CXCR4 antagonistic lipid nanoparticles loading siRNA combat refractory AML through AML1-ETO depletion and homoharringtonine sensitization Authors: Xuelu Peng, Runxia Gu, Tao Wang, Chuanmei Jiang, Haiyan Xing, Jie Meng, Tao Wen, Jianxiang Wang, Jian Liu, Haiyan Xu Published: 2026 ABSTRACT Acute myeloid leukemia (AML) is characterized by uncontrolled malignant clonal proliferation of leukemic cells resulting from the blockade of myeloid hematopoietic stem/progenitor cell differentiation. The disease-causing fusion protein AML1-ETO (also known as RUNX1-ETO or RUNX1-RUNX1T1) and the chemokine (C-X-C motif) receptor 4 (CXCR4) have been recognized as crucial effectors. Strategies targeting each individual factor have been applied to develop new therapeutic approaches; however, clinical demands remain unmet, and much is still unknown about the crosstalk between the two factors. In this study, we utilized a lipid nanoparticle platform to carry AML1-ETO siRNA and the CXCR4 antagonistic peptide E5 (E5-LNP@siAE) to simultaneously deplete the fusion protein and inhibit CXCR4 activation, aiming to elucidate the crosstalk between the two factors and to develop a novel dual-functional therapeutic approach based on lipid nanoparticles. The resulting nanoparticles were investigated in a refractory AML mouse model (AML1-ETO & C-KITD816V) with a high level of CXCR4 and in the t(8; 21)-positive AML cell line Kasumi-1. It was shown that E5-LNP@siAE effectively achieved RNAi of AML1-ETO and antagonism of CXCR4, thereby synergistically inducing effective multi-lineage differentiation, leading to significantly enhanced differentiation-post apoptotic responses of AML cells to homoharringtonine and remarkably prolonged survival in refractory AML mice. Introduction Acute myeloid leukemia (AML) is the most common acute leukemia in adults, characterized by impaired differentiation of myeloid hematopoietic stem/progenitor cells, accompanied by uncontrolled malignant clonal proliferation of leukemic cells [1]. Although initial remission can be achieved with combination chemotherapeutic regimens in some young patients, AML remains high relapse rates and multi-drug resistance, leading to a very poor prognosis with a 5-year overall survival rate of approximately 32%. In the case of elderly patients, tolerance and responses to chemotherapy are extremely low, which leads to a 5-year survival of less than 10% [[2], [3], [4]]. AML is associated with a variety of chromosomal abnormalities and gene mutations [5]. Among these, the chromosomal translocation t(8; 21) (q22; q22) is one of the most prevalent and typically occurs in the French-American-British (FAB)-M2 subtype [6]. This translocation generates the well-known fusion protein AML1-ETO, which plays a pivotal role in the blockade of multiple lineage differentiation, including granulocytic [7,8], megakaryocytic and erythroid [9,10], as well as in leukemogenesis, disease progression and drug resistance development [11]. Moreover, patients with AML1-ETO are more susceptible to progressing to refractory leukemia when co-existing with other gene mutations (such as CKIT) synergistically, characterized by a significantly increased relapse rate and reduced survival rate [12,13]. RNA interference (RNAi) is the most extensively investigated approach for targeting AML1-ETO, primarily due to its effectiveness in silencing disease-causing genes [14,15]. Some pioneering studies use short hairpin RNA to silence the AML1-ETO to enhance leukemic cell apoptosis [16], the depletion of AML1-ETO by electroporation to increase median survival in a xenotransplantation AML mouse model [17], and the interference with leukemic propagation in engrafted immunodeficient mice using chemically modified RNA encapsulated in lipid nanoparticles [18]. The C–X–C chemokine receptor 4 (CXCR4) is highly expressed by many subtypes of AML cells [19], which enable the cells to interact with its specific ligand CXCL12 that is mainly secreted by stromal cells in the bone marrow and spleen [20,21]. When activated by CXCL12, CXCR4 promotes AML cells trafficking and homing to bone marrow and spleen, acquiring protections and develop drug resistance to multiple treatments, including RNAi as well as chemotherapeutics [[22], [23], [24]]. Hence, CXCR4 not only provides a specific site for targeted delivery, but also act as an attractive therapeutic target in AML therapy. Several well-known CXCR4 antagonists have been developed, such as the small molecule AMD3100 [25], the chemically synthesized peptide LY2510924 [26] and the biologically derived peptide BKT140 [27]. E5 is a CXCR4 antagonistic peptide that has been demonstrated high affinity for CXCR4 and antagonistic effect of inhibiting CXCL12-mediated CXCR4 activation, showing anti-leukemic activity in the refractory AML animal model either as monotherapy or in the form of nano-micelles loaded with doxorubicin [28,29]. Nevertheless, therapeutic outcomes of suppressing each single target have not yet met clinical demands, while the crosstalk between the two targets has not been uncovered or received adequate investigation, though crosstalk between the two targets have been indicated exist [30]. Aiming to figure out whether AML1-ETO and CXCR4 support each other and develop a more efficacious treatment strategy, in this work, we employed lipid nanoparticle (LNP) technology to combine RNAi of AML1-ETO and antagonism of CXCR4 in a unified platform, termed E5-LNP@siAML1-ETO, that simultaneously inhibits both targets. In the dual functional lipid nanoparticle, peptide E5 was conjugated to DSPE-PEG2000 to interfere with the CXCR4/CXCL12 axis and act as a targeting arrow to anchor CXCR4-expressing AML cells, while AML1-ETO siRNA form complexes with ionizable lipid molecules to have the gene silenced. The resulting nanoparticles were investigated in a refractory AML mouse model (AML1-ETO & C-KITD816V) with a high level of CXCR4 and in the t(8; 21)-positive AML cell line Kasumi-1. We demonstrated that E5-LNP@siAML1-ETO, administered as monotherapy, induced effective multi-lineage differentiation of AML cells. This effect was achieved by simultaneously downregulating AML1-ETO and interfering with the CXCR4/CXCL12 axis, which significantly inhibited bone marrow colonization and extramedullary infiltration of AML cells. More importantly, E5-LNP@siAML1-ETO strongly enhanced the anti-leukemic effect of homoharringtonine (HHT), markedly prolonging survival in refractory AML mice. Materials and methods Materials Dlin-MC3-DMA (MC3) was obtained from Sinopeg Biotech Co., Ltd. (Xiamen, China). Other lipid components—DSPC, cholesterol and DMG-PEG2000—were obtained from Southeast Nanomaterials Co., Ltd. (Jiangsu, China). The small interfering RNA (siRNA) and primers were synthesized by GenePharma Co., Ltd. (Suzhou, China) (Tables S1 and S2). The siRNA targeting AML1-ETO was abbreviated as siAE and the siRNA for negative control was abbreviated as siNC. DSPE-PEG-E5 was synthesized by Guoping Pharmaceutical Co., Ltd. (Anhui, China) and the sequence of E5 was GGRSFFLLRRIQGCRFRNTVDD. Homoharringtonine (HHT) was procured from Xi'an Hao-Xuan Bio-Tech Co., Ltd. (Shanxi, China). Cell culture The cell line with t(8; 21) (q22; q22) chromosomal translocation, Kasumi-1, was a kind gift from Professor Nanao Kamada (University of Hiroshima, Japan) [31,32]. The cells have been authenticated and tested for mycoplasma by Tsingke Biotech Co., Ltd (Beijing, China). The cells were maintained at 37 °C in a 5% CO2 incubator. The medium was RPMI 1640 (Hyclone, USA), supplemented with 10% fetal bovine serum (FBS, Gibco, USA), 100 U/mL penicillin and 100 U/mL streptomycin (Hyclone, USA). Preparation of lipid nanoparticles (LNPs) LNPs were prepared with microfluidic method and the device was supplied by Prinzen Biomedical Inc. (Shanghai, China). The siRNA and lipid components were dissolved in 50 mM citrate buffer and ethanol, respectively. The molar ratio of the lipids was 50:10:38.5:1.5 (MC3:DSPC:cholesterol:DMG-PEG2000). The weight ratio of MC3 to siRNA was maintained at 10:1. The ethanolic and aqueous fractions were mixed at a volume ratio of 1:3 with a total flow rate of 6 mL/min. The generated LNPs were dialyzed against PBS using 10K MWCO cassettes (Thermo Scientific, USA) for 2 h at 4 °C. DSPE-PEG-E5 was dissolved in PBS and incubated with the resulting LNPs at 45 °C for 30 min, with the molar ratio of DSPE-PEG-E5 to total lipid at 0.5:100. Following these steps, the LNPs were stored at 4 °C until they were used in subsequent experiments. LNPs encapsulating siNC or siAE were named as LNP@siNC and LNP@siAE, respectively. Accordingly, E5-modified LNPs were termed as E5-LNP@siNC and E5-LNP@siAE. Characterization of LNPs The morphology was observed by transmission electron microscopy (TEM, JEM-1400 Plus, Japan), and the particle size was quantified using ImageJ software. The particle size distribution, zeta potential, and the stability of LNPs in PBS and 10% FBS were determined via dynamic light scattering (DLS, Brookhaven, USA). The encapsulation efficiency was measured using the Quant-iT™ RiboGreen Assay Kit (Thermo Scientific, USA). During the assessment of the shelf-life of the LNPs, the storage temperature was at 4 °C, and the hydrodynamic diameters were determined based on the intensity distribution. Laser-scanning confocal microscopy Kasumi-1 cells (2 × 105/mL) were treated with LNPs for 6 h. The concentration of siRNA was 4 μg/mL (307.7 nM). After fixation with 4% PFA, the cells were stained with DAPI (ZSGB-Bio, China) for nuclear staining. The fluorescence intensity of FAM-siNC were observed using Laser Scanning Confocal Microscope (Leica, Germany). Flow cytometry assay Cellular uptake Kasumi-1 cells (4 × 105/mL) were incubated with LNP@FAM-siNC or E5-LNP@FAM-siNC (siRNA 1 μg/mL, 76.9 nM) for 6 h. The cells were collected and rinsed with PBS. Subsequently, the cells were incubated at room temperature for 3 min with or without 0.1% trypan blue. After removal of trypan blue by PBS washing, the mean fluorescence intensity of the cells was detected by flow cytometry. Surface CXCR4 expression Kasumi-1 cells (4 × 105/mL) were incubated with LNP@siNC, E5-LNP@siNC, LNP@siAE or E5-LNP@siAE (siRNA 4 μg/mL, 307.7 nM) for 2 h. Or the cells were incubated with different LNPs for 6 h, and then the medium containing LNPs was replaced with fresh medium and the cells were cultured until 24 h. To examine the effect of E5-LNP@siAE on chemotherapeutic agent-induced elevation of CXCR4 expression on cell surface, Kasumi-1 cells were incubated with HHT at 12.5 nM for 48 h at 37 °C. Subsequently, the cells were incubated with E5-LNP@siAE (siRNA 4 μg/mL, 307.7 nM) for 24 h at 37 °C. Finally, the cells were collected and rinsed with PBS, followed by the incubation of APC or PE labeled CXCR4 antibody (BioLegend, USA) for 30 min at 4 °C. The cells were analyzed using Attune® NxT flow cytometer (Thermo Scientific, USA). Cell apoptosis and differentiation assay Kasumi-1 cells (4 × 105/mL) were treated with different LNPs or E5-LNP@siAE combined with HHT for 6 h. The concentration of siRNA and HHT was 4 μg/mL (307.7 nM) and 12.5 nM, respectively. Then the cells were transferred to fresh culture medium and cultured until 72 h. After that, the cells were incubated with APC-conjugated Annexin V (Dojindo, Japan) or the antibodies against differentiation markers, including FITC labeled CD11b, CD41a and APC labeled CD235a (BioLegend, USA) for the detection of cell apoptosis and differentiation, respectively. The collected cells were analyzed using Attune® NxT flow cytometer (Thermo Scientific, USA). Cell proliferation assay Kasumi-1 cells (4 × 105/mL) were pre-incubated with carboxyfluorescein succinimidyl ester (CFSE, Dojindo, Japan). After incubation with LNPs (siRNA 4 μg/mL, 307.7 nM) for 6 h, LNPs were discarded and the cells were cultured for an additional 90 h. Afterwards, the cells were collected and washed with PBS for flow cytometer analysis. ROS measurement Kasumi-1 cells (4 × 105 cells/mL) were incubated with LNPs for 6 h at a siRNA concentration of 4 μg/mL (307.7 nM). Then LNPs were discarded and cells were cultured until 24 h. Following washing with PBS, cells were incubated in 2′, 7′-dichlorodihydrofluorescein diacetate (DCFH-DA, Germany) (10 μM) for 30 min at 37 °C. Afterwards, cells were washed with PBS again and subjected for flow cytometer analysis. Cell migration assay The LNP@siNC or E5-LNP@siNC at the siRNA concentration of 6 μg/mL (461.5 nM) was pretreated with Kasumi-1 cells for 2 h and added to the upper chambers. For RNAi experiment, Kasumi-1 cells (4 × 105/mL) were pretreated with LNP@siAE or E5-LNP@siAE (siRNA 4 μg/mL, 307.7 nM) for 6 h and the LNPs were discarded. Then the cells were cultured until 24 h and added to the upper chambers. The bottom chambers