FBXO3-mediated DUSP9 ubiquitination promotes leukemia stem cell maintenance and tyrosine kinase inhibitor resistance in chronic myeloid leukemia
Xudong Li, Shiyu Zuo, Yanli Zhang, Zexing Liu, Na Shen, Qingqing Ma, Mingxia Sun, Binglei Zhang, Mengjia Li, Hong Huang, Mengya Gao, Zhenghua Huang, Huifang Zhao, Yilin Chen, Fengcai Gao, Wenjuan Fan, Zhen Zhang, Yuhan Hu, Yu An, Siyue Li, Miao Liu, Yupeng Liu, Yuxuan Liu
TL;DR
This study shows that FBXO3 helps leukemia stem cells survive and resist treatment in chronic myeloid leukemia by modifying DUSP9, suggesting a new way to target these cells.
Contribution
FBXO3 is identified as a novel marker and therapeutic target for CML-LSCs through its role in DUSP9 ubiquitination and MAPK activation.
Findings
FBXO3 is highly expressed in TKI-resistant CML stem cells and promotes their survival.
FBXO3 mediates DUSP9 ubiquitination, activating the MAPK pathway essential for CML progression.
Targeting FBXO3 eliminates LSCs without harming normal blood stem cells.
Abstract
Eradicating leukemia stem cells (LSCs) and overcoming tyrosine kinase inhibitor (TKI) resistance is urgent for chronic myeloid leukemia (CML) treatment. We find that F-box protein 3 (FBXO3) is highly upregulated in CD34+ CML stem cells from TKI-resistant patients and identify it as an innovative CML-LSC marker via single-cell RNA sequencing (scRNA-seq). FBXO3 deficiency induces apoptosis and reduces proliferation of CML cell lines and LSCs in vitro and in vivo, with minimal effects on normal CD34+ hematopoietic stem cells (HSCs). Mechanistically, FBXO3 interacts with DUSP9 to promote its ubiquitination and activate the MAPK pathway, critical for CML cell activity. DUSP9 knockdown partially reverses FBXO3-deficiency-mediated LSC elimination. Furthermore, FBXO3 inhibitor monotherapy or combination with imatinib effectively eradicates CML-LSCs, overcomes TKI resistance, and spares normal…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsUbiquitin and proteasome pathways · Chronic Myeloid Leukemia Treatments · Acute Myeloid Leukemia Research
Introduction
The introduction of tyrosine kinase inhibitors (TKIs) targeting BCR::ABL1 has transformed chronic myeloid leukemia (CML) from a fatal malignancy into a clinically manageable disease for most patients.1^,^2 Nevertheless, TKI resistance remains a formidable challenge in CML management due to its complexity.3^,^4 Therefore, elucidating the mechanisms underlying TKI resistance is essential for developing more effective therapeutic strategies and ultimately improving patient outcomes.
CML progression is fundamentally attributed to the constitutive activity of BCR::ABL1, either through acquired TKI resistance, such as that conferred by the T315I mutation, which imparts resistance to imatinib and other TKIs,5^,^6^,^7^,^8 or via upregulated BCR::ABL1 expression mediated by factors like PABPC1.9 In addition to these BCR::ABL1-dependent processes, BCR::ABL1-independent mechanisms that activate alternative signaling cascades enabling leukemic cell survival despite TKI treatment have also emerged as critical contributors to TKI resistance.10 For instance, activation of the RAS-MAPK-ERK pathway has been identified as a potential contributor to imatinib resistance in CML with unaltered BCR::ABL1 function.11^,^12 Furthermore, mTOR inhibitors have been shown to overcome TKI resistance by enhancing autophagy independently of BCR::ABL1.13 Epigenetic modifications and alterations in gene expression, including PRMT1, PRMT5, PRMT7, and SMYD3, are also involved in TKI resistance.14^,^15^,^16^,^17 Moreover, the persistence of leukemia stem cells (LSCs), which are intrinsically quiescent and thus resistant to TKIs, substantially contributes to disease relapse.16^,^18^,^19^,^20^,^21^,^22^,^23 These findings underscore that CML relapse is primarily driven by LSC persistence and TKI resistance, highlighting the imperative to elucidate and develop CML-LSC-specific targets.
Moreover, accumulating evidence suggests that the dysregulation of protein stability control systems, particularly the ubiquitin-proteasome system (UPS), plays a pivotal role in mediating both intrinsic and acquired resistance mechanisms in various cancer types, which has emerged as a clinically validated approach.24^,^25 E3 ubiquitin ligases, in particular, are responsible for the specificity of substrate recognition and ubiquitination, which marks proteins for degradation by the proteasome.26 This system is intricately involved in various cellular processes, including cell-cycle regulation, apoptosis, and signal transduction, making it a significant modulator of therapeutic resistance in cancer and other diseases.27^,^28^,^29^,^30 Among E3 ligases, FBXO3, an F-box protein within the FBXO family, is widely involved in these processes.31 However, the specific role of FBXO3 in the maintenance of LSC survival and TKI resistance remains unknown.
In this study, we use single-cell RNA sequencing (scRNA-seq) to identify FBXO3 as an innovative marker of CML-LSCs that is frequently upregulated in patients exhibiting TKI resistance. We identify that FBXO3 plays a critical role in sustaining CML-LSC survival and promoting TKI resistance. FBXO3 deficiency significantly inhibits the survival of CML cells and CML-LSCs, while sparing normal CD34^+^ hematopoietic stem cells (HSCs) in vitro and in vivo. Mechanistically, FBXO3 facilitates TKI resistance via the ubiquitin-mediated degradation of DUSP9, thereby activating the mitogen-activated protein kinase (MAPK) signaling. Notably, the combination of the FBXO3 inhibitor with imatinib enhanced the anti-leukemia effect against CML-LSCs with minimal effect on normal hematopoiesis. Our findings elucidate previously unrecognized roles of FBXO3 in LSCs and TKI resistance in CML and highlight its potential as a therapeutic target to improve outcomes in CML.
Results
FBXO3 expression identifies therapy-resistant CML-LSCs and is particularly elevated in TKI-resistant patients
To investigate the molecular mechanism of TKI resistance, we analyzed transcriptomic datasets and revealed that FBXO3 was significantly upregulated in primary CML CD34^+^ cells compared to control (Figure S1A). Notably, patients with blast crisis CML (CML-BC) demonstrated higher FBXO3 expression than those in the chronic phase (CML-CP) (Figure S1A). Further analysis revealed that TKI treatment at both 3 months and 6 months attenuated the expression of FBXO3, and patients with good responses to TKI therapy exhibited a significant decline in FBXO3 expression, whereas those with poor responses still maintained higher levels (Figure S1B).
To further elucidate the role of FBXO3 in CML, we performed scRNA-seq analyses on bone marrow CD34^+^ cells from CML patients (N = 17, GEO: GSE236233) and healthy controls (HCs, N = 13, GEO: GSE173076). After quality control, 43,385 cells were clustered into 16 groups and annotated into 9 hematopoietic lineages (Figures S1C–S1E, 1A; Table S1). Notably, the levels of FBXO3 in CML-derived HSCs were significantly higher compared to those in the control group (Figure 1B). Given the heterogeneity of primitive CML stem cells, which encompasses both BCR::ABL1^+^ and BCR::ABL1^−^ subpopulations, we re-clustered CML-LSCs using differentially expressed genes (DEGs) identified by Giustacchini et al.,32 identifying seven distinct subpopulations (Figure S1F; Tables S2 and S3). Clusters 1 and 3 were identified as BCR::ABL1^−^ subsets, while the other five were identified as BCR::ABL1^+^ (Figures S1G–S1J). The BCR::ABL1^+^ LSCs were further divided into FBXO3^+^ and FBXO3^−^ subsets (Figures S1K–S1N; Table S4). Gene set expression comparison between FBXO3^+^ LSCs and FBXO3^−^ LSCs indicated that genes associated with stemness, leukemia persistence, and TKI resistance were predominantly enriched in FBXO3^+^ LSCs (Figures 1C and S2A; Table S5). To validate whether FBXO3^+^ cells conform to the established LSCs transcriptional program, we validated the expression of previously characterized eight upregulated (IL3RA, FCGR2A, ITGAX, ITGB7, IL1RAP, DPP4, IL2RA, and C1QBP) and one downregulated (KIT) LSC-related genes.32^,^33^,^34 Our results revealed that eight genes have the same expression pattern (Figure 1D). We next investigated the correlations between the nine genes and FBXO3. Our results showed that the levels of IL3RA, FCGR2A, ITGAX, ITGB7, IL1RAP, DPP4, and IL2RA are positively correlated with FBXO3 expression, whereas C1QBP was negatively related (Figure 1E). Although KIT was enriched in FBXO3^−^ LSCs, its expression is positively correlated with FBXO3 expression (Figure S2B).Figure 1FBXO3 overexpression in CML correlates with poor prognosis(A) UMAP projection of 43,385 BM CD34^+^ cells (17 CML patients and 13 HCs), including CLP, EBMP, Ery, GMP, HSCs, MEP, MKP1, MKP2, and MPP.(B) FBXO3 expression in HSCs (HCs vs. CML). Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001.(C) BCR::ABL1-related gene signature scores in FBXO3^−^ vs. FBXO3^+^ SCs (∗∗∗∗p < 0.0001, Wilcoxon test; boxplots show median/IQR).(D) Heatmap of 10 LSC-associated genes in FBXO3^+/−^ stem cells.(E) Significant Spearman correlations between FBXO3 and eight LSC markers (R and p values shown).(F) FBXO3 mRNA levels in PBMCs (N = 9) and CML cell lines detected by RT-qPCR. Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(G and H) FBXO3 protein levels in PBMCs (N = 9) and CML cell lines detected by WB analysis. Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(I) FBXO3 mRNA in CD34^+^ cells from HCs (N = 8) vs. CML patients (optimal response, N = 10; sub-optimal response, N = 8; TKI treatment failure, N = 8). Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance.(J and K) FBXO3 protein in CD34^+^ cells (HCs: N = 8 vs. CML: N = 26; WB). Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(L and M) FBXO3 protein in BM cells from BCR::ABL1-induced CML mice vs. controls. Data are represented as mean ± SEM; t test; ∗p < 0.05.
