Single-cell transcriptome profiling of m6A regulator-mediated methylation modification patterns in elderly acute myeloid leukemia patients Authors: Zhe Wang, Xin Du, Peidong Zhang, Meiling Zhao, Tianbo Zhang, Jiang Liu, Xiaolan Wang, Doudou Chang, Xiaxia Liu, Sicheng Bian, Xialin Zhang, Ruijuan Zhang Published: 2024 ABSTRACT AbstractMillions of people worldwide die of acute myeloid leukaemia (AML) each year. Although N6-methyladenosine (m6A) modification has been reported to regulate the pathogenicity of AML, the mechanisms by which m6A induces dysfunctional hematopoietic differentiation in elderly AML patients remain elusive. This study elucidates the mechanisms of the m6A landscape and the specific roles of m6A regulators in hematopoietic cells of elderly AML patients. Notably, fat mass and obesity-associated protein (FTO) was found to be upregulated in hematopoietic stem cells (HSCs), myeloid cells, and T-cells, where it inhibits their differentiation via the WNT signaling pathway. Additionally, elevated YT521-B homology domain family proteins 2 (YTHDF2) expression in erythrocytes was observed to negatively regulate differentiation through oxidative phosphorylation, resulting in leukocyte activation. Moreover, IGF2BP2 was significantly upregulated in myeloid cells, contributing to an aberrant chromosomal region and disrupted oxidative phosphorylation. m6A regulators were shown to induce abnormal cell-cell communication within hematopoietic cells, mediating ligand-receptor interactions across various cell types through the HMGB1-mediated pathway, thereby promoting AML progression. External validation was conducted using an independent single-cell RNA sequencing (scRNA-Seq) dataset. The THP-1 and MV411 cell lines were utilized to corroborate the m6A regulator profile; in vitro experiments involving short hairpin RNA (shRNA) targeting FTO demonstrated inhibition of cell proliferation, migration, and oxidative phosphorylation, alongside induction of cell cycle arrest and apoptosis. In summary, these findings suggest that the upregulation of m6A regulators in HSCs, erythrocytes, myeloid cells, and T-cells may contribute to the malignant differentiation observed in AML patients. This research provides novel insights into the pathogenesis of AML in elderly patients and identifies potential therapeutic targets.Supplementary InformationThe online version contains supplementary material available at 10.1186/s43556-024-00234-7.Keywords: M6A, Acute myeloid leukemia, Malignant differentiation, WNT signaling, Single-cell RNA-seq Introduction Acute myeloid leukaemia (AML) is a haematological malignancy characterized by strong heterogeneity and high mortality, and the incidence rate of AML continues to increase year by year [1]. Approximately 75% of patients affected by this disease are aged 65 and older [2]. Currently, the composite complete remission (cCR) rate in older adults with AML is only 17–66%, and the 5-year overall survival (OS) remains poor, reflecting the adverse disease biology and clinical symptoms of this population [3–5]. Numerous studies have identified that RNA modification is a posttranscriptional change in the chemical composition of RNA molecules, has the potential to alter diverse biological functions [6]. N6-methyladenosine (m6A) is the most prevalent modification found in eukaryotic mRNA; m6A can be stalled by methyltransferases (writers), namely, METTL3, METTL14 and WTAP, reversed by demethylases (erasers), including FTO and ALKBH5, and recognized by m6A binding proteins (readers), such as YTHDC1/2, YTHDF1-3 and IGF2BP1-3 [7]. Recently, it has been reported that the dysregulation of m6A modification is involved in the occurrence and development of various cancers [8]. For instance, METTL3 promotes the translation of the downstream genes MYC, BCL2, and PTEN through m6A modification and inhibits the pAKT pathway, while METTL14 promotes the expression of the target genes MYB and MYC, thus promoting cell proliferation and inhibiting differentiation/apoptosis [9, 10]. Additionally, FTO can enhance the stability of MYC and CEBPA mRNA, thereby promoting cell proliferation [11], and METTL14 and YTHDF2 promote the self-renewal of leukaemia stem cells (LSCs)/leukaemia initiating cells (LICs), while ALKBH5K promotes LSC proliferation and induces apoptosis [10, 12, 13]. Moreover, m6A regulators are closely associated with the survival of AML patients [14]. The cellular hierarchy in AML patients is probably similar to the normal haematopoietic differentiation hierarchy, accompanied by functional impairment of various cell types [15]. The balance between the proliferation and differentiation of haematopoietic stem cells (HSCs), Myeloids, TCells and BCells is dysfunctional in AML patients [16, 17]. Furthermore, the immune function of Myeloids, TCells and BCells is changed in AML patients, especially the transformation of erythrocytes into immunosuppressive myeloid cells [17, 18]. The signaling pathways and metabolism of each cell type are also altered [19, 20]. At present, the regulatory mechanisms underlying the aberrant differentiation of AML by m6A regulators remain unclear. Single-cell RNA sequencing (scRNA-seq) can potentially be used to solve this problem by identifying the unique characteristics of each cell type in tissue samples and providing a more detailed understanding of intercellular communication networks [21]. In this study, we revealed a cell-type-specific m6A modification pattern in elderly patients with AML involving the distinct roles of FTO in different cell types, including HSCs, Myeloids and TCells. Importantly, we observed that the elevated expression of YTHDF2 in Erythrocytes is involved in regulating erythrocyte differentiation. Additionally, CellChat analysis confirmed alterations in ligand-receptor pairs. The objective of this study is to identify novel therapeutic targets and to facilitate the development of more comprehensive treatment strategies for AML patients. Our study also aims to provide new approaches for enhancing clinical responses to immunotherapy, characterizing the tumor immune phenotypes of diverse cell types, and advancing personalized immunotherapy for AML. Results Distribution and reduction profiles of m6A regulator pattern through single-cell transcriptome atlas in AML patients To understand the cellular diversity and molecular features in AML patients, we used standard bone marrow (BM) aspiration methods to isolate hematocytes from the BM of three AML patients and two control individuals. We analyzed scRNA-seq datasets consisting of 54,809 cells, of which 42,953 were derived from AML samples. To visualize the cluster analysis, we performed uniform manifold approximation and projection (UMAP) and t-distributed stochastic neighbor embedding (t-SNE) using the RunTSNE function in Seurat. The visualizations included the sample type of origin (malignant or non-malignant), the corresponding patient, the associated cell type, and the unique molecular identifier (UMI) detected in each cell (Fig. 1a). Five known cell types were identified based on previously reported cell markers and other AML single-cell sequencing results [22], including 18,491 HSCs corresponding to cell markers