contained the complete medium with or without CXCL12 (200 ng/mL). The cells in the bottom chambers were collected and counted after 4 h incubation. Quantitative Real-time PCR (qRT-PCR) Kasumi-1 cells (4 × 105/mL) were incubated with LNPs (2 μg/mL siRNA, 153.8 nM for RNAi experiments and 4 μg/mL, 307.7 nM for differentiation tests) for 6 h. The culture medium was then replaced, and the cells were further cultured until reaching 24 or 48 h. Total RNA was isolated using Trizol reagent (Merck, Germany) and reverse transcribed into cDNA. Relative mRNA levels of AML1-ETO, CEBPA, PU.1 and GATA1 were measured via qRT-PCR with TB Green Premix Ex Taq (Takara, Beijing) and quantified using the comparative threshold cycle method (2−ΔΔCt). Western blot Kasumi-1 cells (4 × 105/mL) were incubated with LNPs (siRNA 2 μg/mL, 153.8 nM for RNAi experiments; siRNA 4 μg/mL, 307.7 nM for BCL-2 and CCND2 expression) for 6 h. The culture medium was then replaced, and the cells were further cultured until 24 or 48 h. Total intracellular proteins were extracted using RIPA lysis buffer (Solarbio, Beijing). Protein concentration quantification was performed using the BCA Protein Assay Kit (Thermo Scientific, USA). After following the manufacturer's instruction of western blot experiment, antibodies targeting AML1, BCL-2, CCND2 and β-actin (Cell Signaling Technology, USA) were utilized to detect the expression of the corresponding proteins. The signals were visualized using ECL Western detection reagents (Millipore, USA) with the optical imaging system (Tanon, Shanghai). Evaluation of pharmacological effects Establishment of AML1-ETO & CKITD816V AML mouse model The model was established in accordance with the procedure described previously [33]. Leukemia cells were isolated from the AML1-ETO & CKITD816V mouse model, developed by viral transduction of C57BL/6 mouse hematopoietic stem cells with MSCV-AML1/ETO-IRES-GFP and MSCV-HyC-KITD816V-IRES-GFP. All mice involved in this study were maintained under specific pathogen-free (SPF) conditions and performed in compliance with the guidelines approved by the Animal Care and Use Committee of the Institute of Basic Medical Sciences, Peking Union Medical College, Chinese Academy of Medical Sciences (ACUC-XMSB-2024-094). The C57BL/6 mice aged 6–8 weeks were subjected to X-ray irradiation at a dose of 4.5 Gy, firstly. Frozen splenic AML (AML1-ETO & C-KITD816V) cells were subsequently thawed and administered to the mice through tail vein injection, generating the primary experimental animals. These mice were euthanized upon reaching a moribund state. Next, the spleen was isolated, ground and sieved to prepare into single-cell suspension, which was then injected into the mice received irradiation, generating the formal experimental animals. Administration regimen The mice were assigned to different groups once the proportion of AML cells in peripheral blood (PB) reached 10-30%. The grouped mice received daily intraperitoneal injections of PBS, LNP@siNC, E5-LNP@siNC, LNP@siAE, or E5-LNP@siAE (siRNA dosage: 0.6 mg/kg) for a 4-day period (Ctrl group: n = 4; other groups: n = 6). For the combined therapy regimen, AML-bearing mice were administered intraperitoneal injections of both E5-LNP@siAE (siRNA 0.6 mg/kg) and HHT (0.25 mg/kg) (Ctrl group: n = 4; other groups: n = 6). The mice were sacrificed after treatment and the bone, spleen, lung, and liver were isolated. Flow cytometry Grind the spleen, lung, and liver from the AML mice received treatment to prepare a single-cell suspension for flow cytometry analysis. In particular, the red blood cells (RBC) in PB were lysed with lysing solution (BD Biosciences) prior to flow cytometric analysis. Hemogram assay The counts of white blood cells (WBC), RBC, platelets (PLT) and hemoglobin (HGB) in PB were detected by a hematology analyzer (Sysmex XT-2000i, Japan). Hepatorenal function detection After treatment, the serum was collected from the mice via orbital blood sampling, and alanine aminotransferase (ALT), alkaline phosphatase (ALP), total bilirubin (TBIL), total bile acid (TBA) and blood urea nitrogen (BUN) were subsequently detected using a biochemical analyzer (Beckman Coulter, USA). Western blot The spleens of AML mice in each group were isolated, ground, and filtered to prepare single-cell suspensions. Total protein was extracted from mouse splenocytes according to the method described in Section 2.9, and the expression levels of AML1-ETO in each group were subsequently determined. H&E and immunofluorescence staining The bone, spleen, liver, and lung from the AML mice received treatment were fixed in 4% paraformaldehyde (PFA). The tissue slides were subjected to hematoxylin-eosin (H&E) staining following standard protocols (Servicebio, Wuhan). The results were visualized using the microscope (EVOS M7000, Invitrogen, China). The staining of differentiation markers and AML1-ETO was performed using CD11b antibody (Servicebio, Wuhan), CD41a antibody (Abcam, UK), Ter119 antibody (Abcam, UK) and AML1 antibody (Cell Signaling Technology, USA) according to the manufacturers' instructions. Survival experiments The AML-bearing mice were assigned to 5 groups and received intraperitoneal injections of PBS, LNP@siNC, E5-LNP@siNC, LNP@siAE or E5-LNP@siAE once every 2 days (siRNA dosage: 0.6 mg/kg), with a total of 3 doses administered (n = 10). For the combined treatment cohort, the AML-bearing mice were assigned to 3 groups and received intraperitoneal injections of PBS, homoharringtonine (HHT) or E5-LNP@siAE combined with HHT once every 3 days, with a total of 2 doses administered (n = 10). The siRNA dose was set at 0.6 mg/kg. The administration dose of HHT was 0.25 mg/kg. Safety assessment of LNPs in vivo The C57BL/6 healthy mice aged 6–8 weeks were assigned to 3 groups and received daily intraperitoneal injections of PBS, LNP@siNC or E5-LNP@siNC (siRNA dosage: 0.6 mg/kg) for a 4-day period (n = 5). The counts of WBC, RBC, PLT and HGB in PB of each group were determined on day 2 and 4 post-administration. Subsequently, the serum was collected from the mice via orbital blood sampling on day 6, and ALT, ALP, TBIL, TBA and BUN were detected using a biochemical analyzer (Beckman Coulter, USA). Statistics The quantitative data in this study were presented as the mean ± SD. One-way ANOVA was used for result analysis when comparing three or more groups. Tukey's multiple comparison