
==== Front
Hereditas
Hereditas
Hereditas
0018-0661
1601-5223
BioMed Central London

39300580
338
10.1186/s41065-024-00338-8
Research
Glutamine metabolism-related genes predict the prognostic risk of acute myeloid leukemia and stratify patients by subtype analysis
Zhou Jie zhoujie55441@163.com

1
Zhang Na 1
Zuo Yan 1
Xu Feng 1
Cheng Lihua 1
Fu Yuanyuan 1
Yang Fudong 1
Shu Min 1
Zhou Mi 1
Zou Wenting 1
Zhang Shengming 13922468988@163.com

2
1 https://ror.org/02sx09p05 grid.470061.4 Department of Hematology, Deyang People’s Hospital, No. 173 Taishan North Road, Section 1, Jingyang District, Deyang, 618000 Sichuan China
2 grid.413405.7 0000 0004 1808 0686 Department of health management, Guangdong Second Provincial General Hospital, Guangzhou, 510317 Guangdong China
19 9 2024
19 9 2024
2024
161 3510 5 2024
11 9 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Background

Acute myeloid leukemia (AML) is a genetically heterogeneous disease in which glutamine (Gln) contributes to AML progression. Therefore, this study aimed to identify potential prognostic biomarkers for AML based on Gln metabolism-related genes.

Methods

Gln-related genes that were differentially expressed between Cancer Genome Atlas-based AML and normal samples were analyzed using the limma package. Univariate, least absolute shrinkage, selection operators, and stepwise Cox regression analyses were used to identify prognostic signatures. Risk score-based prognostic and nomogram models were constructed to predict the prognostic risk of AML. Subsequently, consistent cluster analysis was performed to stratify patients into different subtypes, and subtype-related module genes were screened using weighted gene co-expression network analysis.

Results

Through a series of regression analyses, HGF, ANGPTL3, MB, F2, CALR, EIF4EBP1, EPHX1, and PDHA1 were identified as potential prognostic biomarkers of AML. Prognostic and nomogram models constructed based on these genes could significantly differentiate between high- and low-risk AML with high predictive accuracy. The eight-signature also stratified patients with AML into two subtypes, among which Cluster 2 was prone to a high risk of AML prognosis. These two clusters exhibited different immune profiles. Of the subtype-related module genes, the HOXA and HOXB family genes may be genetic features of AML subtypes.

Conclusion

Eight Gln metabolism-related genes were identified as potential biomarkers of AML to predict prognostic risk. The molecular subtypes clustered by these genes enabled prognostic risk stratification.

Highlights

Eight genes were identified as potential prognostic biomarkers of acute myeloid leukemia (AML).

The prognostic and nomogram models can accurately predict the AML prognostic risks.

Eight prognostic signatures stratified AML patients into two subtypes with different prognostic patterns and immune profiles.

HOXA and HOXB family genes may be genetic features of AML subtypes.

Keywords

Acute myeloid leukemia
Glutamine
Prognostic model
Molecular subtype
Immune infiltration
issue-copyright-statement© Mendelian Society of Lund and BioMed Central Ltd. 2024
==== Body
pmcBackground

Acute myeloid leukemia (AML) is a fatal cancer characterized by increased self-renewal and uncontrolled proliferation of malignant bone marrow stem cells, accompanied by infection, hemorrhage, and organ infiltration [1]. A small percentage of cases have been determined to be affected by causative factors such as chemotherapy or chemical exposure, but the vast majority develop due to chromosomal abnormalities and gene mutations [2]. As a genetically heterogeneous disease, more than 97% of AML cases have recognizable somatic mutations [3]. Therefore, cytogenetic markers are currently the most important indicators for the risk stratification and treatment of patients with AML [4]. However, AML is still associated with relatively poor survival, and recent data have reported a 5-year overall survival rate of 21% for AML, similar to that of solid organ malignancies with a high fatality rate [5]. This difficulty lies in the fact that the prognosis of AML is closely related to the genetic characteristics of the disease, leading to variability in treatment and prognosis. This study summarized a series of gene mutations, including tumor protein p53 (TP53), nucleophosmin, fms-related receptor tyrosine kinase 3, and CCAAT enhancer binding protein alpha, that may serve as potential prognostic markers and targets for AML [6, 7]. However, the complexity and specificity of each patient’s genetic profile have forced researchers to continually identify novel prognostic markers to predict an individual’s response to treatment, thereby enabling effective personalized treatment.