CD34, SPINK2, and CYTL1, 20,274 Myeloids corresponding to cell markers LYZ, CST3, and MS4A7, 13,640 erythrocytes corresponding to cell markers HBB, HBA1, and GATA1, 1459 TCells corresponding to cell markers CD3D, CD3E, and TRAC, and 945 B cells corresponding to cell markers MS4A1, CD79A, and CD79B (Fig. 1b) (Fig. S1). Analysis of cell proportions indicated that HSCs, erythroblasts, macrophages (Mps), and B cells were more abundant in the AML patient group (Fig. 1c). Diverse cell types in the AML delineated by Single-cell RNA-seq(scRNA-seq) analysis. a t-Distributed Stochastic Neighbor Embedding(tSNE) of the 54,809 cells profiled here, with each cell color coded for (left to right): its sample type of origin (malignant or non-malignant), the corresponding patient, the associated cell type and the unique molecular identifier (UMI) detected in that cell (log scale as defined in the inset). b The plot shows identified cell types and annotated using classical marker genes. c Data of the five cell clusters of 54,809 cells from 5 samples (from left to right): the fraction of cells originating from each patient. d Differential expression of genes in different cell types of AML patients compared with control samples. e and g Kaplan - Meier curves showing progression-free survival in GEPIA 2 in AML Cohort stratified according to high vs. low expression of NHP2 (e) and PLIN5 (g). f and h Heatmap showing the representative gene ontology enriched in upregulated (f) or downregulated (h) genes in each cell type (p < 0.05). i Differential expression of N6-methyladenosine (m6A) regulators in different cell types of AML patients compared with control samples from respective cell-type assignments; The size of the dots indicates the average multiple of difference, and the color of the dots indicates up-regulated (red) or down-regulated (purple) To further explore how diverse cell types impact AML progression, we examined transcriptional alterations in AML-derived cell types. We found that differentially expressed genes (DEGs) in erythrocytes were predominantly upregulated in AML samples compared to healthy samples, while others were downregulated (Fig. 1d). In AML compared to normal group, the top dysregulated genes across the five diverse cell types were identified (Fig. S1). We also identified DEGs in AML samples associated with patient prognosis. High-risk scores based on NHP2 expression correlated with lower survival rates (Fig. 1e), while higher PLIN5 expression correlated with higher survival rates (Fig. 1g). Notably, functional enrichment analysis revealed that hematopoietic stem cells (HSCs) in AML samples were associated with chemical synaptic transmission and mitochondrial homeostasis. Erythrocytes were linked to the regulation of dendrite extension and the positive regulation of protein export from the nucleus. T cells were associated with chromosome segregation, while B cells were linked to the apoptotic signaling pathway (Fig. 1f and h). To understand the cellular features and mechanisms of m6A regulators in the bone marrow (BM) of AML patients, we annotated using classical marker genes. The marker genes for the cell populations are clearly shown. We then analyzed the expression levels of different m6A regulators (Fig. S3a) and defined 23 regulator signatures using the AddModuleScore. The expression of m6A regulators across diverse cell types was illustrated using a UMAP plot based on gene expression profiles. Additionally, we measured the percentage of m6A expression in each cell type in AML (Fig. S2). To elucidate the mechanism of m6A regulation in diverse cell types in AML patient samples, we identified DEGs of m6A regulators. We found that FTO, YTHDF2, and IGF2BP2 were mainly upregulated, while WTAP, ALKBH5, and YTHDF1 were mainly downregulated in diverse cell types of AML (Fig. 1i). These findings suggest the intricate role of dysregulated m6A regulator function in AML progression. m6A regulators are involved in malignant differentiation in diverse cell types in AML patients To define gene expression changes globally and at the cellular level, we performed bulk RNA-seq on AML and control samples. Unsupervised clustering revealed two distinct patterns (Fig. 2a). In AML samples, HNANPC, IGF2BP2, METTL3, and LRPPRC were upregulated, while FMR1 was downregulated (Fig. 2b). Expression patterns of m6A regulator in AML patients and healthy individuals. a Principal Component Analysis (PCA) was employed on 23 m6A regulators to differentiate between tumor and normal samples. b Volcano-plot representation of differentially expressed genes (DEGs). Red, up-regulation; blue, down-regulation. c Heatmap of the DEGs between the gene clusters and different clinical data was shown in the annotation. d Expression patterns of m6A regulators in AML patients and healthy individuals. Heatmap of DEGs between AML and control groups; the m6A regulators in each module were annotated; the line graph showed the trend in the gene module expression, the text on the right showed the enriched pathways for each module gene. e The DEGs patterns of m6A regulators of diverse cell types. f Gene set variation analysis (GSVA) enrichment analysis showing the association between m6A regulators and HSC differentiation in AML. g Spearman’s correlation was used to analyze the correlation between the m6A regulators and classical AML related pathways. Red, positive correlation; blue, negative correlation To explore these m6A modification phenotypes, we focused on control and AML groups. Unsupervised clustering showed distinct patterns of m6A regulator expression (Fig. 2c). Further analysis across four AML subtypes (C1-C4) revealed that C2 contained IGF2BP3, RBM15, HNRNPC, IGF2BP2, METTL14, RBM15B, and IGF2BP1; C3 contained YTHDF2, FMR1, CBLL1, YTHDC2, YTHDC1, and YTHDF1; C1 contained HNRNPA2B1, WTAP, YTHDF3, FTO, METTL3, and LRPPRC; and C4 contained VIRMA and ALKBH5. Based on mRNA profiles, we examined differences in m6A regulators between clusters C2 and C3, focusing on pathways related to chromosomal abnormalities and cellular immune function. We found that most pathways related to immune response activation, such as those involving immune response-activating cells and signal transduction, were downregulated in C3. This suggests that the dysfunction in immune pathways in C3 may be due to m6A regulators like YTHDF2 (Fig. 2d). Next, to reveal the differential expression of m6A molecules between the AML and control groups, we found that FTO was overexpressed in all cell types, YTHDF2 in erythrocytes, and IGF2BP2 in myeloid cells (Fig. 2e). To study the link between m6A regulation and AML progression, we performed GSEA-based functional enrichment analysis and correlated m6A regulators with AML-related pathways. Our analysis showed that upregulated m6A regulator genes were associated with HSC differentiation, the p53 pathway, ribonucleoprotein complexes, T-cell differentiation, monocyte differentiation, AML activation, and the Wnt pathway (r > 0.2). Conversely, downregulated m6A regulator genes were associated with negative senescence, negative erythrocyte differentiation, and leukocyte migration (r < −0.2) (Fig. 2g). Additionally, our findings were supported by further analyses