test was performed as a post-hoc analysis. Survival curves were estimated using the Kaplan-Meier method and compared with the log-rank test. The statistical analysis was carried out by GraphPad Prism 10.1.2. P < 0.05 was considered statistically significant. Results Preparation and characterizations of E5-LNP@siRNA To prepare lipid nanoparticles loaded with siRNA and E5, DSPE-PEG-E5 was synthesized and then mixed with lipid nanoparticles pre-encapsulating siRNA, in this process, DSPE-PEG-E5 inserted into the pre-prepared nanoparticles by self-assemble. The resulting product was referred to as E5-LNP@siRNA, while lipid nanoparticles without E5 were referred to as LNP@siRNA. In following experiments, siNC or siAE was used according to the experimental purposes. The siNC refers to the negative control siRNA, and siAE refers to the siRNA which targets AML1-ETO for silencing. Results of transmission electron microscopy (TEM) showed that both E5-LNP@siNC and LNP@siNC were homogeneous spheres with average diameters of 121.8 ± 25.1 nm and 71.0 ± 13.7 nm, respectively, indicating that the insertion of DSPE-PEG-E5 increased nanoparticle size to a certain degree (Fig. 1A). The average hydrodynamic diameters of E5-LNP@siNC and LNP@siNC were measured to be 172.0 nm and 139.8 nm, respectively, and the two LNPs remained stable for at least 7 days when stored in PBS or 10%FBS (Fig. 1B–C and S1). Results of zeta potential measurement indicated that E5-LNP@siNC was less negatively charged than LNP@siNC (Fig. 1D), because the negative charges were partly neutralized by the positively charged E5. Therefore, it is plausible to infer that E5 was displayed on the surface of the LNPs. When incubated with Kasumi-1 cells over-expressing CXCR4 (Fig. S2), E5-LNP@FAM-siNC was internalized much more than LNP@FAM-siNC, evidenced by confocal microscopy with stronger green fluorescence in the cells after a 6 h incubation (Fig. 1E). Meanwhile, this was also validated by flow cytometry analysis. Trypan blue was used to quench the fluorescence on the cell surface to determine the actual intracellular uptake of LNPs. The results showed that Kasumi-1 cells exhibited significantly higher uptake of E5-LNP@siNC compared with LNP@siNC, including both the fraction adsorbed on the cell surface and the internalized fraction inside the cells (Fig. 1F). The encapsulation efficiency (EE) of siRNA was determined to be 97% for both LNP@siNC and E5-LNP@siNC (Fig. S3), indicating that DSPE-PEG-E5 did not affect the encapsulation efficiency of siRNA in the lipid nanoparticles.Fig. 1Characterization of E5-LNP@siNC. (A) Representative TEM images and size distributions based on TEM of LNP@siNC and E5-LNP@siNC. Scale bar: 100 nm. (B) Size distributions of LNP@siNC and E5-LNP@siNC dispersed in PBS. (C) Hydrodynamic diameters and PDI of E5-LNP@siNC in PBS or 10%FBS for 7 days (n = 3). (D) Zeta-potentials of LNP@siNC and E5-LNP@siNC (n = 3). (E) Representative confocal images of Kasumi-1 cells incubated with different LNPs for 6 h. Scale bar: 10 μm. Green: FAM; Blue: cell nucleus. (F) The MFI of Kasumi-1 cells with or without trypan blue treatment after incubation with LNP@FAM-siNC or E5-LNP@FAM-siNC for 6 h (n = 3). **P < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Characterization of E5-LNP@siNC. (A) Representative TEM images and size distributions based on TEM of LNP@siNC and E5-LNP@siNC. Scale bar: 100 nm. (B) Size distributions of LNP@siNC and E5-LNP@siNC dispersed in PBS. (C) Hydrodynamic diameters and PDI of E5-LNP@siNC in PBS or 10%FBS for 7 days (n = 3). (D) Zeta-potentials of LNP@siNC and E5-LNP@siNC (n = 3). (E) Representative confocal images of Kasumi-1 cells incubated with different LNPs for 6 h. Scale bar: 10 μm. Green: FAM; Blue: cell nucleus. (F) The MFI of Kasumi-1 cells with or without trypan blue treatment after incubation with LNP@FAM-siNC or E5-LNP@FAM-siNC for 6 h (n = 3). **P < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) E5-LNP@siAML1-ETO induced effective RNAi and CXCR4 antagonism In this study, Kasumi-1 cell line was employed because it is a typical cell model of AML1-ETO positive and represent one of the most aggressive AML cells [34]. The siRNA specifically targeting the fusion site of AML1-ETO (siAML1-ETO, siAE) [18] was applied in this study. It was shown that either LNP@siAE or E5-LNP@siAE resulted in a decrease in AML1-ETO mRNA of Kasumi-1 cells detected at 24 h and 48 h (Fig. 2A), and the effect of E5-LNP@siAE was slightly stronger than that of LNP@siAE at the same time points. Consistently, western blotting analysis indicated that AML1-ETO protein was downregulated by E5-LNP@siAE and LNP@siAE at 48 h (Fig. 2B) as well; the former was stronger than the latter.Fig. 2E5-LNP@siAE achieved effective RNAi and CXCR4 antagonism. (A) Relative AML1-ETO mRNA levels of Kasumi-1 cells treated with different LNPs at 24 or 48 h (n = 3). (B) Western blot analysis of AML1-ETO protein of Kasumi-1 cells treated with different LNPs at 48h. (C) The MFI of CXCR4 on Kasumi-1 cells incubated with different LNPs at 2 h or 24 h (n = 3). (D) The relative migration rate of Kasumi-1 cells induced by CXCL12 and pretreatment of LNP@siNC and E5-LNP@siNC for 2 h, or LNP@siAE and E5-LNP@siAE for 6 h, followed by LNP removal and further cell culture for 24 h (n = 3). (E) Intracellular ROS levels in the Kasumi-1 cells measured by DCF fluorescence. (n = 3) *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. E5-LNP@siAE achieved effective RNAi and CXCR4 antagonism. (A) Relative AML1-ETO mRNA levels of Kasumi-1 cells treated with different LNPs at 24 or 48 h (n = 3). (B) Western blot analysis of AML1-ETO protein of Kasumi-1 cells treated with different LNPs at 48h. (C) The MFI of CXCR4 on Kasumi-1 cells incubated with different LNPs at 2 h or 24 h (n = 3). (D) The relative migration rate of Kasumi-1 cells induced by CXCL12 and pretreatment of LNP@siNC and E5-LNP@siNC for 2 h, or LNP@siAE and E5-LNP@siAE for 6 h, followed by LNP removal and further cell culture for 24 h (n = 3). (E) Intracellular ROS levels in the Kasumi-1 cells measured by DCF fluorescence. (n = 3) *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Next, we examined the influence of AML1-ETO knock-down on the CXCR4 expression. CXCR4 is an endocytic recycling receptor whose ligand engagement triggers rapid internalization and subsequent efficient recycling to the cell membrane [35]. It could