Metabolic reprogramming is a key manifestation of AML and is closely associated with clinical diagnosis, risk stratification, and targeted drug development [8]. Cellular metabolism in AML is genotype-specific and is accompanied by epigenetic changes, somatic mutations, and activation of downstream cancer-promoting pathways [9]. Amino acid metabolism plays a role in regulating redox homeostasis and maintaining cell proliferation [10]. Glutamine (Gln), a non-essential amino acid, is the most abundant amino acid in human blood. However, when the energy requirement for the rapid proliferation of cancer cells is not met, Gln can be converted to be conditionally essential and contribute to AML cell proliferation [11]. Removal of Gln induces apoptosis in AML cells by inhibiting the mechanistic target of rapamycin complex 1 pathway [12]. Therefore, screening for molecular targets of Gln metabolism may help develop novel AML treatment strategies and improve patient prognosis.

However, no systematic study has been conducted to comprehensively screen biomarkers of AML from the perspective of Gln metabolism to predict prognosis. Therefore, we conducted a series of bioinformatics analyses to screen prognostic signatures from Gln metabolism-related genes to predict prognostic risk and stratify patients to further identify their genetic and immune characteristics. The potential prognostic markers identified in this study may help optimize treatment choices for patients, reduce prognostic risk, and deepen the biological understanding of AML.

Methods

Data search and information

Gene expression profiles of AML were obtained from the Cancer Genome Atlas (TCGA) [13]. Based on the available clinical information, AML samples with prognostic information and survival time over 30 days, totaling 149 cases, were enrolled in this study. In addition, 337 whole blood samples from the Genotype-Tissue Expression (GTEx) database were used as normal controls. These 486 samples were used as the training set for subsequent analyses.

The validation set, GSE71014, was obtained from the Gene Expression Omnibus (GEO) database [14]. It was detected on the GPL10558 Illumina HumanHT-12 V4.0 expression bead chip and comprised 104 AML samples. Finally, 96 AML samples with a survival time of more than 30 days were included in this study.

Screening of differentially expressed genes (DEGs) related to gln metabolism

In GeneCards, 704 protein-coding genes with correlation coefficients greater than eight for Gln metabolism were defined as Gln-related genes, of which 639 were matched to the training set. By comparing the gene expression profiles of these 639 genes between AML and normal groups, Gln-related DEGs (Gln-DEGs) were selected using the limma package (Version 3.52.4) [15], at thresholds of adj.p < 0.05 & |log2fold change (FC)| > 2.

Genetic mutation of Gln-DEGs

The somatic mutation maf files of AML, processed using Mutect software, were downloaded from TCGA. The oncoplot function of the R package maftools was used to plot the mutation waterfall of the TOP10 mutated genes in the Gln-DEGs.

Identification of prognostic signatures to construct the risk score-based model

Based on the Gln-DEGs expression matrix, univariate Cox regression analysis in the R survival package was employed to screen for genes significantly correlated with prognosis at the expression level with a cutoff of p < 0.01. The least absolute shrinkage and selection operator (LASSO) Cox regression analysis in the R. glmnet package (version 1.2) [16] was used to further screen key genes by penalization parameter tuning through 10-fold cross-validation. Finally, the prognostic signatures were screened using stepwise Cox regression analysis in the R. survminer package (version 0.4.9) to construct a prognostic model according to the risk score, which was calculated as follows:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text{risk score}} = \exp(\upbeta 1 \times 1 + \upbeta 2 \times + \cdots + \upbeta {\text{n X n}})$$\end{document}

In this formula, exp indicates the expression level of prognostic signatures, while β represents the stepwise regression coefficient of this gene. The high- and low-risk groups were bounded by the median risk score in the training and validation cohorts. The association between risk score and the actual prognosis was assessed using Kaplan-Meier (KM) analysis in the survival package of R3.6.1 (version 2.41-1). In contrast, the predictive efficacy of the risk score for prognosis at 1, 3, and 5 years was estimated using receiver operating characteristic (ROC) curves.

Screening of independent prognostic factors to establish a nomogram model

This study further included clinical information (age, race, sex, and FAB subtype) as well as risk scores in the univariate and multivariate Cox regression analyses to screen for independent prognostic factors with a p-value less than 0.05. These factors were used to construct a nomogram model to evaluate the predictive relationship between these factors and prognosis using the rms package in R (version 5.1-2) [17]. KM and ROC curves were used to validate the predictive efficacy of the nomogram.

Immuno-analysis of prognostic signatures

The CIBERSORT algorithm [18] was used to estimate the infiltration abundance of key immune cells in high- and low-risk groups. The cor function in R was used to evaluate the relationship between the prognostic signatures and immune cells by calculating the Spearman correlation coefficient. Expression data for 36 immune checkpoint genes and 15 human leukocyte antigen (HLA) family genes were also extracted to compare their expression differences between high- and low-risk groups using the Wilcoxon test.