presented (Fig. S3c and Fig. S3d), which detail the correlations between m6A regulators and specific AML pathways (Fig. 2f). FTO cell-type-specific upregulation impaired haemocyte differentiation and suppressed oxidative phosphorylation metabolism via WNT signaling in HSCs To further investigate the distinctions between the identified HSC subpopulations in the pathological process of elderly AML, we compared the expression levels of the top 20 marker genes in HSCs between normal and AML samples. All HSCs were further classified into six distinct subclusters (Fig. 3a). We found six signature genes (PCLAF, TYMS, TOP2A, ASPM, ATP8B4 and AHNAK) that were significantly upregulated in HSCs. The upregulation of PCLAF, TYMS, TOP2A, and ASPM began at the malignant stage and was significantly higher in subclusters C1, C3, and C4, while the benign marker genes ATP8B4 and AHNAK were significantly higher in subcluster C2 (Fig. 3b). To investigate the differences between the identified HSC subpopulations, we analyzed copy number aberrations (CNA) using InferCNV. Malignant subclusters C1, C3, and C4 exhibited copy number gains on chromosome 1, whereas subcluster C2 had low CNA levels, indicating a well-differentiated state (Fig. 3c). Further analysis using AUCell revealed that m6A regulators play crucial roles in regulating malignant HSCs, excluding subcluster C2 (Fig. 3d). Overall, we concluded that m6A regulator genes, except for ATP8B4 and AHNAK, could serve as markers for AML progression. FTO upregulated caused impaired hemocyte differentiation in malignant haematopoietic stem cells (HSCs). a tSNE plot showing HSC subclusters. b Relative expression levels of the PCLAF, TYMS, TOP2A, ASPM, ATP8B4 and AHNAK gene across 5 HSC subclusters. c Inferred Copy Number Variants (CNV) levels in each HSC cluster across 22 chromosomes. d The correlation between the key genes(ATP8B4 and AHNAK) of HSC subclusters2 and m6A score signatures in HSCs. e tSNE plot showing the benign and malignant HSC subclusters. f Comparison of cumulative probability of m6A score signatures between benign and malignant in HSCs. (Wilcox test, p < 0.001). g and h According to the expression level of FTO in HSCs, the patients were divided into two groups; red lines represent stronger communication in FTO high expression group, and blue lines represent weaker communication in FTO high expression group (left), rank signaling networks based on the information flow or the number of interactions (right). i and j The correlation analysis between the key genes or pathways and m6A regulators expression in HSCs. k and l The genes of HSCs proliferation and negative regulation of cellular senescence had a significant correlation with FTO expression level; m Spearman correlation between FTO and HSCs proliferation. n Networks of weighted gene co-expression network analysis (WGCNA) module which included FTO in HSCs. o Enrichment analysis for FTO module by Gene Ontology (GO) enrichment Next, using reclustering and UMAP dimensionality reduction techniques, we observed greater heterogeneity in both benign and malignant cells in AML samples compared to control samples (Fig. 3e). To understand specific changes in m6A regulator expression in HSCs, A previous study assessed 23 regulators of m6A modification and further constructed a regulator signature, named the m6A score [23]. We found that the m6A score was upregulated in malignant HSCs (Fig. 3f), especially observed for FTO. To explore the mechanisms underlying HSC-FTO + cell communication in AML, CellChat was used to examine the interactions between FTO expression in HSCs and other cell types (Fig. 3g). It was observed that the FTO low-expression group displayed significant interactions with specific genes in other cell types in patients with AML (Fig. 3h). Moreover, we performed correlation analysis between m6A regulator expression levels and AML pathways. The results showed that FTO upregulation positively regulated the Wnt pathway but negatively regulated and correlated with monocyte differentiation, glycoprotein metabolic processes, MHC II receptor activity, and T-cell differentiation (Fig. 3j). We also revealed significant negative associations between FTO expression and monocyte differentiation (Fig. S3i) and significant positive associations between FTO expression and HSC proliferation (r = 0.29) (Fig. 3m). Notably, key genes involved in monocyte differentiation and HSC proliferation, such as KAT7 and N4BP2L2, were positively correlated with FTO expression (Fig. 3k and Fig. S3g). Key regulators of cellular senescence, BCL6 and SIRT1, were also positively correlated with FTO (Fig. 3l). Furthermore, key genes that positively regulate HSC-mediated dysfunction in elderly AML patients were positively correlated with FTO expression (Fig. 3i). Additionally, we found that WTAP upregulation was positively associated with dysregulation in HSCs (r = −0.42) (Fig. S3e, f, and h). To further unravel the regulatory roles of the m6A regulator gene FTO in HSCs, we performed DEGs analysis and identified 20 genes related to FTO based on gene network biological validity and gene-gene interaction relevance. Weighted Gene Co-expression Network Analysis (WGCNA) showed that these genes were closely related to FTO expression in HSCs of AML samples (Fig. 3n). Functional enrichment analysis revealed that these genes were primarily enriched in hematopoietic cell proliferation and differentiation in AML HSCs (Fig. 3o), suggesting that FTO is a potentially important player in HSC regulation in AML. YTHDF2 cell-type-specific upregulation in erythrocytes caused negative differentiation through oxidative phosphorylation To investigate changes in erythrocyte and erythroblast proportions between AML patients and controls, We observed substantial heterogeneity in AML samples by UMAP (Fig. 4a). We also identified YTHDF2 as a key gene in regulating erythrocyte processes in elderly AML patients. We determined the expression of YTHDF2 was higher in Erythrocytes by MetaCell (Fig. 4b). To explore the cell-type-specific expression pattern of YTHDF2 dysregulation-triggered differentiation dysfunction, correlation showed that YTHDF2 negatively regulated erythrocyte differentiation, cellular senescence (Fig. 4d and e), and erythrocyte homeostasis (r=−0.47) (Fig. 4c), suggesting that YTHDF2 cell-type-specific upregulation caused dysfunction of Erythrocytes. Key regulators of the negative regulation of erythrocyte differentiation, including HSPA9, LDB1, and STAT5B, were positively correlated with YTHDF2 expression (Fig. 4d). Key regulators of cellular senescence, were also positively correlated (Fig. 4e). Next, WGCNA was also used to identify the role of YTHDF2 in erythrocytes from elderly AML patients (Fig. 4f). Metascape analysis showed these genes were linked to leukocyte activation, cancer pathways, and phosphate metabolism (Fig. 4g). This suggests YTHDF2 upregulation causes dysfunctional erythrocyte differentiation and affects leukocyte activation in elderly AML patients. Specific upregulation of YTHDF2 in Erythrocytes resulted in negative regulation of erythrocyte differentiation. a Reclustering of erythrocytes and erythroblasts reveals different distribution of cells. b The