We also validated the expression of FBXO3 in CML cell lines (K562, KBM5, KBM5-T315I, Ku812, KCL22, and BV173), primary CML CD34^+^ cells with distinct response to TKI treatment and CML mice model. All the CML cell lines used in the present study demonstrated higher FBXO3 expression compared to those in control group (Figures 1F–1H). Consistently, FBXO3 expression was significantly higher in patients with sub-optimal response or treatment failure of TKI therapy compared to those with an optimal response or HCs (Figures 1I–1K). The Fbxo3 expression was also elevated in BCR::ABL1-induced CML mice compared to WT mice (Figures 1L and 1M). Collectively, our data revealed that FBXO3 may serve as a potential biomarker for identifying and targeting CML-LSCs to overcome TKI resistance.
FBXO3 deficiency selectively impairs survival and self-renewal of CML cells but not normal CD34+ cells
To investigate the role of FBXO3 in CML, we conducted short hairpin RNA (shRNA)-mediated knockdown of FBXO3 in K562, KBM5, and KBM5-T315I cells and primary CML CD34^+^ cells. Then we verified that FBXO3 was efficiently reduced at both mRNA and protein levels (Figures S3A–S3C), which led to a significant reduction in proliferation (K562 and KBM5 by day 5; KBM5-T315I by day 7) and impaired colony-forming capacity across all cell lines (Figures S3D and S3E). To define the basis for the reduced cell growth, we measured apoptosis and mitochondrial membrane potential by flow cytometry. Our results indicated that FBXO3 knockdown significantly increased the percentage of annexin V^+^ cells and mitochondrial membrane potential levels in CML cells (Figures S3F and S3G). To further demonstrate the roles of FBXO3 in CML, we generated CRISPR/Cas9-mediated FBXO3 knockout CML cells. We found that FBXO3 knockout led to cell proliferation and colony formation inhibition, as well as enhanced cell apoptosis in K562 and KBM5-T315I cells (Figures 2A–2E).Figure 2FBXO3 deletion suppressed IM-sensitive/-resistant cells and primary CML CD34^+^ cells through PARP1-mediated endogenous apoptosis in vitro(A) CRISPR-Cas9-mediated FBXO3 knockout in K562/KBM5-T315I cells confirmed by western blot (left: representative blot; right: quantification, n = 3). Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001.(B) Growth curves of control/sgFBXO3 cells over 7 days (n = 3). Data are represented as mean ± SD; two-way ANOVA; ∗∗∗p < 0.001.(C) Soft agar colony formation (≥50 cells/clone, scale bars: 100 μm). Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01.(D and E) Enhanced apoptosis in sgFBXO3 cells: (D) annexin V/DAPI staining; (E) JC-10 mitochondrial depolarization. Data are represented as mean ± SEM; t test; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 3.(F) Western blot analysis of apoptosis-related proteins in control/sgFBXO3 cells. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns indicates no significance.(G) shRNA-mediated FBXO3 knockdown efficiency in CML/normal CD34^+^ cells. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(H) Colony formation assays in primary CML and normal CD34^+^ cell; scale bars, 100 μm, n = 3. Data are represented as mean ± SEM; t test; ∗∗p < 0.01; ns indicates no significance.(I) Serial replating capacity of CD34^+^ cells (three generations). Data are represented as mean ± SEM; t test; ∗∗p < 0.01; ∗∗∗p < 0.001; ns indicates no significance.(J and K) Apoptosis in FBXO3-knockdown CD34^+^ cells: (J) annexin V/DAPI (n = 3); (K) annexin V/CFSE (n = 3). Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01; ns indicates no significance.(L and M) WB analysis of PARP1-mediated apoptotic proteins in CD34^+^ cells; n = 3. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01; ns indicates no significance.
To elucidate the mechanistic basis for the increased apoptosis, we investigated the activation of the intrinsic apoptosis pathway. Western blot (WB) analysis demonstrated elevated apoptosis-related proteins, including PARP1, cytochrome c, and cleaved caspase-3, alongside reduced anti-apoptotic proteins, such as BCL-xL, BCL-2, and MCL-1, in both FBXO3 knockdown and knockout cell lines (Figures S3H, S1I, and 2F). We also used shRNA to knockdown FBXO3 in primary CD34^+^ cells from CML patients and normal CD34^+^ cells from HCs (Figure 2G). Similarly, FBXO3 knockdown in primary CML CD34^+^ cells also promoted apoptosis and inhibited colony formation (Figures 2H–2K). Importantly, FBXO3 knockdown had no significant effect on the survival and self-renewal of normal CD34^+^ cells (Figures 2H–2K). Further validation confirmed that FBXO3 deficiency activated the intrinsic apoptotic pathway, as evidenced by elevated PARP1 cleavage, caspase-3 activation, cytochrome c release, and downregulation of anti-apoptotic Bcl-2 family members (Figures 2L and 2M). These results indicate that FBXO3 is a critical factor for maintaining the survival and self-renewal of CML CD34^+^ cells.
FBXO3 deficiency suppresses leukemia progression in CML-CDX models and prolongs survival and eradicates LSCs in BCR::ABL1-driven CML mice
To verify the in vivo therapeutic potential of targeting FBXO3, we developed a CML-CDX model via tail vein injection of K562 cells. FBXO3 knockout significantly reduced leukemic burden in peripheral blood (PB), bone marrow (BM), and spleen (SP) in CML murine model (Figures 3A–3D). Notably, FBXO3 knockout prolonged median survival from 59 days (controls) to 83 days in CML-bearing mice (Figure 3E). To further evaluate the role of Fbxo3 in CML, we established a BCR::ABL1-induced CML murine model using BM and SP cells transduced with Fbxo3-targeting shRNA (Figure 3F).35 We confirmed the effective knockdown of Fbxo3 by RT-qPCR and WB analysis (Figures S4A and S4B). Moreover, Fbxo3 deficiency during disease progression significantly prolonged survival, reduced leukemia burden in both BM and SP, and alleviated splenomegaly (Figures 3G–3K and S4C). We also assessed the effect of Fbxo3 on LSCs14^,^16^,^36^,^37^,^38^,^39^,^40^,^41 and found that compared to control CML mice, Fbxo3-deficient mice exhibited a reduction in leukemia stem/progenitor cells (LSCs/LSPCs), including GFP^+^ Lin^−^Sca-1^+^c-Kit^+^ (LSK) cells, GFP^+^ long-term HSCs (LT-HSCs), GFP^+^ short-term HSCs (ST-HSCs), GFP^+^ granulocyte-macrophage progenitors (GMPs), and GFP^+^ common myeloid progenitors (CMPs) in both BM and SP (Figures 3L–3R). Importantly, Fbxo3 deficiency had no significant effect on normal hematopoiesis (Figures S4D and S4E). Moreover, in vitro colony-forming cell (CFC) and serial replating assays using sorted GFP^+^ c-kit^+^ cells demonstrated that Fbxo3 deficiency impaired the self-renewal capacity of LSCs (Figure 3S). In conclusion, FBXO3 deficiency suppressed CML progression and improved survival.Figure 3. Depletion of FBXO3 suppressed the growth of CML cells and diminished survival and self-renewal of primary CML CD34^+^ cells in vivo(A) CML-CDX model workflow: NPG mice (n = 5) injected with control/FBXO3-KO K562 cells (1 × 10^7^/mouse); leukemia burden and survival assessed at 6 weeks(B) The proportion of human CD45^+^ cells in PB and BM was measured by flow cytometry. Left: typical flow scatterplot. Right: statistical analysis chart of the proportion of hCD45^+^ cells. Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗p < 0.001 (n = 5).(C) Photos and H&E staining of spleen of two groups of mice. Scale bars, 20 μm.(D) Statistical analysis chart of spleen weight and relative spleen weight. Data are represented as mean ± SEM; t test; ∗p < 0.05.(E) Statistical charts of survival time of two groups of mice. Gehan-Breslow-Wilcoxon test;∗∗p < 0.01 (n = 5).(F) A flow chart for the construction of a BCR::ABL1-induced CML mouse model.(G) Percentage of GFP^+^ cells in the PB of CML mice detected by flow cytometry. Left: typical flow contour plot. Right: graph of statistical analysis of the percentage of GFP^+^ cells in PB. Data are represented as mean ± SEM; t test; ∗∗∗p < 0.001 (n = 5).(H) Survival analysis post-leukemia onset. Gehan-Breslow-Wilcoxon test, ∗∗p < 0.01 (n = 5).(I) Spleen pathology comparison. Left: photographs of typical spleens. Right: graph of statistical analysis of spleen weight and relative spleen weight. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001 (n = 5).(J) H&E staining of spleen from the two groups of mice. Scale bars, 20 μm.(K) Proportion of BM leukemia cells in CML mice. Left: typical flow pseudo-color map. Right: statistically analyzed graph. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001 (n = 5).(L) Proportion of BM LSK cells in CML mice. Left: typical flow pseudo-color map. Right: statistically analyzed graph. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001 (n = 5).(M) Proportion of total HSCs, LT-HSCs, and ST-HSCs in the BM of CML mice. Left: typical flow pseudo-color map. Right: statistically analyzed graph. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗∗∗p < 0.0001 (n = 5).(N) Proportion of BM GMPs and CMPs in CML mice. Left: typical flow pseudo-color map. Right: statistically analyzed graph. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001 (n = 5).(O–R) Plot of statistical analysis of the proportion of splenic GFP^+^ cells (O), LSK cells (P), HSC subpopulations (Q), GMPs (R), and CMPs (R) in the two groups of mice. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance (n = 5).(S) GFP^+^ c-kit^+^ cells (5,000 cells/well) sorted from sh-control or Sh-FBXO3 CML mice (n = 3) were seeded in MethoCult M3434 methylcellulose medium for three rounds of CFC/replating assay). Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗p < 0.001.
FBXO3 regulates DUSP9 in a ubiquitination-dependent manner in CML
To elucidate the molecular mechanism of FBXO3 in promoting CML progression, an unbiased liquid chromatography-mass spectrometry (LC-MS) screen in K562 cells was conducted and identified dual-specificity phosphatase 9 (DUSP9) as a prominent candidate for interaction with FBXO3 (Figure 4A; Table S6). Reciprocal co-immunoprecipitation assays in imatinib-sensitive/-resistant cell lines and primary CD34^+^ cells in both exogenous and endogenous contexts confirmed the interaction between DUSP9 and FBXO3 in CML cells (Figures 4B and 4C). As FBXO3 functions as an E3 ubiquitin ligase, we further investigated whether DUSP9 is the ubiquitination substrate of FBXO3. We treated CML cells with the 26S proteasome inhibitor MG132, which inhibits the UPS,42^,^43 and observed an accumulation of endogenous DUSP9 (Figures 4D and S5A), suggesting the regulation of DUSP9 by the UPS. Additionally, FBXO3 knockdown increased DUSP9 protein levels, whereas FBXO3 overexpression decreased DUSP expression (Figures 4E, 4F, S5B, and S5C). FBXO3 knockdown also extended the half-life of endogenous DUSP9 (Figure 4G). An in vivo ubiquitination assay further demonstrated that FBXO3 deficiency markedly impaired DUSP9 polyubiquitination (Figure 4H), with no significant changes of DUSP9 mRNA levels (Figures S5D–S5F). Moreover, the mRNA expressions of DUSP9 in CML CD34^+^ cells, as well as CML cell lines showed no significant differences compared to those control groups (Figures S6A–S6C). Strikingly, the protein levels of DUSP9 in CML CD34^+^ cells, as well as CML cell lines significantly decreased compared to those control groups (Figures 4I, 4J, S6D, and S6E). Furthermore, we found that the protein levels of DUSP9 were negatively correlated with FBXO3 expression and TKI treatment failure, suggesting the potential ubiquitination degradation by FBXO3(Figures 4K and 4L). Consistently, CML mice also showed downregulated protein level of DUSP9 compared to WT mice (Figures S6F–S6G). These findings demonstrate that FBXO3 regulates DUSP9 in a ubiquitination-dependent manner in CML.Figure 4FBXO3 binds to and promote ubiquitination of DUSP9 via the ApaG domain(A) Tandem affinity purification and mass spectrometry screening of potential interacting proteins of FBXO3.(B) The interaction of exogenous FBXO3 with endogenous DUSP9 in IM-sensitive cell lines (K562 and KBM5) and IM-resistant cell lines (KBM5-T315I).(C) Detection of endogenous FBXO3 interaction with endogenous DUSP9 in CML cell lines and CML CD34^+^ cells.(D) WB analysis showing DUSP9 degradation via proteasome (MG132 vs. DMSO treatment) in CML cell lines and CML CD34^+^ cells.(E) Lentiviruses targeting FBXO3 were added to CML cell lines and CML CD34^+^ cells, and cells were collected and then subjected to western blot experiments to detect the protein expression levels of FBXO3 and DUSP9.(F) SFB-FBXO3 plasmids were transfected into CML cell lines, and cells were collected and then subjected to western blot experiments to detect the protein expression levels of FBXO3 and DUSP9.(G) The expression level of DUSP9 was detected by western blot in control and FBXO3-knockdown K562 cells treated with CHX for 9, 6, 3, 1.5, and 0 h. Left: graph of typical western blot results. Right: three independent replicate experiments were performed, and the half-life curves were plotted after analyzing the gray values of the bands for statistical analysis.(H) The corresponding plasmids were transfected in control and FBXO3 knockdown K562 cell lines, and western blot experiments were performed to detect the ubiquitination level of DUSP9.(I–K) Elevated DUSP9 in CML BM CD34^+^ cells (N = 28, optimal response, N = 9; sub-optimal response, N = 7; TKI treatment failure, N = 12) vs. normal (N = 8). Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(L) Correlation analysis (Spearman’s method) between FBXO3 and DUSP9 with correlation coefficients R and p values indicated.(M) Schematic diagram of the FBXO3 structural domain deletion mutant.(N) Overexpression of different FBXO3 structural domain deletion mutants in K562 cells and KBM5-T315I cells for Co-IP assay to detect their interaction with DUSP9.(O) Vector, Flag-FBXO3, and Flag-FBXO3-ΔApaG plasmids were transfected respectively in CML cell lines, and western blot experiments were performed to detect DUSP9 protein expression. Left: typical western blot results. Right: three independent replicate experiments were performed, and the bands were analyzed for gray value and then statistically analyzed. t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.(P) The corresponding plasmids were overexpressed in K562 cells. CHX was added at the corresponding time, and the protein was extracted and then subjected to western blot to detect the expression level of DUSP9.(Q) The corresponding plasmids were overexpressed in the K562 cell line, respectively, and the ubiquitination level of DUSP9 was detected by Co-IP assay using Myc beads after protein extraction.
FBXO3 interacts with DUSP9, leading to its ubiquitination and degradation via the FBXO3 ApaG domain
We further mapped the DUSP9-binding domain in FBXO3 using a series of deletion mutants, including ΔF-box, Δ57-121, ΔSUKH, ΔApaG, and Δ409-471 (Figure 4M). Interestingly, except for ΔApaG that was unable to bind DUSP9, other mutant proteins bound to DUSP9 similar to that of wild-type FBXO3 (Figure 4N). Notably, wild-type FBXO3, but not FBXO3ΔApaG mutant, had significant effects on degradation of DUSP9 (Figure 4O). Consistently, cycloheximide chase assays confirmed that intact FBXO3, but not the ApaG-deleted mutant, destabilized DUSP9 (Figure 4P). Moreover, in vivo ubiquitination assays revealed that only intact FBXO3, but not the ΔApaG or ΔF-box mutants, facilitated DUSP9 ubiquitination (Figure 4Q). These results indicate that ΔApaG in the FBXO3 domain is indispensable for its interaction with DUSP9.
FBXO3-mediated DUSP9 degradation promotes CML cell proliferation and survival
To further elucidate the functional implications of FBXO3-mediated DUSP9 degradation in CML, we overexpressed DUSP9 in both imatinib-sensitive and -resistant cell lines (Figure 5A). Consistent with FBXO3 deficiency phenotypes in CML, DUSP9 overexpression also inhibited cell proliferation, reduced colony formation, enhanced apoptosis, and activated the endogenous apoptotic cascade in both K562- and T315I-mutant KBM5 cells (Figures 5B–5F). Furthermore, we performed DUSP9 and/or FBXO3 knockdown in K562 cells (Figure 5G) and found that DUSP9 knockdown counteracted the phenotypic defects caused by FBXO3 depletion, restoring cell proliferation, enhancing colony formation, reducing apoptosis, and inactivating the endogenous apoptotic cascade in FBXO3-knockdown K562 cells (Figures 5H–5L). These findings were recapitulated in primary CML CD34^+^ cells (Figures 5M–5P).Figure 5FBXO3-mediated destruction of DUSP9 determines the role of FBXO3 in CML in vitro(A) DUSP9 overexpression validation in K562/KBM5-T315I by western blot assay.(B) Growth curves of DUSP9-overexpressing cells (1 × 10^4^ seeded; 7-day quantification; n = 3). Data are represented as mean ± SD; two-way ANOVA; ∗∗p < 0.01, ∗∗∗p < 0.001.(C) Soft agar colony formation of control and DUSP9-overexpressing cells; scale bars, 100 μm. The number of clones were counted (more than 50 cells are regarded as one cell clone). Data are represented as mean ± SEM; t test; ∗∗p < 0.01.(D) Apoptosis of control and DUSP9-overexpressing cells was detected using annexin Ⅴ/DAPI apoptosis assay. Data are represented as mean ± SEM; t test; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.(E) The mitochondrial membrane potential of control and DUSP9-overexpressing cells was measured using JC-10 dye. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001 (n = 3).(F) Western blot analysis of apoptosis-related proteins in control and DUSP9-overexpressing cells. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns indicates no significance, n = 3.(G) The K562 cell lines that knock down FBXO3, DUSP9, or both were constructed respectively, and western blot assays were performed to detect the knockdown effects.(H) Growth curves of four groups K562 cells (n = 3). Two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01.(I) Soft agar colony formation of four groups K562 cells; scale bars, 100 μm. The number of clones were counted. Data are represented as mean ± SEM; t test; ∗∗p < 0.01; ns indicates no significance.(J) Apoptosis of the cells of rescue experiments was detected using annexin Ⅴ/DAPI apoptosis assay. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001; ns indicates no significance (n = 3).(K) The mitochondrial membrane potential of the cells of rescue experiments was measured using JC-10 dye. Data are represented as mean ± SEM; t test, ∗∗p < 0.01; ns indicates no significance (n = 3).(L) Western blot analysis of apoptosis-related proteins in the cells of rescue experiments. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns indicates no significance (n = 3).(M) The CML CD34^+^ cells that knock down FBXO3, DUSP9, or both were constructed respectively, and western blot assays were performed to detect the knockdown effects.(N) Soft agar colony formation of four groups CML CD34^+^ cells; scale bars, 100 μm. The number of clones were counted. Data are represented as mean ± SEM; t test, ∗∗p < 0.01; ns indicates no significance (n = 3).(O) Apoptosis of the cells of four groups of CML CD34^+^ cells was detected using annexin Ⅴ/DAPI apoptosis assay. Data are represented as mean ± SEM; t test; ∗∗∗∗p < 0.0001; ns indicates no significance (n = 3).(P) According to the method of primary cell colony formation experiment, three successive generations of culture were carried out, and the number of colonies with cell numbers greater than 50 were counted respectively. Three independent replication experiments were carried out for statistical analysis and t test. Data are represented as mean ± SEM; ∗p < 0.05, ∗∗p < 0.01; ns indicates no significance.
Based on our finding that the ApaG domain of FBXO3 mediates its interaction with DUSP9, we respectively overexpressed Flag-FBXO3 or Flag-FBXO3-ΔApaG plasmids in FBXO3-knockdown K562 cells. Our results indicated that overexpressing Flag-FBXO3 reverses the phenotypic defects caused by FBXO3 knockdown while exogenous expression of Flag-FBXO3-ΔApaG had no effects (Figures S7A–S7E). We further mapped the FBXO3-binding domain in DUSP9 using a series of deletion mutants, including Δ18-139, Δ140-202, ΔTyrosine-protein phosphatase, and Δ348-384 (Figure S7F). Interestingly, co-immunoprecipitation assays revealed that only Δ348–384 abolished FBXO3 binding, indicating that this region is essential for the DUSP9 ubiquitination (Figure S7G). Next, we compared the role of Flag-DUSP9 and Flag-DUSP9-Δ348-384 in CML. K562 cells transfected with Flag-DUSP9-Δ348-384 had a slower growth rate, fewer cell clones, a higher proportion of apoptotic cells, and a greater change in mitochondrial membrane potential (Figures S7H–S7K). A possible reason is that the overexpressed Flag-DUSP9-Δ348-384 cannot be degraded by FBXO3 in vivo, thus triggering a stronger biological effect than the wild-type Flag-DUSP9. Taken together, these in vitro results revealed that FBXO3 played its role in CML via mediating the ubiquitination of DUSP9.