Identification and comparison of molecular subtypes of AML

In this study, we used the R. ConsensusClusterPlus package (Version:1.58.0) [19] to perform a consistent clustering analysis of AML samples using hierarchical clustering based on Spearman correlation coefficients to cluster and identify different AML subtypes. Similarly, the KM curve was used to compare the survival differences between the two risk groups. The abundance of immune cell infiltration in each subtype sample was calculated using CIBERSORT and was compared between the two subtypes.

Screening of key genes related to AML subtypes using weighted gene co-expression network analysis (WGCNA)

This study performed WGCNA on the top 20% of genes with an absolute deviation from the median of expression value screening in AML samples. WGCNA utilizes module eigengenes to differentiate modules. By calculating the correlations between modules and modules as well as between modules and traits, modules that were highly correlated with the traits were screened, and key genes were selected from the modules. This study used subtypes as phenotypic traits, and key genes were selected from modules related to AML subtypes using the WGCNA package in R (version 1.71) [20].

Protein-to-protein interaction (PPI) analysis

The STRING database (version 11.5) [21] was used to predict the interactions between gene-encoded proteins. In this study, PPI analysis was carried out on subtype-related module genes based on the STRING database, with the species as homo sapiens and the parameter set as an interaction score of 0.4.

Results

Mutational analysis on Gln-DEGs

By comparing the expression data of 639 Gln-related genes in AML and normal samples, 387 Gln-DEGs were identified (Fig. 1A). Genetic level-based analyses suggested that most somatic variants in these genes were missense mutations (Fig. 1B). Among these, the TOP10 genes with the highest mutation frequencies in AML included IDH, TP53, WT1, IDH1, KRAS, PTPN11, ACACB, APC, NPC1, and QRICH2 (Fig. 1C).

Fig. 1 Analysis of somatic variants in Gln-DEGs. A. The volcano plot showed 387 Gln-DEGs between AML and normal samples. B. The summary plot showed information on the variant classification, variant types, variant numbers in each sample, and the top10 mutated genes. C. Waterfall plot depicted the top10 mutated Gln-DEGs in terms of tumor mutation burden. IDH, isocitrate dehydrogenase; TP53: tumor protein p53; WT1: transcription factorWT1; IDH1, isocitrate dehydrogenase (NADP(+)) 1; KRAS, KRAS proto-oncogene; PTPN11, protein tyrosine phosphatase non-receptor type 11; ACACB: acetyl-CoA carboxylase beta; APC, APC regulator of WNT signaling pathway; NPC1, NPC intracellular cholesterol transporter 1; QRICH2: glutamine rich 2

Identification of prognostic signatures from Gln-DEGs

Of the 387 Gln-DEGs, 26 genes significantly correlated with prognosis were screened using univariate Cox regression analysis (Fig. 2A). The expression of these genes was significantly different between AML and whole blood samples (p < 0.0001) (Fig. 2B). Then, the optimal gene list comprising 16 genes was selected from the LASSO analysis (Fig. 2C), followed by the further identification of eight prognostic signatures (hepatocyte growth factor (HGF), angiopoietin-like 3 (ANGPTL3), myoglobin (MB), coagulation factor II (F2), calreticulin (CALR), eukaryotic translation initiation factor 4E binding protein 1 (EIF4EBP1), epoxide hydrolase 1 (EPHX1), and pyruvate dehydrogenase E1 subunit alpha 1 (PDHA1)) using the stepwise Cox regression analysis (Fig. 2D).

Fig. 2 Screening of prognostic signatures using the univariate, LASSO, and stepwise Cox regression analyses. A. The univariate Cox analysis of 26 Gln-DEGs. B. Differences in gene Expression between AML and whole blood samples. ****p < 0.0001. C. Left panel: Distribution of the LASSO coefficients. Right panel: Selection of lambda min based on the likelihood deviation of the LASSO coefficient distribution. D. Eight prognostic signatures identified using stepwise Cox regression analysis

Construct a risk score-based model to predict prognostic risks

Based on the expression data and stepwise regression coefficients of the eight prognostic signatures, the risk score for each sample in the training and validation sets was calculated to establish the prognostic models. After defining the high- and low-risk groups in the training set, it was suggested that patients who died were more frequently distributed in the group with a high prognostic risk (Fig. 3A). The KM curves also confirmed a significant difference in survival probability between the two groups (Fig. 3A). ROC curves were then plotted to assess the sensitivity and specificity of the model based on the training set in predicting AML prognostic risks. The areas under the curve (AUCs) of 1-, 3-, and 5-year ROC were 0.85, 0.846, and 0.875, respectively (Fig. 3A), suggesting the predictive potential of this prognostic model. Furthermore, the model was reconstructed using the validation set, and the results were consistent with the above findings. The validation set-based prognostic model also significantly distinguished the survival status of samples under different risk groups and accurately predicted AML prognostic risks (Fig. 3B).

