS accumulation and SYK inhibition, we performed rescue experiments by overexpressing SYK. SYK overexpression significantly rescued Visomitin-induced CD14 expression and apoptosis in AML cells (Figures 7F,G). To further explore the specificity of SYK regulation by Visomitin-induced ROS, we compared the effects of doxorubicin, a well-known ROS-inducing chemotherapeutic agent, on SYK activation (Chen et al., 2020). Notably, doxorubicin treatment led to increased SYK phosphorylation in U937 cells, consistent with previous reports linking doxorubicin-induced ROS to SYK activation (Ma et al., 2020) (Figure 7H). In contrast, Visomitin treatment resulted in SYK inactivation despite also elevating ROS levels (Figure 7A). These opposing outcomes suggest that SYK inhibition may be a Visomitin-specific effect, rather than a general consequence of ROS accumulation. Taken together, these findings suggest that ROS accumulation induced by Visomitin regulates the SYK activation, promoting both differentiation and apoptosis in AML cells.

Visomitin exerted anti-tumor effects by inducing differentiation and apoptosis in vitro. To investigate whether Visomitin’s anti-cancer effects would be recapitulated in vivo, immunocompromised athymic nude mice were inoculated with U937 AML cells, followed by administration of Visomitin via intraperitoneal injection for 10 days and evaluation of its effects on tumor growth. Although SkQ1 is primarily developed as an ophthalmic solution, previous studies have shown that it is well-tolerated in rodents even at high oral doses (up to 100 mg/kg) (Petrov et al., 2016). While precise pharmacokinetic parameters in mice are not fully reported, SkQ1 selectively accumulates in mitochondria and distributes broadly in tissues due to its lipophilic cation structure (Saric et al., 2016). Considering that in vitro activity was observed at ∼500 nM, we selected a higher systemic dose of 4 mg/kg to achieve sufficient exposure in the xenografted tumors. Visomitin-treated mice showed markedly reduced tumor sizes and volumes compared to the control mice (Figures 8A,B). To further evaluate the mechanism of tumor growth inhibition, Ki-67 immunohistochemistry (IHC) was performed on tumor tissues. In vehicle-treated tumors, numerous nuclei were positively stained for Ki-67, indicating active proliferation (Sun and Kaufman, 2018), whereas Visomitin-treated tumors exhibited almost no Ki-67-positive nuclei (Figure 8C), suggesting that Visomitin directly suppresses tumor cell proliferation rather than causing a general growth delay. Importantly, when the ROS levels in tumor cells from the Visomitin-treated group were measured, we observed a significant increase in intracellular ROS accumulation compared to the control, consistent with the data indicating that visomitin increased ROS levels in AML cells (Figures 5A, 8D). Evaluation of Visomitin’s cytotoxicity was conducted through H&E staining of kidney, heart, lung, and liver tissues. No tissue damage induced by Visomitin was observed in the major organs (Figure 8E). In addition, the body weights were not different between the Visomitin-treated group and the control group, which indicates that Visomitin does not systemic cytotoxicity (Figure 8F). Taken together, these results clearly show that Visomitin exhibits anti-AML effects in vivo by inducing ROS production.