DUSP9 silencing reverses the elimination of LSCs in FBXO3-deficient CML mice
Lentiviral shRNAs targeting FBXO3, DUSP9, or both were transduced into BM and SP cells derived from first-generation CML mice to determine whether FBXO3 modulates CML-LSCs via DUSP9 (Figure 6A). Knockdown efficiency was confirmed by RT-qPCR and WB analysis (Figures S8A and S8B). DUSP9 deficiency alone resulted in an increased leukemia burden (Figure 6B). Moreover, DUSP9 deficiency abolished the reduction in leukemia burden induced by FBXO3 knockdown and counteracted the extended survival, alleviated splenomegaly, and reduced tumor cell infiltration induced by FBXO3 deficiency (Figures 6C, 6D, and S8C). The FBXO3-knockdown-induced reduction in GFP^+^ cells in BM and SP was also reversed by simultaneous DUSP9 knockdown (Figure 6E). Flow cytometry revealed that DUSP9 knockdown in FBXO3-deficient BM and SP cells increased proportions of GFP^+^ LSK cells, GFP^+^ ST-HSCs, and GFP^+^ HSCs compared to FBXO3 knockdown alone (Figures 6F and 6G). In addition, FBXO3 knockdown reduced GFP^+^ GMPs and GFP^+^ CMPs, while DUSP9 knockdown rescued these populations (Figure 6H). Notably, normal hematopoietic cell populations were not significantly affected under conditions of either FBXO3 or DUSP9 knockdown (Figures S8D and S8E). In vitro CFC and serial replating assays with sorted GFP^+^ c-kit^+^ cells further demonstrated that DUSP9 knockdown rescued the impaired serial plating capacity induced by FBXO3 knockdown (Figure S8F). These data suggest that DUSP9 silencing counteracts the effect of FBXO3 deficiency on LSCs and leukemia burden in CML.Figure 6DUSP9 knockdown partially reverses the FBXO3-depletion-mediated elimination of LSCs in CML mice(A) A flow chart for the construction of BCR::ABL1-induced CML mouse model for rescue experiments.(B) Percentage of GFP^+^ cells in the PB of CML mice detected by flow cytometry. Left: typical flow contour plot. Right: graph of statistical analysis of the percentage of GFP^+^ cells in PB. Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗∗p < 0.0001 (n = 5).(C) After the onset of the disease in mice, the survival time (days) of each mouse in each group was started to be recorded and survival curves were plotted, Gehan-Breslow-Wilcoxon test; ∗∗p < 0.01; ns indicates no significance (n = 5).(D) CML mice and normal C57 mice of equal weekly age were dissected, and spleens were photographed and weighed. Scale bars, 20 μm. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 (n = 5).(E) Proportion of BM and spleen leukemia cells in CML mice in sh-control, sh-FBXO3 groups, sh-DUSP9 groups, and sh-FBXO3+sh-DUSP9 groups. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 (n = 5).(F) Proportion of BM and spleen LSK cells in CML mice in four groups. Data are represented as mean ± SEM; t test; ∗∗p < 0.01, ∗∗∗∗p < 0.0001 (n = 5).(G) Proportion of total HSCs, LT-HSCs, and ST-HSCs in CML mice (BM and spleen) in four groups. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance (n = 5).(H) Proportion of GMPs and CMPs in CML mice (BM and spleen) in four groups. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance (n = 5).(I) Bubble map of differential gene signaling pathway enrichment analysis. p < 0.05 was considered to be significantly enriched.(J) Column chart of GSVA analysis.(K and L) FBXO3, DUSP9, or both were knocked down separately at the same time in K562 cells and CML CD34^+^ cells. Western blot experiment was performed to detect the expression level of genes related to the MAPK signaling pathway. (K) Typical western blot results. (L) Three independent replication experiments were performed, and the gray release values of the bands were analyzed for statistical analysis. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01; ns indicates no significance.
DUSP9 deficiency activates MAPK pathway in FBXO3-deficient CML cells
To uncover the mechanism by which DUSP9 regulates CML cells, we performed RNA-seq analysis in DUSP9-knockdown K562 cells. Our results showed that MAPK signaling was significantly upregulated (Figures S9A–S9C, 6I, and 6J). Previous studies have reported that DUSP9 participates in diverse biological processes, such as metabolism, cell-cycle regulation, apoptosis, drug sensitivity, and cell stemness, by regulating the MAPK signaling pathway.44 Thus, we hypothesized that DUSP9 may also exert its functions in CML by regulating the activity of MAPK signaling. The results further showed that loss of DUSP9 markedly increased the phosphorylation of p38 and c-Jun N-terminal kinase (JNK) in both K562 cells and primary CML CD34^+^ cells. Conversely, FBXO3 knockdown suppressed these phosphorylation events, whereas simultaneous DUSP9 deficiency reversed this effect (Figures 6K–6L). An investigation of upstream signaling events revealed no alterations in the activation status of ASK1, MKK3/6, or MKK4/7 (Figures S9D–S9E),45 which suggested that DUSP9 may primarily exert its effects through dephosphorylation of p38 and JNK. Previous research has demonstrated that maintaining a delicate equilibrium between pro-survival BCL family members and pro-apoptotic mediators, such as Bim, is crucial for regulating apoptosis and cell survival. Additionally, the MAPK pathway has been shown to modulate Bim activity through phosphorylation.46^,^47^,^48 In light of this, we investigated whether FBXO3 or DUSP9 influences the levels of both Bim protein and its phosphorylated form. Our findings revealed no significant differences in the expression levels of Bim protein and phosphorylated Bim protein (Figures S9F–S9G). These results suggest that DUSP9 modulates CML cells via the MAPK pathway.
Pharmacological inhibition of FBXO3 selectively eliminates CML cells in vitro and disrupts the interaction between FBXO3 and DUSP9
We next investigated the therapeutic effects of BC1215, a small molecule inhibitor targeting FBXO3 ApaG domain49 to treat K562, KBM5, and KBM5-T315I cells and found that the half-maximal inhibitory concentration (IC50) of BC1215 was in a dose- and time-dependent manner (Figure S10A). BC1215 also inhibited cell proliferation, reduced colony formation, induced apoptosis, and activated the intrinsic apoptotic pathway in these cell lines (Figure S10B–S10I). Moreover, BC1215 significantly inhibited cell growth, reduced colony formation, and increased apoptosis in primary CML CD34^+^ cells, whereas normal CD34^+^ cells exhibited only modest growth and colony formation inhibition at higher doses (Figures 7A–7C), indicating preferential toxicity toward leukemic cells.Figure 7BC1215 alone or in combination with IM eliminates LSCs and prolongs survival of CML mice(A) Soft agar colony formation of the CML and normal CD34^+^ cells (after BC1215 treatment); scale bars, 100 μm. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01; ns indicates no significance.(B) According to the method of primary cell colony formation experiment, three successive generations of culture were carried out, and the number of colonies with cell numbers greater than 50 were counted respectively. Data are represented as mean ± SEM; t test; ∗p < 0.05; ns indicates no significance (n = 3).(C) Apoptosis of CML CD34^+^ cells added with BC1215 was detected using annexin Ⅴ/DAPI double staining flow cytometry. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (n = 3).(D) CML and normal CD34^+^ cells were added with IM, BC1215, or IM + BC1215 for 24 h. Apoptosis was detected using annexin Ⅴ/DAPI double staining flow cytometry. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗∗p < 0.001; ns indicates no significance (n = 3).(E) A flow chart for the construction of BCR::ABL1-induced CML mouse model for drug experiments.(F) Western blot experiment was performed to test the expression level of FBXO3 and DUSP9 in BM cells of four groups.(G) Percentage of GFP^+^ cells in the PB of CML mice detected by flow cytometry. Data are represented as mean ± SEM; t test; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance (n = 5).(H) CML mice and normal C57 mice of equal weekly age were dissected and spleens were photographed and weighed; scale bars, 20 μm. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 (n = 5).(I) After six drug injections, the survival time (days) of each mouse in each group was started to be recorded and survival curves were plotted; Gehan-Breslow-Wilcoxon test; ∗p < 0.05, ∗∗p < 0.01 (n = 5).(J–M) Proportion of BM leukemia cells (J), LSK cells (K), HSC subpopulation (L), GMPs (M), and CMPs (M) in CML mice in control, IM, BC1215, and IM + BC1215 groups. Data are represented as mean ± SEM; t test; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns indicates no significance (n = 5).
We have demonstrated that FBXO3 interacts with DUSP9 through its ApaG domain. Furthermore, BC1215 also disrupted this binding effectively in a dose-dependent manner (Figure S11A). Treatment with BC1215 led to a significant, time- and dose-dependent upregulation of DUSP9 in both imatinib-sensitive and -resistant cells, accompanied by prolonged DUSP9 half-life and reduced ubiquitination (Figures S11B–S11D). Downstream signaling analyses revealed significant suppression of p38 and JNK phosphorylation in the MAPK pathway (Figure S11E). These data suggest that BC1215 disrupts the FBXO3-DUSP9 interaction, highlighting its potential as a targeted therapeutic agent in CML.
BC1215 alone or in combination with IM exerts remarkable toxicity on CML CD34+ cells in vitro and in vivo
To assess whether inhibiting FBXO3 could improve the efficacy of imatinib in CML, we treated primary CML CD34^+^ and normal CD34^+^ cells with imatinib (IM), BC1215, or their combination. Notably, BC1215 monotherapy induced significantly higher apoptosis in CML CD34^+^ cells relative to imatinib alone, whereas the combination treatment resulted in the highest levels of apoptosis (Figure 7D). However, normal CD34^+^ cells were minimally affected. These data establish BC1215 as a potential synergistic partner for IM in overcoming therapeutic resistance, while preserving normal HSC/progenitor cell function.