be noticed that a 2 h co-incubation of LNP@siNC group had the CXCR4 level decreased slightly, and along with the treatment time extended, the CXCR4 level recovered to even slightly higher than that of control. This is because nanoparticles usually trigger cellular endocytosis, therefore leading to a transient internalization of some CXCR4 receptors nonspecifically. E5-LNP@siNC also had the CXCR4 expression decreased slightly after the 2 h co-incubation and the level returned to that of control, which suggested E5-LNP@siNC could not induced sufficient and long effect of CXCR4 antagonism. Differently, E5-LNP@siAE group achieved the strongest and most durable inhibitory effect after 24 h incubation (Fig. 2C). The CXCR4/CXCL12 axis serves as a core regulatory pathway driving the migration of AML cells, which directly mediates the homing of AML cells to the bone marrow niche [36]. It could be seen that the CXCR4 internalization induced by E5-LNP@siAE was accompanied with the inhibition of cell migration. Compared to the control, CXCL12, the ligand of CXCR4, effectively induced cells migration, and 2 h incubation of E5-LNP@siNC was able to inhibit the CXCL12-induced migration to a certain extent but LNP@siNC was not (Fig. 2D), attributable to the CXCR4 antagonistic effect of the E5. However, the AML1-ETO fusion gene can modulate the expression of migration-related genes and further enhance the migratory capacity of AML cells [37]. Therefore, LNP@siAE and E5-LNP@siAE, both produced stronger inhibitory effects in the long-term co-incubation, with E5-LNP@siAE achieving the strongest effect among them (Fig. 2D) as it simultaneously inhibited the two targets. Additionally, both LNP@siAE and E5-LNP@siAE reduced intracellular ROS levels (Fig. 2E). Although E5-LNP@siNC did not have this effect, E5-LNP@siAE achieved the greatest reduction, suggesting that the E5 modification enhanced the efficacy of LNP@siAE. It is known that abnormally high intracellular ROS levels are closely associated with drug resistance in AML [38,39]. Therefore, these results supported that targeting CXCR4 and AML1-ETO simultaneously was potent against refractory AML. E5-LNP@siAE reduced leukemia burden In the present study, a refractory AML mouse model featured with AML1-ETO-GFP gene and C-KITD816V-GFP gene was employed, referred to as AML1-ETO & C-KITD816V, and the AML cells in this model highly express CXCR4 [28]. Therapeutic administration was initiated when the proportion of leukemia cells reached approximately 10-30% in peripheral blood (PB) after the intravenous implantation of AML1-ETO & C-KITD816V cells (Fig. S4). In AML mice, hepatosplenomegaly accompanied by a marked increase in organ weight represents a typical visceral manifestation of extensive leukemic cell infiltration and disease progression [40]. Visually, the spleen volume of mice in the E5-LNP@siAE group was significantly smaller than that of the other groups (Fig. 3A), with remarkably reduced spleen and liver weights (Fig. 3B and C), indicating substantial alleviation of disease burden by E5-LNP@siAE. Meanwhile, E5-LNP@siAE had the white blood cells (WBC) reduced to the lowest level among the tested groups. Besides, red blood cells (RBC), hemoglobin (HGB) and platelets (PLT) of E5-LNP@siAE group returned to the levels comparable to those of the control (Fig. 3D). In particular, the platelet count was significantly higher than that of LNP@siAE. These results strongly suggested that the CXCR4 antagonism via E5 mitigated hematopoietic toxicity. Additionally, the H&E staining illustrated visually that E5-LNP@siAE markedly inhibited the colonization of AML cells in BM (epiphysis and diaphysis) and infiltration in the spleen, liver and lung, leading to pronounced restoration of normal tissue architecture (Fig. 3E). Specifically, AML cells exhibit a high nucleocytoplasmic ratio and are deeply stained with an overall bluish-purple color, whereas normal cells appear pink due to their abundant cytoplasm. Compared with other groups, the E5-LNP@siAE group showed a marked reduction in AML cells in BM, an expanded, pink-stained area, and the appearance of adipose vacuoles. In liver tissue, the number of pink-stained normal hepatocytes increased and arranged in a stellate pattern. The alveolar structure in lung tissue became clear, and the lung tissue regained a loose morphological appearance. In the spleen, the proportion of AML cells decreased while RBCs were increased.Fig. 3E5-LNP@siAE reduced leukemia burden. (A) The picture of spleens taken from AML mice with or without LNPs treatment. Average weight of the spleens (B) and livers (C) of AML mice in each group (n = 4-6). (D) The WBC, RBC, PLT and HGB counts analyzed by a hematology analyzer (n = 4-6). (E) The H&E staining of BM, liver, lung and spleen from the AML mice received different treatments. Scale bar: 100 μm *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. E5-LNP@siAE reduced leukemia burden. (A) The picture of spleens taken from AML mice with or without LNPs treatment. Average weight of the spleens (B) and livers (C) of AML mice in each group (n = 4-6). (D) The WBC, RBC, PLT and HGB counts analyzed by a hematology analyzer (n = 4-6). (E) The H&E staining of BM, liver, lung and spleen from the AML mice received different treatments. Scale bar: 100 μm *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Importantly, it was observed that both LNP@siAE and E5-LNP@siAE effectively down regulated the level of AML1-ETO in the spleen of the AML mice compared with the other LNP groups, with E5-LNP@siAE reaching the lowest AML1-ETO expression, which were examined by immunofluorescence staining (Fig. 4A and B) and Western blotting (Fig. 4C). These results indicated that E5-LNP@siAE was able to have AML1-ETO gene silenced in vivo.Fig. 4E5-LNP@siAE achieved effective RNAi in vivo. (A) Immunofluorescence staining of AML1-ETO and GFP in the spleen from AML mice received different LNP treatments. Scale bar: 100 μm. Blue: cell nucleus; Green: GFP; Red: AML1-ETO. (B) Quantification of the mean fluorescence intensity of AML1-ETO in each group. (C) Western blot analysis of AML1-ETO of the splenic cells from AML mice received different LNP treatments. *P < 0.05, ****P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) E5-LNP@siAE achieved effective RNAi in vivo. (A) Immunofluorescence staining of AML1-ETO and GFP in the spleen from AML mice received