Fig. 3 Construction a risk score-based model in the training set (A) and the verification of the model in the validation cohort (B). Left panel: risk score distribution and survival status of all AML samples; Middle panel: KM curve displaying the difference in survival probability between high- and low-risk groups; Right panel: ROC curves showing the ability of the model to predict 1-, 3-, and 5-year survival prognosis

Screening for independent prognostic factors and constructing a nomogram to predict survival

By integrating the clinical information and risk scores, independent prognostic factors were selected using univariate and multivariate Cox regression analyses. Age (HR = 1.020, 95% Cl = 1.005–1.036, p = 0.010) and risk score (HR = 4.905, 95% Cl = 3.063–7.856, p < 0.001) were determined to have prognostic independence (Fig. 4A, B). These two factors were incorporated into the nomogram model to predict survival (Fig. 4C). The fitting curves suggested that the overall survial predicted by the model converged with the actual survival (Fig. 4D). The KM curve further demonstrated that patients with higher risk scores had significantly worse prognoses (Fig. 4E). The AUCs of the ROC curves were all over 0.84, indicating that the nomogram was highly sensitive and specific in predicting the 1-, 3-, and 5- survival statuses (Fig. 4F).

Fig. 4 Screening of independent prognostic factors using the univariate and multivariate Cox regression analyses and the construction of the nomogram model to predict prognosis. A: Forest maps showed independent prognostic factor screening by univariate Cox regression analysis. B. Forest maps showed independent prognostic factor screening using multivariate Cox regression analyses. C: Age and risk scores were used to create a nomogram model for predicting survival. D. The fitness of the model predicted the overall survival to actual survival. E. The KM curve showed the survival difference between samples with high- and low-risk groupings using the nomogram model. F. ROC curves confirmed the power of the nomogram model in predicting the 1-, 3-, and 5- survival statuses

Evaluation of immune landscape

In this study, the CIBERSORT algorithm was used to quantify immune cells in all samples. By comparing the high- and low-risk groups, we found five types of immune cells (CD8 T cells, γδT cells, resting NK cells, activated NK cells, and resting mast cells) were significantly different infiltrated between the two groups, and the proportions of these five cells were all significantly decreased in the high-risk group (Fig. 5A). The correlation between the expression levels of the eight prognostic signatures and the infiltration levels of these cells was shown in Fig. 5B. Furthermore, the majority of immune checkpoint genes and HLA family genes were found to have significant differences in expression between the high- and low-risk groups (Fig. 5C and D), suggesting different immune statuses between the two groups.

Fig. 5 Comparison of infiltration of immune cells and expression of immune-related genes between the high- and low-risk groups. A Differences in the infiltration of 22 types of immune cells between the high- and low-risk groups. B: Correlation between immune cell infiltration and expression of eight prognostic signatures. C Box plot showed differences in the expression of immune checkpoint genes between the two groups. *p < 0.5; **p < 0.01; ***p < 0.001; ****p < 0.0001. D. The box plot presented the expression differences in HLA family genes between the two groups. *p < 0.5; **p < 0.01; ***p < 0.001; ****p < 0.0001

Identification and comparison of AML molecular subtypes

Unsupervised cluster analysis was performed based on the expression data of the eight prognostic signatures in the AML samples. By setting the range of K values from 2 to 6, the optimal K = 2 was selected (the curve was more stable) (Fig. 6A), and two clusters were obtained (Fig. 6B). By comparing the survival status between clusters 1 and 2, the KM curve revealed that patients in cluster 1 were prone to a favorable prognosis (Fig. 6C). The Sankey diagram also showed that patients who died were more distributed in Cluster 2, which comprised a greater proportion of patients with high prognostic risks (Fig. 6D). Using CIBERSORT, we identified six types of immune cells showing differences in infiltration between the two subtypes (Fig. 6E). Among them, CD8 T cells, γδT cells, and resting mast cells were significantly decreased in infiltration in cluster 2, which was more distributed in the high-risk group.