To evaluate the efficacy of FBXO3 inhibition in vivo, we treated BCR::ABL1-driven CML mice with vehicle, IM, BC1215, or IM plus BC1215 (Figure 7E). IM monotherapy and IM plus BC1215 both reduced FBXO3 expression, whereas BC1215 alone had no significant effect (Figure 7F). In addition, DUSP9 expression was elevated after treatment with IM and BC1215 monotherapy, as well as IM plus BC1215 combination therapy (Figure 7F). Subsequent analyses revealed that BC1215 monotherapy and IM plus BC1215 combination significantly suppressed the proliferation of GFP^+^ tumor cells and reduced splenomegaly (Figures 7G, 7H, and S12A). Histopathological analysis revealed that the IM plus BC1215 group had the least splenic tumor infiltration and that BC1215 monotherapy substantially reduced infiltration as well (Figure 7H). Survival analysis showed that both BC1215 alone and in combination with IM extended survival compared to vehicle- and IM-treated mice (Figure 7I). Flow cytometric analysis of LSCs/LSPCs revealed that IM plus BC1215 significantly reduced GFP^+^ cells in BM and SP (Figures 7J and S12B). BC1215 monotherapy also decreased GFP^+^ LSK cells, ST-HSCs, GMPs, and CMPs in BM and SP (Figures 7K–7M, and S11C–S11E), while leaving normal GFP^−^ HSC/HSPCs unaffected (Figures S12F–S12G). These findings suggest that targeting FBXO3 effectively eradicates LSCs and extends survival in CML, while preserving normal hematopoiesis.
BC1215 demonstrates minimal toxicity to normal hematopoiesis in mice
To evaluate the potential cytotoxicity of BC1215 on normal hematopoiesis, adult C57BL/6 mice were treated with BC1215 or vehicle for 2 weeks using the same dosing regimen as in CML mice. Our results showed no significant differences in body weight, spleen weight, or PB cell counts including red blood cells (RBCs), white blood cells (WBCs), platelets (PLTs), and hemoglobin (HGB) (Figures S13A–S13C). Flow cytometric analysis of PB revealed that BC1215 treatment had no effect on the percentages of T cells, B cells, natural killer (NK) cells, and myeloid cells. Moreover, it also had no effect on HSCs/progenitor cells, megakaryocytes, macrophages, monocytes, and burst-forming unit-erythroid (BFU-E)/colony-forming unit-erythroid (CFU-E) in BM (Figure S13D). These results suggest that BC1215 is well tolerated with minimal toxicity to normal hematopoiesis in adult mice.
Discussion
LSCs are a small subset of cells with stemness characteristics within the leukemia cell population, possessing three core properties: “self-renewal capacity,” “multilineage differentiation potential,” and “chemoresistance.” They are one of the fundamental causes of leukemia initiation, recurrence, and drug resistance.50^,^51^,^52^,^53^,^54 Although the advent of TKIs has transformed the therapeutic landscape for patients with CML, the persistence of LSCs leads to TKI resistance in a subset of patients, resulting in disease recurrence and ultimately therapeutic failure. Therefore, developing strategies that target LSCs while sparing normal HSCs is a critical unmet need in CML. Recent studies have implicated FBXO3 in promoting stemness and metastasis in solid tumors and hematological malignancies, including breast cancer,55 non-small cell lung cancer,56 and acute promyelocytic leukemia.57 In CML, quiescent LSCs evade TKI-mediated eradication, thus contributing to disease relapse.20^,^58 However, comprehensively characterizing CML-LSCs and elucidating the mechanisms driving their resistance to TKIs remain challenging, primarily due to the difficulty of distinguishing quiescent LSCs from normal HSCs. Here, we demonstrate that FBXO3 serves as an innovative marker of CML-LSCs, as evidenced by enhanced stemness and quiescence signatures as well as enrichment of several previously characterized LSC-associated genes in FBXO3^+^ LSCs at the single-cell level.32^,^33^,^34 Inhibition of FBXO3 enhances the clearance of CML-LSCs through DUSP9-mediated inactivation of the MAPK signaling pathway, which is critical for sustaining LSC self-renewal, quiescence, and resistance to TKIs. Importantly, this mechanism offers a promising strategy to overcome TKI resistance by selectively eradicating LSCs, while preserving normal hematopoiesis. These findings establish FBXO3 as a promising therapeutic target for CML and provide a rationale for clinical exploration of FBXO3 inhibitors as alone agents or in combination with existing TKIs to achieve deeper, more durable remissions.
Approximately 72 F-box proteins have been identified in mammals, each modulating diverse cellular pathways through recognizing specific substrates.59 Of these F-box proteins, only FBXW7 has been suggested as potential target for CML-LSC clearance.60^,^61 Nevertheless, other F-box proteins also could become potential CML-LSC target. Among them, we revealed that FBXO3 promotes CML-LSC survival but has had little to no effect on normal hematopoiesis. FBXO3 yielded this effect through ubiquitin-dependent degradation of DUSP9, a dual-specificity phosphatase that inactivates MAPK signaling by dephosphorylating p38 and JNK.44^,^62 Functional analyses revealed that DUSP9 overexpression inhibits CML cell proliferation and clonogenicity, while its knockdown modestly promotes these properties via MAPK signaling activation. Conversely, DUSP9 knockdown restores the survival of CML-LSCs in FBXO3-deficient cells both in vitro and in vivo, confirming the dependency of FBXO3’s oncogenic function on DUSP9 degradation. These results are consistent with previous work linking MAPK activation to tumor stem cell maintenance in CML and other cancers.63 Reduced DUSP9 expression has also been observed in other tumor types and correlates with increased MAPK activity, which in turn promotes cellular proliferation, differentiation, and therapeutic resistance.64^,^65^,^66 These data support a central role for the FBXO3-DUSP9 axis in supporting LSC survival and TKI resistance in CML.
Current TKI therapies have effectively improved the prognosis of CML patients67^,^68^,^69 but fail to eliminate quiescent residual CML-LSCs, limiting the potential for achieving durable treatment-free remission (TFR).70^,^71 Our in vivo data demonstrate that BC1215 monotherapy or combination with imatinib impairs the self-renewal capacity of CML-LSCs and effectively eradicates them in mouse models, reducing residual disease and prolonging survival. Notably, the combination regimen induces more profound LSC elimination than either agent alone, suggesting synergistic activity. The impaired self-renewal capacity leads to a progressive depletion of the residual LSC pool after treatment discontinuation. Critically, BC1215 spares normal hematopoiesis in mice, underscoring its therapeutic potential in further CML clinical trials to eliminate LSCs, thereby reversing TKI resistance and preventing relapse upon TKI discontinuation.
In conclusion, our study identifies FBXO3 as a critical regulator of CML-LSC maintenance and TKI resistance. Mechanistically, FBXO3 promotes ubiquitin-mediated degradation of DUSP9, thereby activating the MAPK signaling to sustain LSC stemness and quiescence. Both genetic deletion and pharmacological inhibition of FBXO3 via BC1215 effectively eliminate CML-LSCs while sparing normal hematopoiesis. These findings establish FBXO3 as a validated therapeutic target and highlight the potential of combining FBXO3 inhibitors with TKIs to eliminate residual LSCs, paving the way for improved strategies to achieve durable treatment-free remission and cure in CML.
Limitations of the study
Several limitations of this study should be considered. The upstream regulatory mechanisms governing both the expression and activity of FBXO3 in quiescent CML-LSCs remain to be elucidated. Understanding how FBXO3 is selectively upregulated in CML-LSCs versus normal HSCs is critical to grasping its role in CML. The translation of these findings to the clinic will depend on conducting clinical trials to validate the safety and efficacy of FBXO3 inhibitor.
Resource availability
Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Wei Li ([email protected]).
Materials availability
This study did not generate new unique reagents.
Data and code availability
RNA-seq data have been deposited at Gene Expression Omnibus (GEO) database and are accessible through GEO: GSE309070. The accession number is also listed in the key resources table. The mass spectrometry proteomics data have been deposited to the PRIDE repository with the dataset identifier PXD073522. The full-length, unprocessed western blots are included in the paper in Data S1.
This paper does not report original code.
Any additional information or code required to reanalyze the data reported in this work paper is available from the lead contact upon request.
Acknowledgments
This study was supported by the 10.13039/501100001809National Natural Science Foundation of China (82270149 [W.L.], 82370144 [Y.S.], 82170211 [Z.B.], 82470217 [Z.B.], 82370219 [W.C.], 82200190 [N.S.], 82400149 [M. Li], and 82400233 [B.Z.]), The National 10.13039/501100006407Natural Science Foundation of Henan Province (242300421080), The Leading Talent Project of Henan Province (nos. LJRC2023004 and LJRC2024015), a 10.13039/501100002858China Postdoctoral Science Foundation Funded Project (2024M752955), Category C of 10.13039/501100002858China Postdoctoral Science Foundation Funding (GZC20232411 and GZC20232429), the Henan Provincial Joint Construction Project of Medical Science and Technology Research Program (LHGJ20240207), Key Research and Development Projects of the Ministry of Science and Technology (2024YFA1108303), and Funding for the Scientific Research and Innovation Team of the First Affiliated Hospital of Zhengzhou University (QNCXTD2023003). We thank Jingxuan Pan from Sun Yat-sen University for his donation of the KBM5 and the KBM5-T315I cell lines. The graphical material in this article was supplied by FigDraw for free.
All experiments involving human samples were conducted in compliance with all relevant ethical regulations and were approved by the ethics committees of First Affiliated Hospital of Zhengzhou Universities (2021-KY-0575-002).
All animal experiments were performed according to the protocols approved by the Animal Care and Use Committee of Henan China Model Technology Research Institute (202501002).
Author contributions
X.L., S.Z., and W.L. conceived the project, designed experiments, and wrote the manuscript; X.L. performed in vitro functions, mechanism studies, and mouse model studies; S.Z. performed plasmid construction, some mechanism studies, and flow-cytometry-related experiments; Z.L. performed bioinformatic analyses, with conceptual input from W.L.; Yanli Zhang and N.S. provided CML patient samples; and other researchers in the lab (B.Z., M. Li, H.H., M.G., Z.H., M.S., H.Z., Y.C., F.G., W.F., Z.Z., Y.H., Y.A., S.L., M. Liu, YupengLiu, Yuxuan Liu, C.L., Yiguo Zhang, Yingmei Li, W.C., and F.W.) helped with the experiments. W.L., Z.B., L.X., and Y.S. supervised the overall study.