correlation between the key genes and m6A regulators in erythrocytes. c Spearman correlation between YTHDF2 and erythrocyte homeostasis. d and e The genes of erythrocyte differentiation and cellular senescence had a significant correlation with YTHDF2 expression level. f Networks of WGCNA module which included YTHDF2 in Erythrocytes. g And enrichment analysis on YTHDF2 module by Metascape. h and i Metabolic pathway activities in Erythrocytes via YTHDF2 high and low expression. Glycolysis in Erythrocytes via YTHDF2 high and low expression. Statistically non-significant values are shown as blank. j Communication and ligand-receptor interaction between erythrocyte and neighboring cells showing in the dotplot (high YTHDF2 versus low YTHDF2). k and l According to the expression level of YTHDF2 in Erythrocytes, the patients were divided into two groups; red lines represent stronger communication in YTHDF2 high expression group, and blue lines represent weaker communication in YTHDF2 high expression group (left), rank signaling networks based on the information flow or the number of interactions (right) To quantify the activity of metabolic pathways affected by YTHDF2 upregulation in erythrocytes, we used a pathway activity heatmap showing the correlation. We found that YTHDF2 upregulation positively regulated oxidative phosphorylation, glycolysis/gluconeogenesis, folate biosynthesis, and glycosaminoglycan biosynthesis (Fig. 4h). Notably, key genes involved in glycolysis/gluconeogenesis, including PGK1, DHA, ENO1, TPI1, GAPDH, and LDHB, were upregulated by YTHDF2 upregulation (Fig. 4i). To explore the mechanisms underlying erythrocyte-YTHDF2 + cell communication in AML, CellChat was used to examine the interaction between YTHDF2 expression in Erythrocytes and other cells (Fig. 4j and k), and it was observed that the YTHDF2 high-expression group displayed significant interactions with specific genes (Fig. 4l). FTO cell-type-specific upregulation in Myeloids and TCells caused negative regulation of haematopoietic progenitor cell and T-cell-mediated immunity To explore the heterogeneity in myeloid cells in elderly AML patients, we identified patterns using UMAP (Fig. 5a). To determine the causes of dysregulation in myeloid cells by m6A regulators, we performed correlation analysis between m6A expression and AML pathways. We found that key genes positively regulating myeloid-mediated dysfunction in elderly AML patients were positively correlated with FTO expression (Fig. 5b). The results showed that FTO upregulation positively regulated HSC abnormal proliferation and cell aging, while negatively regulating and correlating with pathways such as myeloid cell differentiation, HSC differentiation, apoptotic signaling pathways, and immune response in AML (Fig. 5f). Specifically, FTO upregulation negatively regulated hematopoietic progenitor cell differentiation (r = 0.44) (Fig. 5d) and myeloid differentiation and the p53 pathway (r = −0.5) (Fig. 5e). The expression of key genes involved in negative hematopoietic progenitor differentiation was positively correlated with FTO (Fig. 5c). FTO and IGF2BP2 promoted the increase of pathological myeloid differentiation. a Reclustering of Myeloids reveals different distributions of malignant and normal cells. b The correlation between the key genes and m6A regulators in Myeloids. c The genes of negative regulation of haematopoietic progenitor cell differentiation had a significant correlation with FTO expression level. d and e Spearman correlation between FTO and negative regulation of haematopoietic progenitor cell differentiation, negative regulation of apoptosis pathway. f The correlation analysis between the key pathways and m6A regulators expression in Myeloids. g WGCNA identifies brown modules with correlated gene expression patterns in Myeloids of AML patients. h Enrichment analysis for FTO module by GO enrichment. i The genes of chromosomal region had a significant correlation with IGF2BP2 expression level. j Spearman correlation between IGF2BP2 and chromosomal region. k Networks of PINK module which included IGF2BP2 in Myeloids. l And enrichment analysis for PINK module by GO enrichment. m Metabolic pathway activities in Myeloids via IGF2BP2 high and low expression To elucidate cell-specific ligand-receptor pairs between myeloid cells and other cell types, CellChat analysis indicated that FTO played a regulatory role (Fig. S3j). To identify the function of FTO upregulation in myeloid cells, we divided the WGCNA results into nine modules (Fig. 5g). FTO was identified in the brown module in myeloid cells. Functional enrichment analysis revealed that this module was involved in the regulation of the apoptotic process and negative regulation of myeloid cell differentiation in AML samples (Fig. 5h). Next, we found that IGF2BP2 upregulation induced dysfunctional chromosomal regions (r = 0.4) involved in myeloid cells in AML (Fig. 5i and j). WGCNA showed that 20 genes were closely related to IGF2BP2 in AML myeloid cells (Fig. 5k). Functional enrichment analysis indicated that IGF2BP2 upregulation negatively regulated erythroid differentiation in myeloid cells in AML samples (Fig. 5l). The metabolic pathways dysregulated by IGF2BP2 upregulation in myeloid cells showed that IGF2BP2 upregulation was also positively regulated by oxidative phosphorylation and glycolysis/gluconeogenesis (Fig. 5m). In addition, we determined that FTO was a key gene in the positive regulation of granulocyte-monocyte progenitor (GMP)-mediated processes (Fig. S3k and Fig. S3l). To examine the expression pattern of FTO in T cells, we found that FTO upregulation positively correlated with cellular senescence (r = 0.41) but negatively affected T-cell-mediated immunity (r = −0.4). WGCNA identified biological functions involved in the negative regulation of immune cell differentiation, suggesting that FTO upregulation leads to dysfunctional immune responses in T cells. We also identified ligand-target interactions between T cells and HSCs, highlighting the role of HMGB1-ligands in their communication using NicheNet (Fig. S4). Heterogeneity transitional trajectory and dysfunctional differentiation of haemocyte regulatory networks in AML To elucidate the mechanism of the transition from benign to malignant HSCs, pseudotime analysis was performed on five cell clusters in AML samples (Fig. 6a). For HSCs, a large proportion of cells were in an initial state of differentiation from benign to malignant HSCs (Fig. 6b). Notably, subcluster 1 cells represented malignant HSCs with impaired differentiation, and there was an increase in the number of subcluster 1 cells within the HSC population (Fig. 6d). Additionally, FTO had higher expression levels during the initial stage of disease (Fig. 6c). Pseudotime analysis predicted that the majority of malignant cells originated from initial HSCs (Fig. 6e), and that the majority of malignant HSCs originating from subcluster 1 were dysregulated by FTO, with a few benign cells differentiating into other cell clusters (Fig. 6f and g). Gene Ontology (GO) enrichment analysis of upregulated genes in the initial stage (Fig. 6h) revealed that subcluster genes were involved in pathways such as mRNA methylation (Fig. 6i). Overall, these results contribute to the understanding of the hierarchical structure and heterogeneity of malignant HSCs dysregulated by FTO in AML. m6A regulators dysregulated cytopathologic differentiation in diverse cell types. a Pseudo-time trajectory of diverse cell type, b and Pseudo-time trajectory of cell of AML with gene expression profiles inferred by Monocle 2. Each point corresponds to a single cell. c Fitted curves showing dynamic expression changes in representative m6A genes in each gene group based on AML progressing time. d Pseudotime analysis of each cell type. HSCs, Myeloids, Erythrocytes, TCells and BCells were divided into relatively malignant (Pseudotime < 5) and benign (Pseudotime ≥ 5) groups. e The proportions of HSCs, Myeloids, Erythrocytes, TCells and BCells; pseudo-time was shown on the horizontal axis. f Heatmap of m6A regulator expression profiles based on pseudo-time trajectory of 3 cluster cells of AML. g Heatmap of key genes of 3 cluster cells profiles based on pseudo-time is indicated on the horizontal axis. h Heatmap representing the smoothed expression of pseudo-time-dependent genes along pseudo-timeline of AML cell fate branch; Pseudo-time-dependent genes were grouped into three clusters according to different expression patterns, and the m6A regulators included in each module were annotated. i Bar plot of key genes of HSCs module, Myeloids module, Erythrocytes module enriched pathways. j NicheNet was used to analyze the expression of ligands and receptors to identify the intercellular communication patterns between Erythrocytes and HSCs Similarly, pseudotime analysis was performed on erythrocytes, and subcluster identification was conducted to investigate the transition of YTHDF2-expressing erythrocyte differentiation (Fig. 6c). YTHDF2 had higher expression in subcluster 1 compared to other subclusters (Fig. 6f), and was primarily enriched in functions related to erythrocyte differentiation and erythrocyte homeostasis (Fig. 6i). Pseudotime analysis was also conducted on myeloid cells (Fig. 6a and b). FTO and IGF2BP2 had higher expression levels during the intermediate stage (Fig. 6c). The AML group had a higher number of fate 2 cells compared to the control group (Fig. 6d and h), which were enriched in the positive regulation of leukocyte activation (Fig. 6i). Moreover, we found that FTO, YTHDF2 and IGF2BP2 clustered together in subcluster 1 (Fig. 6f), and common oncogenes in AML, such as IGKC, CXCL2, SOX4 and MYC (Fig. 6g). To elucidate cell-specific ligand receptors among diverse cell types across m6A regulators in AML, intercellular communication between Erythrocytes and HSCs was examined using NicheNet. Ligand activity prediction was performed to identify the most likely YTHDF2-regulating ligands. Notably, one of the top-ranked ligands was HMGB1. Subsequently, we assessed the extent to which HMGB1-ligands may regulate aberrant erythrocyte differentiation by the 20 top-ranked HMGB1-ligands can predict dedifferentiation-related genes. Finally, we identified the interactions between HMGB1 ligands and erythrocyte differentiation genes in AML (Fig. 6j). Overall, we identified the potential active ligand-target links that played specific dedifferentiation regulatory roles of HMGB1-ligands that were dysregulated by YTHDF2. Additionally, NicheNet analysis was performed to identify ligand-target links between Myeloids and others. Similarly, NichNet was used to identify potential active ligand–target links between HSCs and Myeloids, ligand–target links between TCells and Myeloids. Taken together, results suggest the diverse potential of m6A regulators in the regulatory function of specific types of AML (Fig. S5). The validation of mechanism for m6A regulator cell-type-specific upregulation in diverse cell types of AML by scRNA-Seq data, and experiment test in vitro of cell function and oxidative phosphorylation To verify cellular diversity and m6A regulator expression in AML, we analyzed 17 samples with scRNA-Seq (GSE235923). After quality control, 35,203 cells were retained, including HSCs, myeloid cells, erythrocytes, T cells, B cells, and plasma cells (Fig. S6a). Several m6A regulators, showed high expression levels and high proportion in Erythrocytes, HSCs and myeloid cells. Reclustered the expression of YTHDF2 and IGF2BP2 (Fig. S6b, c, d and e), Calculating an m6A score in diverse cell types, suggesting abnormal immune cell differentiation (Fig. S6f). To validate the cell-type-specific mechanisms of m6A regulators in AML, FTO in HSCs were defined as ‘malignant HSC’ (MHSC) signature. FTO and MHSC showed a positive correlation. YTHDF2 in erythrocytes were defined as ‘malignant erythrocytes’ (MERY) signature. YTHDF2 and MERY showed a positive correlation. FTO in myeloid cells were defined as ‘malignant myeloid cells’ (MMYE) signature. FTO and MMYE showed a positive correlation. IGF2BP2 and SMC4 expression showed a positive correlation. (Fig. S6g, h, i and j). The results demonstrated that FTO, YTHDF2, and IGF2BP2 play important cell-specific regulatory roles in AML. To verify cell-type-specific m6A regulators in AML, we used MetaCell and WGCNA. Results showed 25 genes related to YTHDF2 in erythrocytes, indicated these genes affect immune response. Overall, through another datasets of AML, we proved FTO was specifically upregulated in HSCs. YTHDF2 was upregulated in erythrocytes. FTO and IGF2BP2 were upregulated in myeloid cells (Fig. S6k and Fig. S6l). To further verify the role of FTO and its downstream-related mechanisms in AML cells in vitro, we constructed shRNA-FTO−1(sh1) and shRNA-FTO−2(sh2) plasmids for transfection into the THP1 and MV411 cell lines, to measure the cell proliferation, migration and apoptosis. The efficiency of silencing was verified by Quantitative Real-Time Polymerase Chain Reaction (qRT–PCR) after the plasmid transfection is complete (Fig. 7a). To measure the cell proliferation, Results from the Cell Counting Kit-8 (CCK-8) assay demonstrated that the silencing of FTO effectively inhibits the survival rate of THP-1 and MV411 cells (p < 0.001) (Fig. 7b). To begin exploring the mechanisms by which FTO might regulate cell cycle progression in AML cells, we performed cell cycle analysis of THP-1 and MV411 cells by using flow cytometry, cell cycle analysis revealed that silencing FTO led to a decreased percentage of THP-1 and MV411 cells in G0-G1 phase and an increased percentage of cells in S phase compared to the control group (Fig. 7c and d). To investigate the apoptotic effects of FTO on AML cells, We used an apoptosis kit to assess the apoptotic effects of FTO on THP-1 and MV411 cells by flow cytometry. FTO flow cytometer analysis using Annexin V/PI double staining showed a significant increase in the apoptosis rate of cells transfected with sh-FTO (p < 0.001) (Fig. 7e and f). Additionally, to determine the effects of FTO on the invasion of AML cells, transwell assays were conducted in THP-1 and MV411 cells after knockdown of FTO. The results showed that downregulation of FTO significantly attenuated THP-1 and MV411 cells migration and invasion (p < 0.001) (Fig. S7a and Fig. S7b). These results suggested that FTO exhibited important regulation of cell