Fig. 6 Identification of two subtypes for AML and comparison between cluster 1 and 2. A. Left panel: umulative distribution function (CDF) for consensus clustering with K = 2–6; Right panel: relative change in area under the CDF curve at K = 2–6. B. Consensus Clustering Matrix at Optimal K = 2. C. The KM curve showed the survival difference between the two clusters. D. Sankey diagram showed the distribution of samples with different prognostic risks between the two subgroups. E. Box plot showed the infiltration differences in the 22 types of immune cells between clusters 1 and 2. *p < 0.5; **p < 0.01

Screening of hub genes using WGCNA

To screen for key genes further, we used subtype as a trait to construct a WGCNA network. When the power was 4, the network approximated a scale-free network distribution (Fig. 7A). Based on the dynamic tree-cutting algorithm, we set the minimum number of genes in each module to 30 and finally obtained seven modules (Fig. 7B). By calculating the relationships between the modules and subtypes, the pink module was found to have the strongest correlation with both subtypes (Fig. 7C). Next, 98 genes in the pink module were included in the PPI analysis. Based on the STRING database, a PPI network comprising 48 genes and 161 interaction pairs was constructed (Fig. 7D). Among them, the homeobox A cluster (HOXA) and homeobox B cluster (HOXB) family genes were considered to contribute more to AML subtypes because they had more degrees of connection in this PPI network.

Fig. 7 Screening of hub genes related to AML subtypes using WGCNA and PPI analysis. A. The scale-free fit index of β at soft thresholds of 1–20. B. The Genes were categorized into seven modules using hierarchical clustering. C Correlation between modules and subtypes. D. PPI networks comprising 48 genes and 161 interaction pairs

Discussion

Gln inhibition correlates with the level of glutaminase activity, and glutaminase inhibitors can suppress AML cell growth and induce cell apoptosis and differentiation of disease subtypes. Therefore, inhibition of Gln uptake is an attractive new strategy for treating AML [12, 22]. Key targets in the Gln metabolic pathway, such as proto-oncogenes, glutamic-pyruvic transaminase 2, and solute carrier family 1 member 5, are regulated by insulin-like growth factor 2 mRNA-binding protein 2 in an m6A-dependent manner to promote AML development and stem cell self-renewal [23]. Therefore, based on the 704 genes related to Gln metabolism, we screened 387 genes with significant differences in expression between AML and normal controls. Using univariate, LASSO, and stepwise Cox regression analyses, eight genes (HGF, ANGPTL3, MB, F2, CALR, EIF4EBP1, EPHX1, and PDHA1) were identified as a prognostic signature. The prognostic model constructed using the eight genes accurately predicted the prognostic risk in patients with AML, even in the validation cohort. Based on the expression levels of these genes, this study also categorized AML into two molecular subtypes, in which the prognostic risk of patients in cluster 2 was more adverse. Using the molecular subtype as a trait, this study also screened subtype-related module genes through WGCNA, among which the HOXA and HOXB family genes may be the key genetic features of disease subtypes. These results provided a valuable reference for further understanding the molecular mechanisms of these potential markers in AML and their impact on prognosis.

In the present study, we screened 387 DEGs associated with Gln expression in AML samples. Most somatic variations in these genes were missense mutations. Among these, the TOP10 genes with the highest mutation frequencies in AML were IDH, TP53, WT1, IDH1, KRAS, PTPN11, ACACB, APC, NPC1, and QRICH2. IDH, a mutated enzyme in the citric acid cycle, leads to the production of the oncogenic metabolite R-2-hydroxy-glutarate. This arrests the differentiation of hematopoietic stem cells, leading to the promotion of leukemia [24]. TP53-mutated AML is a unique subtype of AML with a poor prognosis [25]. A longitudinal study tracking the evolution of mutations demonstrated that TP53 mutations represent primary mutational events in chemotherapy or radiation therapy-induced AML [26]. A meta-analysis suggested that WT1 and TP53 mutations exhibit a mutually exclusive tendency in AML [27]. WT1 has been reported to function as an oncogene and tumor suppressor in AML [28–30]. IDH1 mutations occur in 6–10% of patients with AML [31], and inhibitors such as Ivosidenib and Azacitidine are currently available for this mutation [32]. A previous study found that KRAS mutations were associated with poor prognosis in AML [33]. Mutations in PTPN11 and KRAS confer resistance to combinations and multiple venetoclax combinations [34]. However, our understanding of the roles of ACACB, APC, NPC1, and QRICH2 in AML is limited. This study identified mutated genes in AML that could contribute to treating AML.

