All bioinformatic analyses in this study were performed using the R software environment (version 4.4.1). The BloodSpot database (https://www.bloodspot.eu/) was utilized to integrate the GSE13159 and GSE42519 datasets and to compare ALKBH5 expression levels between AML patients and normal controls. RNA-Seq expression data and corresponding clinical information for the TCGA-LAML cohort were retrieved from The Cancer Genome Atlas (TCGA) database to conduct survival analysis of ALKBH5. The AML dataset GSE13159 (GPL570 platform) was obtained from the GEO database for differential expression analysis, functional enrichment, and immune infiltration analysis. Differentially expressed genes (DEGs) between high- and low-ALKBH5 expression groups were identified using the limma R package, with the selection criteria defined as |log₂FC| > 0.3 and adjusted P value < 0.05. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of the identified DEGs were conducted using the clusterProfiler package. The CIBERSORT algorithm was applied to the normalized expression matrix of the GSE13159 dataset to estimate the proportions of 22 immune cell subtypes. Spearman correlation analysis was subsequently performed to assess the associations between ALKBH5 expression, immune cell infiltration levels, and the expression of key immune checkpoint molecules.

This study enrolled 77 newly diagnosed AML patients and 13 non-hematologic disease controls who were admitted to the First Affiliated Hospital of Harbin Medical University between January 2020 and June 2024 (Supplementary Table 1). All study protocols were reviewed and approved by the Ethics Committee of Harbin Medical University (approval number: hrbmuecdc20251201), and written informed consent was obtained from all participants. Bone marrow samples were collected from all subjects, and bone marrow mononuclear cells (BMNCs) were isolated using Ficoll (Solarbio, Beijing, China) density gradient centrifugation for subsequent total RNA extraction. Clinical data collected included age, sex, French-American-British (FAB) classification, chromosomal karyotype, proportions of bone marrow and peripheral blood blasts, white blood cell count, and lymphocyte percentages determined by flow cytometry.

The human acute myeloid leukemia cell lines THP-1 and MOLM-13 were generously provided by Tongji Hospital, Huazhong University of Science and Technology. The human bone marrow stromal cell line HS-5 was purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China). THP-1 and MOLM-13 cells were cultured in RPMI-1640 medium (Gibco, Shanghai, China) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Abbkine, USA) and 1% penicillin-streptomycin (Beyotime, Shanghai, China). HS-5 cells were maintained in DMEM medium (Gibco, Shanghai, China) with the same supplements. All cells were incubated at 37 °C in a humidified incubator containing 5% CO₂.

To establish ALKBH5 stable knockdown cell models, three specific short hairpin RNAs (shRNAs) and one negative control shRNA (sh-NC) were synthesized and packaged into lentiviruses by General Biol Co., Ltd. (Hefei, China). Polybrene (General Biol, Hefei, China) was added during transduction to enhance infection efficiency. Seventy-two hours post-transduction, cells were selected in medium containing 4 μg/mL puromycin (Beyotime, Shanghai, China) until the control group cells were completely eliminated. Knockdown efficiency was confirmed by RT-qPCR and Western blot analyses. The sequences of shRNAs used in this study are listed in Supplementary Table 2.

Total RNA was extracted from cells using TRIzol reagent (Takara, Beijing, China). RNA concentration and purity were determined using a NanoDrop 2000C spectrophotometer (Thermo Fisher Scientific, MA, USA). One microgram of total RNA was treated with the PrimeScript™ RT reagent Kit with gDNA Eraser (Takara, Beijing, China) to remove genomic DNA and reverse transcribed into cDNA. Real-time quantitative PCR was performed using the SYBR^® Green Kit (Roche, Basel, Switzerland) on an SLAN-96P real-time PCR system (Hongshi, Shanghai, China). The thermal cycling conditions were as follows: 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. β-actin was used as the internal reference gene, and the relative expression of target genes was calculated using the 2^-ΔΔCt method. The primer sequences used in this study are listed in Supplementary Table 3.

Total protein was extracted from cells using RIPA lysis buffer supplemented with protease inhibitors (Beyotime, Shanghai, China), and protein concentration was determined using a BCA Protein Assay Kit (Beyotime, Shanghai, China). Thirty micrograms of protein were separated on 10% SDS-PAGE gels and then transferred onto nitrocellulose membranes using the wet transfer method (LABSELECT, Shanghai, China). The membranes were blocked with 5% skimmed milk (Biosharp, Beijing, China) for 1 hour and incubated overnight at 4 °C with the respective primary antibodies: rabbit polyclonal anti-ALKBH5 (1:5000, Origene, Wuxi, China), rabbit monoclonal anti-PD-L1 (1:2000, Proteintech, Wuhan, China), and mouse monoclonal anti-GAPDH (1:2000, Origene, Wuxi, China). The following day, membranes were incubated with HRP-conjugated secondary antibody goat anti-mouse IgG (1:10000, Abbkine, Wuhan, China) at room temperature for 1 hour. Signals were detected using ECL chemiluminescent substrate (Abbkine, Wuhan, China) and captured with the Tanon automated chemiluminescence imaging system (Tanon Image System, Shanghai, China). Band intensities were quantified using ImageJ software (NIH, Bethesda, MD, USA), and the signal of each target protein was normalized to the corresponding internal control.

Cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8, Abbkine, Wuhan, China). Cells were seeded into 96-well plates at a density of 1×10⁴ cells per well. At 24, 48, and 72 hours post-seeding, 10 μL of CCK-8 reagent was added to each well. After incubation for 1.5 hours, absorbance was measured at 450 nm. For each time point and condition, six replicate wells were set up, and the entire experiment was independently repeated three times.

Cell migration was assessed using Transwell chambers with 8.0 μm pores (LABSELECT, Shanghai, China). A total of 1×10⁵ cells, starved in serum-free medium for 24 hours, were resuspended and added to the upper chamber. The lower chamber was filled with complete medium containing 10% FBS as a chemoattractant. After 48 hours of incubation, cells that had migrated to the lower chamber were collected and quantified using the CCK-8 assay.

Peripheral blood was collected from healthy donors, and peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll density gradient centrifugation (Solarbio, Beijing, China). PBMCs were resuspended in complete medium and activated with ImmunoCult™ Human CD3/CD28 T Cell Activator (STEMCELL Technologies, Shanghai, China) and recombinant human IL-2 (final concentration 20 ng/mL, Abbkine, Wuhan, China) for 3–5 days for subsequent experiments.

Activated PBMCs were co-cultured with MOLM-13 cells at an effector-to-target (E:T) ratio of 4:1 in 24-well plates, with target cells (MOLM-13) seeded at 1×10⁵ cells per well and effector cells (PBMCs) at 4×10⁵ cells per well. Cells and supernatants were collected at 24 and 48 hours. For flow cytometry analysis, cells were surface-stained with PE anti-CD3, FITC anti-CD4, and APC anti-CD8 antibodies (Elabscience, Wuhan, China), and analyzed using a FACSCanto™ II flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). Concentrations of IFN-γ and TNF-α in the supernatants were measured using ELISA kits (Elabscience, Wuhan, China), and the survival rate of AML cells post co-culture was assessed using the CCK-8 assay.

All statistical analyses were performed using GraphPad Prism 9.0 and R software (v4.4.1). Quantitative data conforming to a normal distribution are presented as mean ± standard deviation (SD), and comparisons between two groups were performed using Student’s t-test, while comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA). Non-normally distributed data are presented as median and compared using the Mann-Whitney U test. Categorical data were analyzed using the chi-square test. Correlation analyses were conducted using Pearson or Spearman correlation based on data distribution. Survival analyses were performed using the Kaplan-Meier method and compared with the log-rank test. P < 0.05 was considered statistically significant.

To systematically investigate the role of ALKBH5 in AML, we first integrated GSE13159 and GSE42519 datasets via the BloodSpot database, which revealed that ALKBH5 expression was markedly elevated in primary AML patients bearing various chromosomal translocations compared to normal hematopoietic stem cells (Figure 1A). Subsequently, we performed bioinformatic analyses. Survival analysis of the TCGA-LAML cohort demonstrated that patients were stratified into high- and low-ALKBH5 expression groups based on the median ALKBH5 expression level, and those with high ALKBH5 expression exhibited significantly shorter overall survival compared with patients in the low-expression group. (P < 0.01, Figure 1B).

To explore the biological functions associated with high ALKBH5 expression, differential gene expression analysis between ALKBH5 high- and low-expression groups was conducted, followed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. The results showed that differentially expressed genes (DEGs) were significantly enriched in immune-related processes, including positive regulation of leukocyte activation, T cell cytokine production, and regulation of cytokine production involved in immune response (Figure 1C). Pathway analysis further revealed enrichment in key immune signaling pathways such as NF-κB signaling, T cell receptor signaling, and IL-17 signaling (Figure 1D), suggesting a potential role for ALKBH5 in modulating immune regulation in AML.

Subsequently, we applied the CIBERSORT algorithm to assess the relationship between ALKBH5 expression and the tumor immune microenvironment. High ALKBH5 expression was significantly positively correlated with the infiltration levels of CD8⁺ T cells, resting NK cells, and M0 macrophages (Figure 2A, Supplementary Figure 1). Notably, ALKBH5 expression was strongly positively associated with several key immune checkpoint molecules, including PD-L1, CTLA-4, and LAG3 (Figures 2B, C). In summary, these bioinformatics results suggest that ALKBH5 is highly expressed in AML and may promote disease progression by regulating the immune microenvironment. Based on this, we hypothesize that ALKBH5 may mediate immune escape in AML by up-regulating PD-L1. To verify this hypothesis, we further conducted in vitro functional and mechanistic studies.