proliferation, cell cycle progression, apoptosis, and cell invasion in THP-1 and MV411 cells. FTO regulated THP-1 and MV411 cells proliferation and oxidative phosphorylation, while inhibited cell cycle arrest and apoptosis. a Quantitative Real-Time Polymerase Chain Reaction (qRT–PCR) analysis of FTO expression in THP-1 and MV411 cells transfected with sh-NC and sh-FTO expression vector. b The growth curves of cells transfected with indicated vectors were evaluated by Cell Counting Kit-8 (CCK-8) assays. c and d The cell cycle progression was analyzed by flow cytometer after being transfected with indicated plasmids. e and f Apoptosis rate was analyzed by flow cytometer after being transfected with sh-NC and sh-FTO expression vector. g and h WNT1 and YTHDF2 expression in THP-1 and MV411 cells transfected with sh-NC and sh-FTO expression vector, ** and *** represent p < 0.01 and p < 0.001, respectively (THP-1 cell on the left and MV411 cell on the right of the bar graphs) To verify the association between FTO and oxidative stress in vitro experiment in AML, the expression level of reactive oxygen species (ROS) in THP-1 and MV411 cells after knockdown of FTO was examined through the ROS assay kit. The results showed that ROS levels were attenuated in both THP-1 and MV411 cells after knockdown of FTO (Fig. S7c). To evaluate the regulatory role of the FTO/WNT pathway and YTHDF2 in oxidative phosphorylation, we performed the PCR to analyse the expression of WNT1 and YTHDF2 after knockdown of FTO. The results showed that knockdown of FTO resulted in lower expression of WNT1 and YTHDF2 in THP1 and MV411 cells than in the non-knockdown group (p < 0.001) (Fig. 7g and h). The results of RNA quantification revealed that knocked down FTO expression in THP-1 and MV411 cells was associated with m6A levels. Overall, FTO exerts significant regulatory effects on the proliferation, invasion, apoptosis, oxidative phosphorylation, m6A regulation and downstream tumor-related signaling pathways in AML cells. Discussion Although considerable progress has been made to alter proliferation/apoptosis, signal transduction, epigenetic modification and immunity in AML, patients still have a poor prognosis and a high mortality rate due to the complexity of the pathogenesis of elderly AML [15, 24–27]. Recently, investigations into epigenetic modification of RNA at the post-transcriptional level have attracted widespread attention, and m6A modification, which is the most common RNA modification, has been extensively studied in the occurrence and development of AML [28]. ScRNA-seq was used to identify differences in m6A modification between AML patients and healthy controls. Our findings showed that m6A regulation varies among different cell types in elderly AML patients. FTO upregulation in HSCs, Myeloids, and T-cells inhibited their differentiation through the WNT pathway. YTHDF2 was elevated in erythrocytes, negatively regulating differentiation via oxidative phmyeloidsosphorylation and causing leukocyte activation. IGF2BP2 was upregulated in, leading to dysfunctional chromosomal regions and dysregulated oxidative phosphorylation. Further analysis with CellChat confirmed changes in ligand-receptor interactions mediated by m6A regulators, activating the HMGB1 pathway and promoting AML progression. In vitro experiments using shRNA targeting FTO demonstrated inhibition of cell proliferation, migration, and oxidative phosphorylation, as well as induction of cell cycle arrest and apoptosis. FTO is an RNA m6A demethylase that belongs to the Alkb protein family [29]. Previous studies have shown that FTO is abnormally upregulated in certain subtypes of AML, such as t (11q23)/MLL rearrangement, t (15; 17), FLT3-ITD, and/or NPM1-mutated AML [30]. Our results further demonstrated that FTO was highly expressed in HSCs and increased with the differentiation of HSCs. In addition, we explained the function of FTO in HSCs. FTO was positively correlated with HSC proliferation genes (KAT7 and N4BP2L2) and negatively correlated with monocyte differentiation genes (MYC, IFI16, HLA-DRB1, and CD74), suggesting that FTO can promote HSC proliferation and inhibit HSC monocyte differentiation. Li et al. previously reported that FTO downregulates the expression of ASB2 and RARA by reducing their m6A levels, thereby promoting the survival/proliferation of AML cells and inhibiting differentiation and apoptosis [30]. In addition, Li et al. demonstrated that this axis can inhibit the differentiation of AML cells induced by ATRA [30]. Subsequently, Su et al. demonstrated that FTO can also promote the proliferation of leukaemia cells through FTO/m6A/MYC/CEBPA signaling [11]. We similarly observed positive correlations between FTO and N4BP2L2 and RUNX1, indicating that FTO was also involved in the WNT signaling pathway in HSCs. Previous studies have demonstrated that in multiple myeloma, IDH2 enhances the expression of WNT7B and activates the WNT signaling pathway by targeting the m6A demethylase FTO, thereby promoting the occurrence and development of multiple myeloma [31]. The importance of FTO in regulating WNT signalling pathways has also been reported in solid tumors, which was consistent with our research results [32–34]. Intriguingly, our study revealed that FTO was positively correlated with BCL6 and SIRT1, suggesting its potential involvement in regulating cellular senescence. Li et al. substantiated this point by demonstrating that FTO induces endothelial cellular senescence [35]. We also proved that FTO was highly expressed in Myeloids, and inhibited haematopoietic progenitor cell differentiation, myeloid differentiation, p53 signaling pathway. Our observations revealed a negative correlation between FTO and the p53 signaling pathway, suggesting that FTO may regulate the proliferation and differentiation of hematopoietic stem/progenitor cells through this pathway. Furthermore, FTO exhibited an impact on immune responses in Myeloids, influencing the therapeutic resistance of AML. The use of small-molecule inhibitors targeting FTO affected the expression levels of downstream genes such as MYC, CEBPA, RARA, and ASB2, thereby exerting anti-leukaemia effects [36]. Moreover, FTO can also inhibit the expression of immune checkpoint genes, especially LILRB4, significantly weakening the self-renewal and reprogramming immune response of leukaemia stem cells/initiating cells, which has been proven to be an effective treatment strategy for AML [36, 37]. Additionally, the upregulation of IGF2BP2 in Myeloids induced a dysfunctional chromosomal region, which is influenced by oxidative phosphorylation and glycolysis/gluconeogenesis. In addition, we serendipitously found significant differences in the Erythrocytes between patients and healthy individuals during cluster analysis, especially with a significant increase in erythroblasts in patients. Previous studies have shown that erythroid progenitor cells derived from tumor hosts lose their erythrocyte potential and differentiate into myeloid cells, named erythroid-differentiated myeloid cells (EDMCs) [18, 38]. Our study suggested that the binding protein YTHDF2 of m6A was involved in the proliferation and differentiation of erythrocytes, regulating