Of the eight prognostic signatures identified in this study, seven genes, excluding MB, were significantly upregulated in AML. HGF, a hepatocyte growth factor, was confirmed to be upregulated in AML samples and cells and to promote the tumor malignancy of AML cells through targeted binding with miR-204 [35]. The mutation load of CALR alleles increases during the transformation from primary thrombocythemia to AML [36]. Activation of EIF4EBP1, eukaryotic translation initiation factor 4E binding protein 1, promotes AML cell proliferation and disease progression [37]. High expression of EPHX1, a microsomal epoxide hydrolase 1, is significantly associated with high recurrence and low overall survival rates in patients with AML [38], consistent with our findings, suggesting that EPHX1 may promote disease progression in AML. ANGPTL3 suppresses the expression of Ikaros, a key regulator of hematopoietic cell differentiation, thereby promoting the expansion and stemness of HSCs. It stimulates cancer growth by promoting angiogenesis, cell proliferation, and migration [39]. Recent studies have proposed that PDHA1, a key cuproptosis gene, is crucial for reprogramming glucose metabolism in tumor cells [40]. One study found that PDHA1 was abnormally overexpressed in AML, consistent with our findings [41]. The core of thrombosis is thrombin, a product of F2, which participates in coagulation. The upregulation of F2 may increase the risk of thrombotic bleeding complications in patients with AML [42]. MB transports and stores oxygen in muscle cells, and a decrease in MB in patients with AML may be related to impaired heart function [43]. The predictive effects of these eight genes on AML prognosis of AML is initially proposed in this study. A prognostic model constructed using eight genes accurately predicted prognostic risk in patients with AML.

Furthermore, this study found that the infiltration levels of CD8 T cells, γδT cells, resting NK cells, activated NK cells, and resting mast cells were significantly reduced in the high-risk prognostic group. Meanwhile, it is also found that the expression level of EPHX1 was negatively correlated with CD8 T cells, γδT cells, resting NK cells, and activated NK cells, and the expression of EIF4EBP1 was negatively correlated with CD8 T cells and γδT cells. Immune deficiency in AML is reflected in T cells and NK cells, where CD8 + T and diseased γδ T cells exhibit an exhausted state in AML at diagnosis [44]. However, the relationship of EPHX1 and EIF4EBP1 expression with these immune cells has not been investigated in AML. Therefore, this study proposed that EIF4EBP1 and EPHX1 contribute to immunodeficiency and further affect the prognostic survival status of AML patients by negatively modulating the infiltration level of CD8 T cells and γδT cells. However, this hypothesis requires further investigation.

Based on consistent cluster analysis, this study identified two molecular subtypes of AML, with cluster 2 tending to have a poorer prognosis. The samples in cluster 2 were also more distributed in the high-risk group. Furthermore, CD8 T cells, γδT cells, and resting mast cells, which were proportionately downregulated in the high-risk group, were similarly infiltrated in reduced abundance in cluster 2. These results suggested an increased prognostic risk for patients in cluster 2, suggesting that stratification based on disease subtypes can identify survival probability in AML. WGCNA further identified the module genes associated with the subtype, and the PPI network suggested that the HOXA and HOXB family genes may be key genetic characteristics of the disease subtype. Gene amplification, deep deletion, and alterations in the mRNA expression of HOXA have been found in approximately 18% of AML samples, and HOXA3-10 serves as a potential AML therapeutic target and prognostic marker [45]. A relevant bioinformatics analysis revealed that six HOXA and three HOXB genes were significantly underexpressed and hypermethylated in AML, accompanied by favorable cytogenetic profiles [46]. Small-molecule inhibitors targeting the menin-lysine methyltransferase 2 A interaction restored normal HOXA expression in mutant AML, with therapeutic implications for AML patients [47]. In addition, HOXA7, HOXA9, and HOXA11 are associated with AML risk status and prognosis [48]. To further support the above findings, this study suggested that HOXA3-7, 9–11, and HOXB2-9 expression were closely associated with the prognostic risk predicted based on eight prognostic signatures. The expression levels of these markers can help stratify patients with AML and predict their prognostic risks.

However, this study had some limitations. First, the predictive performances of these eight prognostic signatures must be independently validated using a cohort of clinical samples. Moreover, the relationship between these prognostic signatures and immune cell infiltration, as well as their potential regulatory mechanisms, needs to be explored through in vivo and in vitro experiments. In the future, we will continue to explore the potential of these eight genes as prognostic targets in AML and investigate the impact of their expression levels on the efficacy of immunotherapy.

Conclusion

Based on 704 genes related to Gln metabolism, this study carried out differential expression analyses, as well as univariate, LASSO, and stepwise Cox regression analyses, and identified eight prognostic signatures. Prognostic and nomogram models constructed based on the expression levels can accurately identify the prognostic risk of AML. These prognostic signatures clustered patients with AML into two molecular subtypes with different prognostic risk patterns and immune profiles. Among the subtype-related module genes, the HOXA and HOXB family genes may be key genetic features of the AML subtypes.