Declaration of interests
The authors declare no competing interests.
STAR★Methods
Key resources table
REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesFBXO3(CO-IP)Santa CruzCat# sc-514625; RRID: AB_3713047GAPDHServicebioCat#GB12002; RRID: AB_3206256FBXO3(WB)ProteintechCat#17803-1-AP; RRID: AB_2278445PARP1ProteintechCat#13371-1-AP; RRID: AB_2160459Caspase-3ProteintechCat#19677-1-AP; RRID: AB_10733244Cleaved caspase-3Cell Signaling TechnologyCat##9661; RRID: AB_2341188BCL-xLProteintechCat#26967-1-AP; RRID: AB_2880702BCL-2ProteintechCat#12789-1-AP; RRID: AB_2227948MCL-1ProteintechCat#16225-1-AP; RRID: AB_2143977BAXProteintechCat#50599-2-Ig; RRID: AB_2061561BADProteintechCat#10435-1-AP; RRID: AB_2061994XIAPProteintechCat#10037-1-Ig; RRID: AB_2215009SurvivinProteintechCat#10508-1-AP; RRID: AB_2064048Cytochrome cProteintechCat#10993-1-AP; RRID: AB_2090467DUSP9ABclonalCat#A3839; RRID: AB_2765337HAProteintechCat#51064-2-AP; RRID: AB_11042321FLAGABclonalCat#AE005; RRID: AB_2770401MYCproteintechCat#60003-2-Ig; RRID: AB_2734122Mouse mAb IgG1 Isotype controlCell Signaling TechnologyCat##5415; RRID: AB_10829607p44/42 MAPK (Erk1/2)Cell Signaling TechnologyCat##4695; RRID: AB_390779Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204)Cell Signaling TechnologyCat##4370; RRID: AB_2315112p38Cell Signaling TechnologyCat##9212; RRID: AB_330713Phospho-p38 MAPK (Thr180/Tyr182)Cell Signaling TechnologyCat##4511; RRID: AB_2139682JNKCell Signaling TechnologyCat##9252; RRID: AB_2250373Phospho-JNK1-T183/Y185+JNK2-T183/Y185+JNK3-T221/Y223ABclonalCat#AP1337; RRID: AB_3697809ASK1ABclonalCat#A12458; RRID: AB_2759301Phospho-ASK1-S1033ABclonalCat#AP0058; RRID: AB_2771298MKK3/MKK6ABclonalCat#A19830; RRID: AB_2895227Phospho-MKK3-S189+MKK6-S207ABclonalCat#AP1512MKK4ABclonalCat#A14781; RRID: AB_2861707Phospho-MKK4-S257/T261ABclonalCat#AP0541; RRID: AB_2771289MKK7ABclonalCat#A12950; RRID: AB_2861679Phospho-MKK7 (Ser271/Thr275) Polyclonal antibodyProteintechCat#29199-1-AP; RRID: AB_2918246BIM/BCL2L11 Polyclonal antibodyProteintechCat#22037-1-AP; RRID: AB_2878978Phospho-Bim (Ser69) (D7E11) Rabbit mAbCell Signaling TechnologyCat#4585; RRID: AB_2065176Phospho-Bim (Ser77) (D4H12) Rabbit mAbCell Signaling TechnologyCat#12433; RRID: AB_2797911PE-Cy7 CD34BioLegendCat#128618; RRID: AB_2721678BV421 lineageBioLegendCat#133311; RRID: AB_11203535APC-Cy7 Ly6A/E(Sca-1)BioLegendCat#108126; RRID: AB_10645327APC CD117(c-Kit)BioLegendCat#105812; RRID: AB_313221PE CD135BioLegendCat#135306; RRID: AB_1877217BV605 CD48BioLegendCat#103441; RRID: AB_2650825PerCP/Cy5.5 CD150(SLAM)BioLegendCat#115921; RRID: AB_2206887BV510 CD16/32BioLegendCat#156625; RRID: AB_3083139BV711 CD127(IL-7Ra)BioLegendCat#135035; RRID: AB_2564577BV421 CD41BioLegendCat#133912; RRID: AB_2650893PE-Cy7 CD34BioLegendCat#128618; RRID: AB_2721678PE-Cy7 CD16/32BioLegendCat#156610; RRID: AB_2800708APC-Cy7 Scal1BioLegendCat#108126; RRID: AB_10645327APC CD117BioLegendCat#105812; RRID: AB_313221Percp Cy5.5 CD71BioLegendCat#113816; RRID: AB_2565482BV421 CD3BioLegendCat#100227; RRID: AB_10900227BV421 CD11bBioLegendCat#101235; RRID: AB_10897942BV421 CD11cBioLegendCat#117329; RRID: AB_10897814BV421 CD19BioLegendCat#115537; RRID: AB_10897814BV421 Ly6GBioLegendCat#127628; RRID: AB_2562567BV421 Ter119BioLegendCat#116234; RRID: AB_2562917BV421 B220BioLegendCat#103239; RRID: AB_10933424BV421 CD49bBioLegendCat#108918; RRID: AB_2265144APC-Cy7 CD11bBioLegendCat#101226; RRID: AB_830642APC-Cy7 Gr-1BioLegendCat#108424; RRID: AB_2137485APC-Cy7 CD45BD BiosciencesCat#557659; RRID: AB_396774Biological samplesIndividuals with CML and healthy adult donors’ peripheral blood or bone marrow samplesThe first affiliated hospital of Zhengzhou University/Affiliated cancer hospital of Zhengzhou UniversityN/AChemicals, peptides, and recombinant proteinsFetal Bovine SerumGibcoCat#A3160902Phosphate Buffered SalineServicebioCat#G4202Opti-MEM mediumGibcoCat#11058021Puromycin Solution (10mg/mL)BiosharpCat#BL528APolybrene(10mg/mL)SolarbioCat#H8761Recombinant human SCFPeprotechCat#300-07Recombinant human IL-3PeprotechCat#200-03Recombinant human IL-6PeprotechCat#200-06Recombinant human GM-CSFPeprotechCat#300-03BC1215SelleckCat#S3022ImatinibMCECat#HY-15463CellTrace CFSEInvitrogenCat#C34554A5-FluorouracilMCECat#HY-90006Recombinant mouse SCFPeprotechCat#250-03Recombinant mouse IL-3PeprotechCat#213-13Recombinant mouse IL-6PeprotechCat#216-16Protein A/G Plus-AgaroseSanta CruzCat#Sc-2003S Protein Agarose BeadsMerckCat#69704-4MG132Sigma AldrichCat#M8699CCK-8ServicebioCat#G4101Cycloheximide (CHX)AladdinCat#C112766Critical commercial assaysFastPure Cell/Tissue Total RNA Isolation Kit V2VazymeRC112-01HiScript II One Step qRT-PCR SYBR Green KitVazymeQ221-01Lipofectamine 3000InvitrogenL3000-001Lentivirus Concentration SolutionYeasen41101ES50CD34 MicroBead KitMiltenyi Biotec130-046-703LS ColumnsMiltenyi Biotec130-042-401Annexin V-APC/DAPI Apoptosis Detection KitWuhan Pricella BiotechnologyP-CA-248Mitochondrial Membrane Potential Assay KitBoxbioAKOP013-2MethoCult M3434STEMCELLCat#03434MethoCult H4434STEMCELLCat#04434Deposited dataRaw dataData S1N/ARNA-seq dataGene Expression Omnibus (GEO)GEO: [GSE309070](GSE309070)CITE-seq data from Warfvinge et al.32GEOGEO: [GSE236233](GSE236233)ScRNA-seq data from Sommarin et al.72GEOGEO: [GSE173076](GSE173076)ScRNA-seq data from Giustacchini et al.32GEOGEO: [GSE76312](GSE76312)Expression profiling by array from Diaz-Blanco et al.73GEOGEO: [GSE5550](GSE5550)Expression profiling by array from Radich et al.74GEOGEO: [GSE4170](GSE4170)CO-IP MS dataProteomeXchange (PRIDE)PXD073522Experimental models: cell linesK562 cellsAmerican Type Culture CollectionN/AKBM5 cellsJin et al.15N/AKBM5-T315I cellsJin et al.15N/A293T cellsAmerican Type Culture CollectionN/AExperimental models: organisms/strainsNOD-Prkdcscid-IL2rgem9 (NSG) miceCharles RiverN/AC57BL/6JNuohang Biotechnology Research Institute Xuzhou Co., Ltd.N/ARecombinant DNAMSCV-BCR-ABL-IRES-GFPJin et al.15N/ApCMV-Gag-PolCell biolabsCat. # RV-111pCMV-VSV-GCell biolabsCat. # RV-110psPAX2Hanbio BiotechnologyN/ApMD2.GHanbio BiotechnologyN/AVector(pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3(pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3ΔF-box (pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3Δ57-121(pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3ΔSUKH (pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3ΔApaG (pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-FBXO3Δ409-471 (pcDNA3.1-CMV-MCS-3xflag-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-DUSP9(pcDNA3.1-CMV-MCS-MYC-EF1-ZsGreen-T2A-Puro)Hanbio BiotechnologyN/Ah-DUSP9(pcDNA3.1-3xflag)Hanbio BiotechnologyN/Ah-DUSP9Δ18-139(pcDNA3.1-3xflag)Hanbio BiotechnologyN/Ah-DUSP9Δ140-202(pcDNA3.1-3xflag)Hanbio BiotechnologyN/Ah-DUSP9Δ203-246(pcDNA3.1-3xflag)Hanbio BiotechnologyN/Ah-DUSP9Δ348-384(pcDNA3.1-3xflag)Hanbio BiotechnologyN/ASoftware and algorithmsPRISMGraphPad SoftwareVersion 9.0ImageJImageJ1.53eR version R 4.2.3The R Project for Statistical Computinghttps://www.r-project.org/Cluster Profiler R package version 4.16.0Bioconductorhttps://www.bioconductor.org/packages/Seurat R package version 5.1.0https://github.com/satijalab/seurathttps://satijalab.org/seurat/articles/Ucell R package version 2.12.0Bioconductorhttps://www.bioconductor.org/packagesrelease/bioc/html/UCell.html
Experimental model and study participant details
Cell lines
The human myeloid leukemia cell lines K562, BV173, and KCL22, the human peripheral blood basophilic leukemia cell line KU812, and the human embryonic kidney cell line HEK293T used in this study were all preserved in our laboratory before.75 The KBM5 cells and the KBM5-T315I cells were kindly donated by Professor Pan Jingxuan from Sun Yat-sen University. K562, BV173, KU812, and KCL22 were cultured in RPMI1640 medium (Transgen biotech, S10226) supplemented with 10% fetal bovine serum (FBS) (Gibco, A3160902) and 1% penicillin-streptomycin mixture (PS) (Servicebio, G4003). HEK293T was cultured in DMEM medium (Transgen biotech, R21017) supplemented with 10% FBS and 1% PS. KBM5 and KBM5-T315I cells were cultured in IMDM medium (Gibco, 2537734) supplemented with 10% FBS and 1% PS. All cells were grown at 37°C in a humidified atmosphere of 5% CO_2_.