erythrocyte homeostasis through oxidative phosphorylation, and YTHDF2 was highly expressed in Erythrocytes rather than in other types of cells. YTHDF2 is significantly more highly expressed in AML samples with multiple cytogenetic abnormalities and is essential for AML cell survival and leukaemia cell implantation, as reported by Paris et al. [12]. Moreover, YTHDF2 reduces the half-life of various m6A transcripts, which are crucial to the functional maintenance of LSCs [12]. Paris et al. also believed that YTHDF2 was not important for steady haematopoiesis, while Mapperley et al. reached the opposite conclusion [12, 39]. It is necessary to further clarify the role of YTHDF2 in HSCs and its function in regulating the deterioration of HSCs in the future. To further validate the function of FTO, we conducted FTO silencing experiments in THP-1 cells. Our findings suggest that silencing FTO suppresses proliferation, induces cell cycle arrest, and promotes apoptosis in THP-1. These results, combined with previous studies, underscore the pro-leukemic role of FTO in vitro [11, 30, 40]. To explore the cell-specific ligand-receptor pairs among diverse cell types, we used NicheNet to examine intercellular communication, and the results showed that HMGB1 ligands dysregulated by YTHDF2 were related to erythrocyte differentiation genes in AML. HMGB1 ligands are also involved in the communication between Myeloids/TCells and other cell types. HMGB1 ligands may activate proinflammatory pathways mediated by IL1β in the HSC-to-myeloid cell transition. This finding is consistent with recent observations that HMGB1 is highly expressed in BM mononuclear cells of AML patients, which affects the activity of the downstream gene TGFBI and the expression levels of TNF-α, IL-1β and SCF, blocking AML cell proliferation and myeloid cell differentiation [41–43]. The study has some limitations. Our research primarily focused on how AML affects blood cell differentiation and the role of m6A regulators in HSCs, myeloid cells, and T-cells, specifically how m6A regulators inhibits their differentiation via various signaling pathway. However, we did not cover the full systemic impact of the disease or conduct animal experiments to validate the impact of cell-cell ligand receptor interactions on AML progression. Additionally, we did not study cell damage caused by AML progression at the spatiotemporal level in diverse cell types, which was due to the small sample size. Moving forward, we plan to investigate more m6A regulators and their effects on cell types involved in AML. We will also perform animal experiments to validate the impact of cell-cell interactions and investigate the full-body effects of AML using multi-omics approaches and clinical data analysis. In conclusion, we showed that the m6A score could be used to assess patients’ clinicopathological subtype features of diverse cell types, including stages of CNV, tumor differentiation levels, and molecular subtypes. Detailed relationships between m6A score and clinicopathological features of AML were revealed in our study. We could also predict the efficacy of chemotherapy and patient clinical response to immunotherapy using m6A regulators, FTO was upregulated in HSCs, Myeloids and TCells, which inhibited the differentiation of diverse cell types through the WNT pathway. YTHDF2 upregulation in Erythrocytes induced the negative regulation of differentiation through oxidative phosphorylation, leading to leukocyte activation. m6A regulators induced aberrant cell-cell communication in haemocytes and mediated ligand-receptor interactions across diverse cell types. Our study provided a new avenue for improving patient clinical responses to immunotherapy, identifying distinct tumor immune phenotypes, and promoting personalized AML immunotherapy. Moving forward, we plan to investigate additional m6A regulators and their effects on various cell types involved in AML. Through these endeavors, we aim to uncover new therapeutic targets and facilitate the development of more comprehensive treatment plans for AML patients. Overall, the findings of the present study provide a novel approach for developing drug targets for patients with AML. Materials and methods Dataset source and preprocessing Data for three elderly AML patients and two healthy donors enrolled in this study were sourced from the Department of Hematology at the Third Hospital of Shanxi Medical University between January 2021 and March 2022. The study protocol was approved by the Ethics Committee of the Third Hospital of Shanxi Medical University, China (Approval No.: SBQKL-2021-051). Bioinformatics analysis was conducted on single-cell RNA sequencing (scRNA-seq) data derived from bone marrow (BM) blood cell samples collected from patients prior to their receiving any treatment. ScRNA-seq datasets of AML BM cells were obtained from the Gene Expression Omnibus (GEO) database under accession number GSE235923 (accessible at www.ncbi.nlm.nih.gov/GEO). Additionally, bulk RNA-seq datasets for both AML BM cells and healthy donor BM cells were also retrieved from the GEO database (accession number GSE116616). Detailed information regarding all samples, including the time elapsed since diagnosis, sex, age, and mutation profiles, is provided in Supplementary Tables 1 and Supplementary Table 3. The BD Rhapsody Analysis pipeline and the BD AbSeq Ab-Oligos Panel were employed to quantify FASTQ data for the single-cell RNA sequencing (scRNA-seq) analysis of the samples collected for this study. The quantified data were aligned against the human reference genome (hg19) to generate a matrix of raw counts. As part of our quality control measures, we filtered out outliers using a threshold of three times the Median Absolute Deviation (MAD) above or below the median. To correct for differences in gene expression levels among cells, we applied the LogNormalize method for global scaling normalization, standardizing both the characteristic and total expression per cell. Analysis of scRNA-seq data The expression data from the control and AML samples were integrated using the FindIntegrationAnchors and IntegrateData functions within the Seurat package. Normalization was performed using the ScaleData function in Seurat, ensuring that each gene was weighted equally, with a mean of zero and a variance of one. To reduce computational complexity and noise, principal component analysis (PCA) was utilized for initial dimensionality reduction. K-nearest neighbors (K-NN) graphs were constructed using the FindNeighbors function based on the Euclidean distances in the PCA space. Cell clustering was performed using the Louvain algorithm. Finally, the annotated information for each cell was visualized using Uniform Manifold Approximation and Projection (UMAP). Pseudotime trajectory analysis of m6A mRNA regulators in AML cells To investigate the relationship between cell pseudotime trajectories and m6A regulators, we utilized the Monocle R package to analyze single-cell RNA data across all cell types. Highly variable genes were identified based on differential expression between two specific cell types. Dimensionality reduction was performed using the DDRTree method. Subsequently, we employed the plot_pseudotime_heatmap function to visualize heatmaps depicting the dynamic expression of m6A regulators along the pseudotime trajectories of different pulmonary cell types in AML patients. Construction of metacell to conduct DEGs and correlation analysis The concrete details can be found in supplementary information. Cell culture and transfection THP-1 cells and MV411 cells (obtained from ATCC) were cultured in RPMI-1640 medium (Gibco) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (Gibco). 