Acknowledgements

None.

Author contributions

Jie Zhou carried out the Conception and design of the research, Na Zhang, Yan Zuo, Feng Xu and Mi Zhou participated in the Acquisition of data. Lihua Cheng, Yuanyuan Fu and Wenting Zou carried out the Analysis and interpretation of data. Fudong Yang and Min Shu participated in the design of the study and performed the statistical analysis. Jie Zhou and Shengming Zhang drafted the manuscript and revision of manuscript for important intellectual content. All authors read and approved the final manuscript.

Funding

This work was supported by Scientific Research Project Program of Sichuan Medical Association (No.S20038), Incubating Project of Deyang People’s Hospital (No.FHT202004).

Data availability

All data generated or analysed during this study are included in this article.

Declarations

Ethics approval and consent to participate

The use of clinical samples was approved by the Ethics Committee of Guangdong Second Provincial General Hospital (No. 2019-BSGZ-008-01). All the participants provided written informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Abbreviations

AML Acute myeloid leukemia

Gln glutamine

TCGA the Cancer Genome Atlas

GTEx Genotype-Tissue Expression

GEO Gene Expression Omnibus

DEGs differentially expressed genes

Gln-DEGs Gln-related DEGs

KM Kaplan-Meier

ROC receiver operating characteristic

PPI Protein-to-protein interaction

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References

1. Shimony S Stahl M Stone RM Acute myeloid leukemia: 2023 update on diagnosis, risk-stratification, and management Am J Hematol 2023 98 3 502 26 10.1002/ajh.26822 36594187
Shimony S, Stahl M, Stone RM. Acute myeloid leukemia: 2023 update on diagnosis, risk-stratification, and management. Am J Hematol. 2023;98(3):502–26.36594187
2. Pelcovits A Niroula R Acute myeloid leukemia: a review Rhode Island Med J (2013) 2020 103 3 38 40
Pelcovits A, Niroula R. Acute myeloid leukemia: a review. Rhode Island Med J (2013). 2020;103(3):38–40.
3. Kayser S Levis MJ The clinical impact of the molecular landscape of acute myeloid leukemia Haematologica 2023 108 2 308 20 10.3324/haematol.2022.280801 36722402
Kayser S, Levis MJ. The clinical impact of the molecular landscape of acute myeloid leukemia. Haematologica. 2023;108(2):308–20.36722402
4. Prada-Arismendy J Arroyave JC Röthlisberger S Molecular biomarkers in acute myeloid leukemia Blood Rev 2017 31 1 63 76 10.1016/j.blre.2016.08.005 27639498
Prada-Arismendy J, Arroyave JC, Röthlisberger S. Molecular biomarkers in acute myeloid leukemia. Blood Rev. 2017;31(1):63–76.27639498
5. Stubbins RJ Francis A Kuchenbauer F Sanford D Management of Acute myeloid leukemia: a review for General practitioners in Oncology Curr Oncol (Toronto Ont) 2022 29 9 6245 59 10.3390/curroncol29090491
Stubbins RJ, Francis A, Kuchenbauer F, Sanford D. Management of Acute myeloid leukemia: a review for General practitioners in Oncology. Curr Oncol (Toronto Ont). 2022;29(9):6245–59.
6. Weinberg OK Porwit A Orazi A Hasserjian RP Foucar K Duncavage EJ Arber DA The International Consensus classification of acute myeloid leukemia Virchows Archiv: Int J Pathol 2023 482 1 27 37 10.1007/s00428-022-03430-4
Weinberg OK, Porwit A, Orazi A, Hasserjian RP, Foucar K, Duncavage EJ, Arber DA. The International Consensus classification of acute myeloid leukemia. Virchows Archiv: Int J Pathol. 2023;482(1):27–37.
7. Padmakumar D Chandraprabha VR Gopinath P Vimala Devi ART Anitha GRJ Sreelatha MM Padmakumar A Sreedharan H A concise review on the molecular genetics of acute myeloid leukemia Leuk Res 2021 111 106727 10.1016/j.leukres.2021.106727 34700049
Padmakumar D, Chandraprabha VR, Gopinath P, Vimala Devi ART, Anitha GRJ, Sreelatha MM, Padmakumar A, Sreedharan H. A concise review on the molecular genetics of acute myeloid leukemia. Leuk Res. 2021;111:106727.34700049