different LNP treatments. Scale bar: 100 μm. Blue: cell nucleus; Green: GFP; Red: AML1-ETO. (B) Quantification of the mean fluorescence intensity of AML1-ETO in each group. (C) Western blot analysis of AML1-ETO of the splenic cells from AML mice received different LNP treatments. *P < 0.05, ****P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) E5-LNP@siAE induced differentiation and apoptosis of AML cells We next examined resulting effects of E5-LNP@siAE on AML cells from the views of myeloid differentiation by AML1-ETO depletion and CXCR4 antagonism. As shown in Fig. 5A, following incubation with E5-LNP@siAE, increases in CEBPA, PU.1 and GATA1 were detected at mRNA levels, indicating differentiations to granulocytic, megakaryocytic and erythroid were initiated respectively.Fig. 5E5-LNP@siAE induced multi-lineage differentiation and cell apoptosis in vitro. (A) Relative mRNA levels of CEBPA, PU.1 and GATA1 of Kasumi-1 cells incubated with LNPs analyzed by qPCR (n = 3). (B) The MFI of CD11b, CD41a and CD235a of Kasumi-1 cells incubated with different LNPs detected by flow cytometry (n = 3). (C) Western blot and quantitative analysis of BCL-2 and CCND2 of Kasumi-1 cells treated with different LNPs (n = 3). (D) Proportion of Annexin V positive cells after LNPs treatment (n = 3). (E) The inhibition of cell proliferation measured by CFSE staining (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. E5-LNP@siAE induced multi-lineage differentiation and cell apoptosis in vitro. (A) Relative mRNA levels of CEBPA, PU.1 and GATA1 of Kasumi-1 cells incubated with LNPs analyzed by qPCR (n = 3). (B) The MFI of CD11b, CD41a and CD235a of Kasumi-1 cells incubated with different LNPs detected by flow cytometry (n = 3). (C) Western blot and quantitative analysis of BCL-2 and CCND2 of Kasumi-1 cells treated with different LNPs (n = 3). (D) Proportion of Annexin V positive cells after LNPs treatment (n = 3). (E) The inhibition of cell proliferation measured by CFSE staining (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Meanwhile, the effects of different LNPs on the expressions of CD11b, CD41a and CD235a of Kasumi-1 cells were also examined using flow cytometry, supplementing results at protein levels (Fig. 5B). The results demonstrated that only LNP@siAE and E5-LNP@siAE could upregulated the expressions of the differentiation markers while the other groups did not, and E5-LNP@siAE induced a more pronounced elevation, while the other LNPs did not. Anti-apoptosis gene (BCL2) and cell cycle regulation gene (cyclin D2, CCND2) are downstream ones of AML1-ETO that contribute to leukemogenesis [11]. Results of western blotting showed that E5-LNP@siAE significantly downregulated the levels of BCL2 and CCND2 in Kasumi-1 cells (Fig. 5C). Consistently, cell apoptosis was increased by E5-LNP@siAE in reference to the other groups (Fig. 5D), and the cell proliferation was suppressed, as evidenced by the increased fluorescence intensity of the cells because of the inhibition of cell division (Fig. 5E). Overall, these results indicated that E5-LNP@siAE was able to drive the myeloid differentiations to different directions by regulating the downstream of AML1-ETO transcription, promoting cell apoptosis and inhibiting cell proliferation. E5-LNP@siAE increased efficacy of chemotherapy HHT is a well-known chemotherapeutic agent against AML [[41], [42], [43]], however, with a quite narrow therapeutic window due to its high toxicity. We next investigated whether E5-LNP@siAE could sensitize AML cells to the drug, since both AML1-ETO and CXCR4 play important roles in chemotherapy resistance [44]. Firstly, the effects of different LNP formulations on the survival of the AML mice were examined. The results showed that among the LNP groups, only E5-LNP@siAE prolonged the survival of the AML mice (Fig. S5). Next, E5-LNP@siAE was applied in combination with HHT. Compared with HHT alone, the combination regimen of E5-LNP@siAE + HHT significantly prolonged the survival of AML mice (Fig. 6A) along with a decrease in AML cells in PB, spleen, liver and lung (Fig. 6B and S6). The H&E staining results clearly showed that E5-LNP@siAE + HHT inhibited the infiltration of AML cells in the BM (epiphysis and diaphysis), spleen, liver and lung, while restoring normal tissue architecture. Compared with the control group, mice treated with HHT monotherapy showed reductions of deeply stained blue-purple AML cells in the major organs, with the combination treatment achieving more pronounced effect. Moreover, many adipose vacuoles appeared in BM of the combination treatment group, suggesting the restoration of BM. The overall staining of the spleen tissue was lighter with an interspersed pink appearance, suggesting an increased proportion of RBCs. The liver tissue displayed more stellate-arranged normal hepatocytes, and the alveolar structure in the lung tissue also became clearly visible (Fig. 6C). Weights of the spleen and liver from the AML mice treated with the combination regimen were decreased significantly in reference to that of HHT alone (Fig. 6D), confirming the sensitization effect resulted from E5-LNP@siAE. In addition, the hepatorenal function indicators in the AML mice decreased after treatment, indicating that the liver infiltration caused by AML was relieved and the hepatorenal function was restored following the combined therapy.Fig. 6E5-LNP@siAE enhanced the anti-leukemic effect of HHT in vivo. (A) Survival curves of AML mice received different treatments (n = 10). (B) The leukemic cell percentages in the PB, spleen, liver and lung analyzed by flow cytometry (n = 4-6). (C) The H&E staining of BM, spleen, liver and lung from the AML mice received different treatment. Scale bar: 100 μm. (D) Average weight of the spleens and livers in each group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. E5-LNP@siAE enhanced the anti-leukemic effect of HHT in vivo. (A) Survival curves of AML mice received different treatments (n = 10). (B) The leukemic cell percentages in the PB, spleen, liver and lung analyzed by flow cytometry (n = 4-6). (C) The H&E staining of BM, spleen, liver and lung from the AML mice received different treatment. Scale bar: 100 μm. (D) Average weight of the spleens and livers in each group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. When AML cells infiltrate the liver, they induce hepatocellular injury and intrahepatic