Primary cells
The bone marrow (BM) samples of CML patients (N = 40) and healthy individuals (N = 10) used in this study were collected from the Affiliated Cancer Hospital of Zhengzhou University and First affiliated hospital of Zhengzhou Universities. The information of all patients was shown in the Table S7. Based on European LeukemiaNet criteria, CML patients were broadly classified into 3 groups according to the increasing degree of TKI resistance: Optimal response (major molecular response, MMR) to first-line imatinib (Optimal response, N = 10); Suboptimal responses to imatinib but favorable responses to second- or third-line TKIs (Sub-optimal responses, N = 15); Pan-TKI resistance with eventual blast phase (BP) progression (Treatment failure, N = 15). All samples were collected strictly according to the medical operation specifications and approved by the hospital ethics committee (2021-KY-0575-002). All patients and healthy subjects signed an informed consent form before sample collection. Mononuclear cells were isolated by density gradient centrifugation. Primary CD34^+^ cells were isolated using CD34 MicroBead Kit (Miltenyi Biotec, 130-046-703). Cells were cultured in IMDM medium supplemented with 10% FBS, 100 ng/mL GM-CSF (Peprotech, 300-03), 100 ng/mL SCF (Peprotech, 300-07), 20ng/mL IL-3 (Peprotech, 200-03), and 20 ng/mL IL-6 (Peprotech, 200-06) at 37°C in a humidified atmosphere of 5% CO_2_.
CML-CDX mouse model
The K562 cells of WT and FBXO3-KO were respectively injected into 4-5-week-old NOD-Prkdcscid-IL2rgem9 (NSG) mice via the tail vein (1×10^7^/mice). The mice were observed and monitored every day. After 6 weeks, the body weights of mice inoculated with WT cells and FBXO3-WT cells were weighed. Bone marrow and peripheral blood were collected and the proportion of human CD45^+^ cells were analyzed. The pictures and weights of spleens were recorded. The remaining mice in the two groups will continue to be raised, their survival was observed, and survival analysis experiments were carried out.
BCR::ABL1-induced CML mouse model
The retroviral construct MSCV-IRES-BCR-ABL-GFP was used to generate a high-titer helper-free retrovirus by transient transfection of HEK293T cells. BM cells from 5-fluorouracil-treated (200 mg/kg) C57BL/6J donor mice (6-8-week) were transduced twice with BCR::ABL1 retrovirus and cultured in DMEM medium supplemented with 15% FBS, 1% PS, 1.0 mg/mL ciprofloxacin, 200 mM L-glutamine, 6 ng/mL recombinant mouse IL-3, 10 ng/mL recombinant mouse IL-6, 50 ng/mL recombinant mouse stem cell factor. C57BL/6J recipient mice of the same sex were irradiated twice with a dose of 550cGy, at an interval of three hours. 5×10^5^ infected BM cells per mouse were injected via the tail vein. After two weeks, flow cytometry was performed to detect the green fluorescent protein myeloid cells and evaluate the progression of CML. The proportion of CML cells in the spleen and BM was analyzed via flow cytometry.
Method details
RNA extracts and qRT-PCR analysis
Total RNA was extracted from CML cell lines and primary CD34^+^ cells using the FastPure Cell/Tissue Total RNA Isolation Kit V2 (Vazyme, RC112-01) following the manufacturer’s protocol. Briefly, cell pellets were resuspended in 500 μL RL buffer, lysates were clarified using the provided spin and adsorption columns, and RNA was eluted in 100 μL RNase-free H_2_O. RNA concentration was quantified using a NanoDrop spectrophotometer, and samples were stored at −80°C used for qRT-PCR. One-step quantitative reverse transcription PCR (qRT-PCR) was performed using the HiScript II One Step qRT-PCR SYBR Green Kit (Vazyme, Q221-01) on an Applied Biosystems Step One PlusTM detection system (ABI 7300). Primer sequences are listed in Table S8. Relative gene expression was calculated using the 2^−ΔΔCt^ method, normalized to GAPDH as an internal control.
Western blot
Cells were collected and lysed with NP-40 protein lysis buffer. After centrifugation, the supernatant was collected and denatured at 100°C for 10 min and stored at −80°C until use. Proteins were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis on 6–15% gels and then transferred to 0.2μm polyvinylidene difluoride membranes (Immobilon). Membranes were blocked with the blocking buffer (20 mL TBST solution+1g skim milk powder) for 1 h at room temperature, followed by incubation with primary antibodies diluted in blocking buffer overnight at 4°C. Then after three washes with TBST, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) and quantified by densitometry (ImageJ v1.53), with GAPDH serving as a loading control.
ScRNA-seq analysis of CML-LSCs
ScRNA-seq data on bone marrow CD34^+^ cells from CML patients (N = 17, GEO: [GSE236233](GSE236233)) and healthy controls (HCs, N = 13, GEO: [GSE173076](GSE173076)) were included in this study. The raw droplet-based 10× Genomics scRNA-seq data were processed using CellRanger (7.0.1 version, 10× Genomics) to align with the human genome reference (GRCh38) and generate the gene×cell UMI matrix for each sample and gradually complete standardization, batch removal and reduction after strict quality control. We then used a list of 746 unique DEGs, which were identified by Giustacchini et al. and associated with the presence of the BCR::ABL1 fusion gene from CP-CML patients at diagnosis, to divide the HSCs of CML patients into BCR::ABL1 positive and negative cells. These 746 genes are assigned to the var. features slot of a Seurat object containing all CML HSCs (3457 cells). The Seurat object was analyzed using the default Seurat pipeline considering 20 PCs for dimensionality reduction and a resolution of 1 for Louvain clustering obtaining a set of clusters denoted by C. Marker genes for the obtained clusters are computed in a 1 vs. All fashion using the FindAllMarkers function with the parameters min.pct = 0.2, and logfc.threshold = 0.5. Finally, we further classified BCR::ABL1^+^ SCs into FBXO3^+/−^ cells based on their FBXO3 expression.
Lentivirus packaging and transduction
ShRNA constructs targeting FBXO3, and DUSP9 (sequences in Table S8) were purchased from Tsingke Biotechnology. The lentivirus supernatant was packaged in HEK293T cells by co-transfecting with pMD2.G and psPAX2. Viral supernatant was harvested 48 h post-transfection and concentrated with the YEASEN Lentivirus Concentration Kit. CML cells were transduced with viral particles in the presence of polybrene (2μg/mL) for 72 h, followed by puromycin selection (2μg/mL). Knockdown efficiency was validated by western blot (GAPDH normalization) and qRT-PCR (GAPDH reference).
CRISPR-Cas9 knockout assay
Single-guide RNAs (sgRNAs) targeting FBXO3 (Table S8) were cloned into the LentiCRISPRv2 vector. Lentiviral particles were generated in HEK293T cells via co-transfection with psPAX2 and pMD2.G packaging plasmids using polyethylenimine (PEI). The concentrated viruses were added into imatinib-sensitive (K562 cells) and imatinib-resistant (KBM5-T315I cells) cells to achieve the knockout of FBXO3.
Transfection of plasmids
K562 and KBM5-T315I cells in the logarithmic growth phase were harvested and seeded into 6-well plates at a density of 2×10^6^ cells per well. On the following day, 2 μg of plasmid was added to 250μL of OPTIMEM reduced serum medium, followed by the addition of 5μL of P3000 reagent. The mixture was gently mixed and allowed to stand. In a separate EP tube, 250 μL of OPTIMEM reduced serum medium was mixed with 5μL of Lipofectamine 3000 transfection reagent, and the mixture was gently vortexed and incubated for 5 min at room temperature. The contents of the two EP tubes were combined, gently mixed, and incubated for 20 min at room temperature. The mixture was then added to the cells in the 6-well plates, gently mixed, and subjected to flat-angle centrifugation at 2000 rpm for 90 min. After centrifugation, the plates were placed in an incubator for overnight culture. On the third day, a second transfection was performed. Cells were harvested 48 h after the second transfection for subsequent experiments.
Clonogenic assay
CML cells were resuspended in a complete medium supplemented with pre-warmed FBS at a density of 5×10^4^ cells/mL. Specifically, 400 μL cell suspension and 1 mL 2× medium was added respectively into a 2 mL EP tube containing 600 μL of soft agar. The mixture was then added to a 24-well plate with 500 μL per well. The plate was placed in an incubator for culturing. After two weeks, the number of clones (with the number of cells greater than 50) was quantified and documented using microscopy for further analysis.
Colony-formation cell/replacing assay
Primary CD34^+^ cells from CML patients and healthy donors were pre-treated according to the experimental protocol (infected with lentivirus/drug treatment). Cells were collected and resuspended at a density of 5×10^5^ cells/mL. Then 60 μL cell suspension was mixed with 540 μL of methylcellulose (with cytokines added in advance). This mixture was then distributed into a 96-well plate, with 100 μL per well. After 14 days, colonies of primary cells were counted using microscopy, and pictures were taken. Upon completion of colony counting, cells were collected. After being centrifuged and resuspended, they were washed once with PBS. Then the above steps were repeated for the second-generation culture.