To validate the aforementioned bioinformatic findings, bone marrow samples from 77 newly diagnosed AML patients and 13 healthy controls were collected for experimental verification. RT-qPCR results confirmed that ALKBH5 mRNA levels were significantly elevated in AML patient samples compared with healthy controls (P < 0.001, Figure 3A), with the highest expression observed in the M1 subtype and the lowest in M4 (Figure 3B). AML patients were stratified into high- and low-ALKBH5 expression groups based on the median expression level, and their clinical characteristics were analyzed. Patients in the ALKBH5 high-expression group exhibited a significantly higher proportion of peripheral blood blasts (58.00% vs. 33.50%, P = 0.02) and a trend toward higher bone marrow blast percentage (51.00% vs. 44.00%, P = 0.05), indicating a close association between ALKBH5 expression and tumor burden in AML (Table 1, Figure 3C, Supplementary Figure 2). Notably, the proportion of lymphocytes in the bone marrow (assessed by flow cytometry) was slightly higher in the ALKBH5 high-expression group than in the low-expression group, though the difference was not statistically significant (6.50% vs. 5.39%, P = 0.72), which may be attributable to limited sample size. More importantly, clinical sample data successfully validated the bioinformatic prediction: ALKBH5 expression was significantly positively correlated with PD-L1 mRNA levels (r = 0.30, P = 0.0073, Figure 3D). Additionally, PD-L1 expression levels were significantly higher in the ALKBH5 high-expression group compared with the low-expression group (P = 0.01). These clinical findings directly link elevated ALKBH5 expression to the key immune evasion molecule PD-L1 in AML, providing a clear direction for subsequent mechanistic investigations.

To validate the expression pattern and functional role of ALKBH5 at the cellular level, we first assessed ALKBH5 expression across multiple cell lines. Both RT-qPCR and Western blot analyses consistently demonstrated that ALKBH5 mRNA and protein levels were significantly upregulated in AML cell lines (THP-1, MOLM-13) compared with normal bone marrow stromal cells (HS-5) (Figures 4A–C), which is highly consistent with our clinical sample findings.

To investigate the biological function of ALKBH5, lentivirus-mediated shRNA was used to knock down ALKBH5 in THP-1 and MOLM-13 cells. Three distinct shRNA sequences (sh-ALKBH5#1, #2, #3) were constructed. Both RT-qPCR and Western blot confirmed that all three shRNAs effectively reduced ALKBH5 expression compared with the negative control (sh-NC), with sh-ALKBH5#3 exhibiting the most stable and pronounced knockdown across all cell lines (Figures 4D–I). Therefore, subsequent experiments employed the stable knockdown model generated using sh-ALKBH5#3. Functional assays revealed that ALKBH5 knockdown significantly inhibited AML cell proliferation (P < 0.01, Figure 4J) and migration (P < 0.001, Figure 4K), confirming the direct role of ALKBH5 in driving malignant progression in AML. These findings establish the role of ALKBH5 in driving the intrinsic malignant behavior of AML cells. Based on pathway enrichment analysis, we further speculate that ALKBH5 may also facilitate the immune evasion of AML cells, which could represent another critical mechanism contributing to poor patient prognosis.

Next, we investigated the molecular mechanism by which ALKBH5 regulates PD-L1. Consistent with clinical observations, ALKBH5 knockdown markedly reduced PD-L1 protein expression (Figures 5B, C, E, F). Interestingly, PD-L1 mRNA levels were not significantly altered (Figures 5A, D), indicating that ALKBH5 primarily regulates PD-L1 at the post-transcriptional level, potentially by affecting mRNA stability or translation efficiency.

Finally, the immunomodulatory function of ALKBH5 was assessed using an in vitro co-culture system. Activated peripheral blood mononuclear cells were co-cultured with control (sh-NC) or ALKBH5 knockdown (sh-ALKBH5) MOLM-13 cells. Compared with sh-NC co-cultures, the proportion of CD8⁺ T cells was significantly increased in the sh-ALKBH5 co-cultures (P < 0.05, Figure 5H), and the secretion of T cell effector cytokines IFN-γ and TNF-α in the supernatant was also markedly elevated (P < 0.05, Figure 5I). Correspondingly, AML cells with ALKBH5 knockdown exhibited reduced survival in the co-culture system (P < 0.01, Figure 5G), indicating enhanced sensitivity to T cell-mediated cytotoxicity.