cholestasis. Concurrently, the metabolism and excretion of bilirubin and bile acids are impaired. These changes ultimately manifested as significant increases in serum ALT, ALP, TBIL, and TBA. Among these markers, TBA is a sensitive signal reflecting hepatic functional recovery. During AML progression, the hypermetabolic state contributes to renal impairment, accompanied by increased urea production, inadequate excretion, and subsequent urea retention in the circulation, ultimately manifesting as markedly elevated serum BUN levels [45,46]. Our results showed that the combination regimen reduced these parameters compared to Ctrl and HHT group, implying improvement of hepatic and renal function in the AML mice (Fig. 7A). Moreover, the WBC count of AML mice was significantly reduced while the RBC and PLT counts showed no significant changes (Fig. 7B).Fig. 7E5-LNP@siAE improved the indicators of hepatic and renal function as well as the hematological parameters in vivo. (A) The ALT, ALP, TBIL, TBA and BUN levels of AML mice received combined regimen detected by a biochemical analyzer (n = 4-6). (B) The WBC, RBC and PLT counts analyzed by a hematology analyzer (n = 4-6). *P < 0.05, ***P < 0.001. E5-LNP@siAE improved the indicators of hepatic and renal function as well as the hematological parameters in vivo. (A) The ALT, ALP, TBIL, TBA and BUN levels of AML mice received combined regimen detected by a biochemical analyzer (n = 4-6). (B) The WBC, RBC and PLT counts analyzed by a hematology analyzer (n = 4-6). *P < 0.05, ***P < 0.001. Furthermore, E5-LNP@siAE combined with HHT increased the expressions of CD11b, CD41a and CD235a (Fig. 8A), as well as the cell apoptosis rate (Fig. 8B) in vitro. It is known that chemotherapeutics usually increases the CXCR4 level of AML cells, which is closely associated with the development of drug resistance [44,47]. As shown that the cell-surface CXCR4 expression was elevated when Kasumi-1 cells were co-incubated with HHT. When these HHT-exposed cells were treated with E5-LNP@siAE, the up-regulation of CXCR4 was reduced (Fig. 8C), which demonstrated that E5-LNP@siAE could reduce the upregulation of cell-surface CXCR4 expression induced by HHT treatment and suggested an attractive inhibitory role in the development of HHT-resistance. Experimental results in vivo provided further evidence that the combination regimen induced remarkable differentiations to different directions in the spleen of the AML mice (Fig. 9), as shown that levels of CD11b, CD41a and Ter119 were increased, indicating the AML cells were induced to differentiate towards granulocytic, megakaryocytic and erythroid directions, respectively.Fig. 8E5-LNP@siAE induced cell differentiation and sensitized therapeutic efficiency of HHT in vitro. (a) The MFI of CD11b, CD41a and CD235a of Kasumi-1 cells in the combination regimen (n = 3). (b) Proportion of Annexin V positive cells in the combination regimen (n = 3). (c) The MFI of CXCR4 on Kasumi-1 cells in the combination regimen (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.Fig. 9The combination regimen induced cell differentiation in vivo. (A) Immunofluorescence staining of CD11b, CD41a and Ter119 in the spleen from AML mice received combination regimen. Scale bar: 100 μm. Blue: cell nucleus; Green: GFP; Red: CD11b/CD41a/Ter119. (B) Quantification of the mean fluorescence intensity of CD11b, CD41a or Ter119 in each group. **P < 0.01, ****P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) E5-LNP@siAE induced cell differentiation and sensitized therapeutic efficiency of HHT in vitro. (a) The MFI of CD11b, CD41a and CD235a of Kasumi-1 cells in the combination regimen (n = 3). (b) Proportion of Annexin V positive cells in the combination regimen (n = 3). (c) The MFI of CXCR4 on Kasumi-1 cells in the combination regimen (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. The combination regimen induced cell differentiation in vivo. (A) Immunofluorescence staining of CD11b, CD41a and Ter119 in the spleen from AML mice received combination regimen. Scale bar: 100 μm. Blue: cell nucleus; Green: GFP; Red: CD11b/CD41a/Ter119. (B) Quantification of the mean fluorescence intensity of CD11b, CD41a or Ter119 in each group. **P < 0.01, ****P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Safety assessment of LNPs in vivo To evaluate the in vivo safety of LNPs, the levels of WBC, RBC, PLT and HGB in peripheral blood of healthy mice in each group were detected. The results showed that there were no significant alterations in hematological parameters in the LNP@siNC and E5-LNP@siNC groups compared with the control group on day 2 (Fig. 10A) and day 4 (Fig. 10B) after administration. Meanwhile, no obvious changes in body weight were observed in all groups (Fig. 10C). In addition, serum was collected from the mice in each group to detect changes in hepatic and renal function markers, including ALT, ALP, TBIL, TBA, and BUN. The results showed that there were no significant abnormalities in hepatic and renal function–related indices in the LNP@siNC and E5-LNP@siNC groups compared with the control group, suggesting negligible systemic toxicity (Fig. 10D).Fig. 10Safety assessment of LNPs in vivo. The WBC, RBC, HGB and PLT counts of the mice in each group on day 2 (A) and day 4 (B) after administration (n = 5). (C) Body weight of the mice in each group on day 2, 4 and 6 after administration (n = 5). (D) The ALT, ALP, TBIL, TBA and BUN levels of the mice in each group on day 6 after administration (n = 5). Safety assessment of LNPs in vivo. The WBC, RBC, HGB and PLT counts of the mice in each group on day 2 (A) and day 4 (B) after administration (n = 5). (C) Body weight of the mice in each group on day 2, 4 and 6 after administration (n = 5). (D) The ALT, ALP, TBIL, TBA and BUN levels of the mice in each group on day 6 after administration (n = 5). Discussion AML is an aggressive hematological malignancy, featured with the uncontrolled and rapid growth of non-functional myeloblasts in the BM. This process is due to the differentiation blockade of normal myeloid precursors and highly occurrence of drug resistance and relapse [1]. The fusion protein AML1-ETO has been demonstrated to play a pivotal role in the inhibition of multiple lineage differentiation, as well as in the development of drug resistance, leukemogenesis and disease progression [11]. Nucleic acid drug therapy has been