Measurement of mitochondrial transmembrane potential
The mitochondrial membrane potential detection kit (Beijing Boxbio Science & Technology) was used. 500 μL of JC-10 staining working solution was added to the cell suspension and were thoroughly mixed. The cells were placed in an incubator at 37°C for 20 min and were mixed once every 5 min. After incubation, cells were centrifuged at 300 g for 5 min. The cells were then washed 3 times with JC-10 Dyeing Buffer (1×) and flow cytometry was used for detection.
Cell apoptosis analysis
The cells were harvested and centrifuged at 1000–1500 rpm for 5 min. Then the cells were resuspended with cold PBS. 5×10^5^ cells were transferred into a 1.5 mL microcentrifuge tube. Annexin V-FITC and PI staining solution were added. The tubes were gently vortexed and incubated in the dark at room temperature for 15–20 min. After incubation, PBS was added to dilute the staining solution. Flow cytometer was used to analyze. The apoptotic cells were defined to be Annexin V+ cells.
Apoptosis analysis of quiescent cells
Freshly prepared 0.1% BSA solution (100 mL PBS+0.1 g BSA) was used to adjust the concentration of CML/normal CD34^+^ cells to 1×10^6^ cells/mL. 1 μL of CFSE (5 mmol/L) was added to 1 mL cell suspension, and then was incubated at 37°C in the dark for 30 min. After 3 times wash, 1 mL cell suspension was added to each well and the plate was incubated at 37°C. After 72 h, cells were harvested as described above and Annexin V-APC was added. CFSE^max^CD34^+^Annexin V^+^ cells were determined as apoptotic quiescent cells.
Co-immunoprecipitation (Co-IP) assay
Cells were lysed in lysis buffer for 30 min at 4°C. The lysate was centrifuged at 12,000–14,000 rpm for 25 min at 4°C to remove cell debris. The supernatant was transferred to a new tube and added with the specific antibody against the target protein and the protein A/G agarose beads (pre-washed with lysis buffer). The mixture was incubated overnight at 4°C with gentle rotation. The next day, the sample was centrifuged at 12,000 rpm for 2 min at 4°C and the supernatant was removed. The beads were washed with lysis buffer 3 times to remove non-specifically bound proteins. The eluted proteins were analyzed by western blotting to identify the interacting proteins.
Liquid chromatography-mass spectrometry (LC-MS)
For the identification of FBXO3-interacting proteins, K562 cells (2×10^7^) were lysed in lysis buffer for 30 min at 4°C. The lysate was centrifuged at 12,000–14,000 rpm for 25 min at 4°C to remove cell debris. The supernatant was transferred to 2 new tubes and added with FBXO3 antibody (SANTA CRUZ, sc-514625) or mouse mAb IgG1 Isotype. Pre-washed protein A/G agarose beads were added and mixture was incubated overnight at 4°C with gentle rotation. The next day, the sample was prepared following CO-IP assay. Protein bands were collected from Coomassie Brilliant Blue-stained SDSPAGE gels. After enzymic hydrolysis and peptide desalination, LC-MS analysis was performed using Ultimate 3000nano ultra-high performance liquid phase tandem Q Exactive plus high resolution mass spectrometer (Oebiotech) (Table S6). The mass spectrometry proteomics data have been deposited to the PRIDE repository with the dataset identifier PXD073522.
Half-life experiment
Cycloheximide (CHX) was added into CML cells at a ratio of 1:1000 according to the time points so that the treatment time of CML cells with CHX was 0, 1.5, 3, 6, and 9 h respectively. The proteasome inhibitor MG132 (10 mM) was added at a ratio of 1:1000 4–6 h before protein extraction. After extracting proteins, western blot experiments were conducted to detect the changes in related proteins.
In vivo ubiquitination assay
K562 cells were transfected with indicated plasmids for 48h and treated with MG132(10 μM) for 6 h before the protein extraction. By using S Protein Agarose Beads, the proteins were processed according to the Co-IP experiment method described above. The eluted proteins were analyzed by western blotting to testify their expression levels.
CCK8 assay
CML cells were seeded in 96-well plates at 2×10^4^ cells/well and incubated overnight at 37°C under 5% CO_2_. Cells were treated with BC1215 at final concentrations of 0.1, 1, 5, 10, 25, and 50 μmol/L. Cell viability was assessed at 24, 48, 72, and 96 h using the Cell Counting Kit-8. Briefly, 10 μL CCK-8 reagent was added to each well, followed by incubation for 2 h at 37°C. Absorbance was measured at 450 nm using a microplate reader. Viability was calculated according to the absorbance.
Flow cytometry analyses of HSC/HSPC
To analyze the populations of myeloid leukemia cells and LSPCs in CML mice, BM and spleen cells from CML mice were analyzed by flow cytometry. For analysis of LSCs, cells were stained with BV421 lineage (BioLegend, 133311), APC-Cy7 Ly6A/E (Sca-1) (BioLegend, 108126), APC CD117 (c-Kit) (BioLegend, 105812), PE CD135 (BioLegend, 135306), BV605 CD48 (BioLegend, 103441), PerCP/Cy5.5 CD150 (SLAM) (BioLegend, 115921); For analysis of leukemia progenitor cells, cells were stained with BV421 lineage (BioLegend, 133311), APC-Cy7 Ly6A/E (Sca-1) (BioLegend, 108126), APC CD117 (c-Kit) (BioLegend, 105812), PE-Cy7 CD34 (BioLegend, 128618), and BV510 CD16/32 antibody (BioLegend, 156625).
RNA sequencing
Sh-control or DUSP9-targeting sh-RNAs were transfected in K562 cells and cells were collected after 48 h. Total RNA was isolated using TRIzol reagent (Takara). RNA sequencing was performed with the MGI platform (Annaroad Gene Technology (Beijing)Co., Ltd). Genes were significantly differentially expressed by the transcript fragments per kilobase million values if they satisfied the cutoff threshold criteria of a fold change of >2.0 and a false discovery rate of <5%. The RNA-seq data have been deposited in the Gene Expression Omnibus database. The access number is GEO: [GSE309070](GSE309070).
Reagents and antibodies
BC1215 (S3022) was purchased from Selleck Chemicals. MG132 (M8699) was purchased from Sigma Aldrich. Imatinib (HY-15463) and 5-Fluorouracil (HY-90006) were purchased from MedChemExpress. Cycloheximide (C112766) was purchased from Aladdin. Recombinant human SCF (300-07), IL-3 (200-03), IL-6 (200-06), and GM-CSF (300-03) were purchased from Peprotech. Recombinant mouse SCF (250-03), IL-3 (213-13), and IL-6 (216-16) were purchased from Peprotech. The information of western blot antibodies and flow cytometry antibodies were shown in key resources table.
Quantification and statistical analysis
Data were presented as means ± SEM. Statistical analyses were performed with GraphPad Prism 9.0 (GraphPad, San Diego, CA). Two-tailed Student’s t test was used to calculate the p values for comparisons between two groups, and one-way ANOVA, post hoc intergroup comparisons, and Tukey’s test were used to calculate the p values for comparisons among multiple groups. Log rank test was used to calculate the p values for Kaplan-Meier survival curves. Sample size (n) for statistical analyses was shown in related figure legends. p < 0.05 was considered statistically significant.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Cortes J.Pavlovsky C.Saußele S.Chronic myeloid leukaemia Lancet 39820211914192610.1016/s 0140-6736(21)01204-634425075 · doi ↗ · pubmed ↗
- 2Bower H.Björkholm M.Dickman P.W.Höglund M.Lambert P.C.Andersson T.M.L.Life Expectancy of Patients With Chronic Myeloid Leukemia Approaches the Life Expectancy of the General Population J. Clin. Oncol.3420162851285710.1200/jco.2015.66.286627325849 · doi ↗ · pubmed ↗
- 3Hochhaus A.Larson R.A.Guilhot F.Radich J.P.Branford S.Hughes T.P.Baccarani M.Deininger M.W.Cervantes F.Fujihara S.Long-Term Outcomes of Imatinib Treatment for Chronic Myeloid Leukemia N. Engl. J. Med.376201791792710.1056/NEJ Moa 160932428273028 PMC 5901965 · doi ↗ · pubmed ↗
- 4Baccarani M.Deininger M.W.Rosti G.Hochhaus A.Soverini S.Apperley J.F.Cervantes F.Clark R.E.Cortes J.E.Guilhot F.European Leukemia Net recommendations for the management of chronic myeloid leukemia: 2013 Blood 122201387288410.1182/blood-2013-05-50156923803709 PMC 4915804 · doi ↗ · pubmed ↗
- 5Senapati J.Sasaki K.Issa G.C.Lipton J.H.Radich J.P.Jabbour E.Kantarjian H.M.Management of chronic myeloid leukemia in 2023 - common ground and common sense Blood Cancer J.1320235810.1038/s 41408-023-00823-9PMC 1012306637088793 · doi ↗ · pubmed ↗
- 6Cortes J.E.Sasaki K.Kim D.W.Hughes T.P.Etienne G.Mauro M.J.Hochhaus A.Lang F.Heinrich M.C.Breccia M.Asciminib monotherapy in patients with chronic-phase chronic myeloid leukemia with the T 315I mutation after ≥1 prior tyrosine kinase inhibitor: 2-year follow-up results Leukemia 3820241522153310.1038/s 41375-024-02278-838755421 PMC 11217003 · doi ↗ · pubmed ↗
- 7Jiang Q.Li Z.Qin Y.Li W.Xu N.Liu B.Zhang Y.Meng L.Zhu H.Du X.Olverembatinib (HQP 1351), a well-tolerated and effective tyrosine kinase inhibitor for patients with T 315I-mutated chronic myeloid leukemia: results of an open-label, multicenter phase 1/2 trial J. Hematol. Oncol.15202211310.1186/s 13045-022-01334-z PMC 938980435982483 · doi ↗ · pubmed ↗
- 8Yang K.Fu K.Zhang H.Wang X.To K.K.W.Yang C.Wang F.Chen Z.S.Fu L.PBA 2, a novel inhibitor of the β-catenin/CBP pathway, eradicates chronic myeloid leukemia including BCR-ABL T 315I mutation Mol. Cancer 23202420910.1186/s 12943-024-02129-1PMC 1143830839342174 · doi ↗ · pubmed ↗