293T cells (donated by the Chinese Academy of Medical Sciences) were maintained in DMEM medium (Gibco) supplemented with 10% FBS. All cell lines used in this study were authenticated by short tandem repeat (STR) analysis to confirm their identity and were found to be mycoplasma-free. All cell cultures were incubated in a constant temperature incubator (Thermo Scientific) at 37 °C in an atmosphere of 5% CO2. FTO short hairpin RNAs (shRNAs) were cloned into the pGPU6/GFP/Neo lentiviral vector (synthesized by GenePharma). All target sequences for the shRNAs are listed in Supplementary Table 2. Lentiviral particles were produced in 293T cells and used to infect THP-1 cells. The efficiency of gene transfer into THP-1 cells was assessed by detecting green fluorescent protein (GFP) using laser confocal microscopy (Olympus) and a flow cytometer (Beckman). Quantitative polymerase chain reaction (qPCR) Total RNA was extracted from cell samples using Trizol reagent (Ambion, USA). The mRNA was then reverse transcribed into cDNA using the cDNA Synthesis SuperMix (TransGen Biotech, Beijing). PCR amplification was conducted using a PCR machine (Bioer Technology, GeneMax Tc-s-B, China) using the following cycling parameters: an initial denaturation at 95 ℃ for 5 min, followed by 40 cycles of denaturation at 95 ℃ for 10 s, and annealing at 60 ℃ for 30 s. Gene expression levels were quantified using the ΔΔCt method. The experiments were performed in triplicates to ensure reproducibility. The primers used for RT-qPCR were as follows: GAPDH forward: 5’-GCTCTCTGCTCCTCCTGTTC-3’; GAPDH forward: 5’-GCTCTCTGCTCCTCCTGTTC-3’; GAPDH reverse: 5’-ACGACCAAATCCGTTGACTC-3’; GAPDH reverse: 5’-ACGACCAAATCCGTTGACTC-3’; FTO forward:5′-AАCАСCAGGCTCTTTACGGTC-3′; FTO forward:5′-AАCАСCAGGCTCTTTACGGTC-3′; FTO reverse:5′-TGTCCGTTGTAGGATGААССС−3′. FTO reverse:5′-TGTCCGTTGTAGGATGААССС−3′. Cell proliferation, apoptosis, cell cycle, and cell migration assays A cell viability assay was performed using the CCK-8 Assay Kit (DOJINDO). Approximately 5,000 cells were seeded per well in a 96-well plate. Apoptosis in THP-1 cells was detected using the Annexin V PE/7-AAD Apoptosis Assay Kit (Bioss) according to the manufacturer’s instructions. Additionally, the cell cycle was analyzed using the Cell Cycle Analysis Kit (Bioss) and evaluated with ModFit software. For cell migration assays, Transwell plates (Corning) were used. Cells were initially incubated in serum-free medium for 24 h. Subsequently, 1 × 105 cells were suspended in 200 µL of serum-free RPMI-1640 medium and added to the upper chamber. Meanwhile, 1 mL of medium containing 20% FBS was placed in the lower chamber. After 24 h of incubation, the migrated cells were photographed and manually counted. Statistical analysis All statistical analyses were conducted using R software (version 4.0). Student’s t test was used to evaluate the statistical significance of differences between two different groups. A p value < 0.05 was considered statistically significant. Supplementary Information Supplementary Material 1. Acknowledgements We would like to thank all members of the R.-J.Z. laboratory for discussion. We are grateful for the contribution of our coworkers involved in conducting the study and writing this article. Abbreviations Haematopoietic stem cells Acute myeloid leukaemia N6-methyladenosine single-cell RNA sequencing short hairpin RNA composite complete remission Overall survival Leukaemia stem cells Leukaemia initiating cells Bone marrow Gene Expression Omnibus Median Absolute Deviation Principal component analysis Uniform manifold approximation and projection k-Nearest Neighbors Weighted gene co-expression network analysis High dimensional weighted gene co-expression network analysis Molecular signature database Differentially expressed genes Green fluorescent protein Quantitative polymerase chain reaction t-distributed stochastic neighbour embedding Unique molecular identifier The Cancer Genome Atlas Granulocyte-monocyte progenitor Copy number aberration Macrophages Cell Counting Kit-8 Gene Set Enrichment Analysis Gene Ontology Authors’ contributions Ruijuan Zhang: Conceptualization, Project administration, Resources, Supervision, Funding acquisition, Writing-review & editing. Zhe Wang: Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft. Xin Du: Investigation, Methodology, Software, Visualization, Writing – original draft. Peidong Zhang: Investigation, Formal analysis, Methodology, Validation, Visualization. Meiling Zhao: Investigation, Writing-review & editing. Tianbo Zhang, Jiang Liu, Xiaolan Wang, Doudou Chang, Xiaxia Liu, Sicheng Bian and Xialin Zhang: Formal analysis, Methodology, Validation, Validation, Writing-original draft. All authors have read and approved the final manuscript. Funding This research was funded by the Basic Research Project of Shanxi Province, China (No. 20210302124037 to X.-L.Z.), Shanxi Bethune Talent Foundation Project (No. 2021RC017 R.-J.Z. and No.2021RC038 to X.-L.Z.), China Postdoctoral Science Foundation (2021M691995 to Z.W.), and The Natural Science Foundation of Shanxi Province, China (202103021224394 to Z.W.). Data availability Single-cell RNA sequencing gene expression data generated in this study has been deposited in the ArrayExpress database with accession of E-MTAB-19818. (https://www.ebi.ac.uk/fg/annotare/edit/19818/#DESIGN:PROTOCOLS). Raw sequencing data and processed data are available through the NCBI Gene Expression Omnibus (GEO) under accession number “GSE116616” and “GSE235923”. All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Declarations Ethics approval and consent to participate This study was performed in line with the principles of the Declaration of Helsinki. All procedures were approved by the Ethics Committee of Shanxi Bethune Hospital (IRB-number: SBQKL-2021-051). Consent for publication Patients signed informed consent regarding publishing their data, including Table S1 and Table S3. Competing interests The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Zhe Wang, Xin Du and Peidong Zhang contributed equally to this work and share first authorship. Contributor Information Xialin Zhang, Email: charlotte007@163.com. Ruijuan Zhang, Email: 13593169668@163.com. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement Single-cell RNA sequencing gene expression data generated in this study has been deposited in the ArrayExpress database with accession of E-MTAB-19818. (https://www.ebi.ac.uk/fg/annotare/edit/19818/#DESIGN:PROTOCOLS). Raw sequencing data and processed data are available through the NCBI Gene Expression Omnibus (GEO) under accession number “GSE116616” and “GSE235923”. All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.