identified as a promising avenue for overcoming the limitations of conventional therapeutic modalities, including resistance to chemotherapy and the inability to correct genetic defects. Rather than merely treating the symptoms of the disease, nucleic acid drug therapy seeks to target the genetic factors that are its underlying cause. In view of the pathogenesis of AML, nucleic acid drug therapy is a highly promising therapeutic avenue. Simultaneously, evidence suggested a close association between drug resistance and relapses of AML and the CXCR4/CXCL12 axis [22]. The axis has been demonstrated to facilitate interactions between leukemic cells and stromal cells, thereby contributing to the development of minimal residual disease in bone marrow [48]. Recent literature has reported several indications of the potential for crosstalk between AML1-ETO and CXCR4 in AML. First, the inhibition of AML1 has been observed to result in the downregulation of CXCR4 and the impairment of its migratory activity [30]. Secondly, CXCR4 expression has been demonstrated to regulate drug resistance via AML1 [49]. However, there has been limited exploration of the implications to clinical applications of the crosstalk. In this work, we fabricated a novel lipid nanoparticle platform to harness AML1-ETO RNAi and CXCR4 antagonistic technologies, referred to as E5-LNP@siAE, aiming to figure out whether combinational effects exist and to develop a new therapeutic strategy of simultaneously depleting the fusion protein AML1-ETO and interfering with the CXCR4/CXCL12 axis to combat refractory AML. Results indicated that E5-LNP@siAE was able to induce the strongest differentiation effect. It is known that differentiation-inducing therapy is an ideal therapeutic strategy that can induce leukemic cells to restore the differentiation program [50], and the differentiated leukemic cells get rid of the non-stopping proliferation at immature status [51] and become more sensitive to chemotherapeutics compared with immature leukemic cells [52]. Functional suppression of the myeloid transcription factors represents a mechanism involved in the differentiation block in AML1-ETO leukemia. Therefore, AML1-ETO knockdown is essential for lifting the transcriptional repression [53]. Our results showed that the leukemic cells became more susceptible to HHT-induced apoptosis when they differentiated to the mature-like cells, evidenced by the increases of CD11b, CD41a and CD235a. As a result, the combination regimen prolonged the survival of the AML animals as well as induced the highest rate of apoptosis. Our results suggested that the CXCR4 antagonism could help to reach a notably better outcome in the differentiation-induction in vitro and in vivo. As a CXCR4 antagonistic peptide, E5 could overcome the stroma-mediated protection of leukemia cells and drive the leukemia cells out of the bone marrow and spleen [28], consequently, these leukemic cells might become more sensitive to the RNAi of AML1-ETO. Besides, as displayed on the LNPs’ surface, E5 enabled the LNPs specifically to bind CXCR4-expressing AML cells, which could increase their targeting ability. It could be noticed that in the cellular experiments, the silencing efficiency of AML1-ETO did not reflect the advantage of E5-modification between LNP@siAE vs E5-LNP@siAE. It is plausible to consider following reasons: one is that the uptake difference for the two LNPs was not as large as expected in vitro (Fig. 1E), and a portion of internalized LNPs were trapped in lysosomes and failed to fully release into the cytoplasm, even more internalization was achieved due to the E5 modification. Therefore, further optimization for the delivery system is necessary. In the in vivo studies, the effects of E5 modification on the AML1-ETO knockdown, differentiation and survival showed significant benefits. Leukemic cells are usually under oxidative stress with abnormally high levels of ROS, which help the cells develop drug resistance [39]. It was shown that CXCR4 antagonism could enhance the function of RNAi for AML1-ETO to downregulate the ROS levels significantly (Fig. 2E), which could be beneficial in sensitizing AML cells to chemotherapeutics. HHT is a cytotoxic alkaloid isolated from Cephalotaxus species, acting as a protein translation inhibitor and has garnered increasing research interest as part of standard therapy [42]. Recent drug screening studies suggest that HHT shows preferential anti-leukemic activity in combination therapies for relapsed/refractory AML [43,54]. However, the therapeutic window of HHT is quite narrow, increasing treatment dose may bring slight improvement of the efficiency, however, serious side effects is increased. Therefore, clinical application of HHT is limited by dose-related toxicities and the development of drug resistance [42,55]. In this study, we chose a relative low dose to investigate whether E5-LNP@siAE could improve the HHT treatment. Our results showed that E5-LNP@siAE was able to increase the efficacy of HHT, resulting in the prolongation of survival of the AML mice, suggesting the combination regimens of E5-LNP@siAE with low dose of HHT is of translational potential in clinical application. Conclusion In summary, this work constructed a lipid nanoparticle platform (E5-LNP@siAE) that harnessed the depletion of AML1-ETO fusion protein by RNAi and blockade of CXCR4/CXCL12 axis by antagonism, which achieved successful combinational therapeutic effects on Kasumi-1 cells and in a refractory AML mouse model. As one monotherapy, E5-LNP@siAE was demonstrated to markedly reduce the leukemia burden. Moreover, when combined with chemotherapeutic drug, HHT, E5-LNP@siAE significantly enhanced the anti-leukemic efficacy of HHT. Therefore, E5-LNP@siAE provided a promising targeted therapy strategy for the treatment of refractory AML. CRediT authorship contribution statement Xuelu Peng: Data curation, Formal analysis, Investigation, Validation, Visualization, Writing – original draft. Runxia Gu: Resources, Writing – review & editing. Tao Wang: Funding acquisition, Investigation, Validation. Chuanmei Jiang: Investigation. Haiyan Xing: Investigation, Resources. Jie Meng: Investigation. Tao Wen: Investigation. Jianxiang Wang: Conceptualization, Resources. Jian Liu: Conceptualization, Funding acquisition, Writing – review & editing. Haiyan Xu: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.