8. Wojcicki AV Kasowski MM Sakamoto KM Lacayo N Metabolomics in acute myeloid leukemia Mol Genet Metab 2020 130 4 230 8 10.1016/j.ymgme.2020.05.005 32457018
Wojcicki AV, Kasowski MM, Sakamoto KM, Lacayo N. Metabolomics in acute myeloid leukemia. Mol Genet Metab. 2020;130(4):230–8.32457018
9. Mishra SK Millman SE Zhang L Metabolism in acute myeloid leukemia: mechanistic insights and therapeutic targets Blood 2023 141 10 1119 35 10.1182/blood.2022018092 36548959
Mishra SK, Millman SE, Zhang L. Metabolism in acute myeloid leukemia: mechanistic insights and therapeutic targets. Blood. 2023;141(10):1119–35.36548959
10. Kreitz J, Schönfeld C, Seibert M, Stolp V, Alshamleh I, Oellerich T, Steffen B, Schwalbe H, Schnütgen F, Kurrle N et al. Metabolic plasticity of Acute myeloid leukemia. Cells 2019, 8(8).
11. Xiao Y Hu B Guo Y Zhang D Zhao Y Chen Y Li N Yu L Targeting glutamine metabolism as an attractive therapeutic strategy for Acute myeloid leukemia Curr Treat Options Oncol 2023 24 8 1021 35 10.1007/s11864-023-01104-0 37249801
Xiao Y, Hu B, Guo Y, Zhang D, Zhao Y, Chen Y, Li N, Yu L. Targeting glutamine metabolism as an attractive therapeutic strategy for Acute myeloid leukemia. Curr Treat Options Oncol. 2023;24(8):1021–35.37249801
12. Willems L Jacque N Jacquel A Neveux N Maciel TT Lambert M Schmitt A Poulain L Green AS Uzunov M Inhibiting glutamine uptake represents an attractive new strategy for treating acute myeloid leukemia Blood 2013 122 20 3521 32 10.1182/blood-2013-03-493163 24014241
Willems L, Jacque N, Jacquel A, Neveux N, Maciel TT, Lambert M, Schmitt A, Poulain L, Green AS, Uzunov M, et al. Inhibiting glutamine uptake represents an attractive new strategy for treating acute myeloid leukemia. Blood. 2013;122(20):3521–32.24014241
13. Tomczak K Czerwińska P Wiznerowicz M The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge Contemp Oncol (Poznan Poland) 2015 19 1a A68 77
Tomczak K, Czerwińska P, Wiznerowicz M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Poznan Poland). 2015;19(1a):A68–77.
14. Barrett T Wilhite SE Ledoux P Evangelista C Kim IF Tomashevsky M Marshall KA Phillippy KH Sherman PM Holko M NCBI GEO: archive for functional genomics data sets–update Nucleic Acids Res 2013 41 Database issue D991 995 23193258
Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, Marshall KA, Phillippy KH, Sherman PM, Holko M, et al. NCBI GEO: archive for functional genomics data sets–update. Nucleic Acids Res. 2013;41(Database issue):D991–995.23193258
15. Ritchie ME Phipson B Wu D Hu Y Law CW Shi W Smyth GK Limma powers differential expression analyses for RNA-sequencing and microarray studies Nucleic Acids Res 2015 43 7 e47 10.1093/nar/gkv007 25605792
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47.25605792
16. Engebretsen S Bohlin J Statistical predictions with glmnet Clin Epigenetics 2019 11 1 123 10.1186/s13148-019-0730-1 31443682
Engebretsen S, Bohlin J. Statistical predictions with glmnet. Clin Epigenetics. 2019;11(1):123.31443682
17. Zhang Z Geskus RB Kattan MW Zhang H Liu T Nomogram for survival analysis in the presence of competing risks Annals Translational Med 2017 5 20 403 10.21037/atm.2017.07.27
Zhang Z, Geskus RB, Kattan MW, Zhang H, Liu T. Nomogram for survival analysis in the presence of competing risks. Annals Translational Med. 2017;5(20):403.
18. Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods in molecular biology (Clifton, NJ) 2018, 1711:243–259.
19. Wilkerson MD Hayes DN ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking Bioinf (Oxford England) 2010 26 12 1572 3
Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinf (Oxford England). 2010;26(12):1572–3.
20. Langfelder P Horvath S WGCNA: an R package for weighted correlation network analysis BMC Bioinformatics 2008 9 559 10.1186/1471-2105-9-559 19114008
Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559.19114008
21. Szklarczyk D Gable AL Lyon D Junge A Wyder S Huerta-Cepas J Simonovic M Doncheva NT Morris JH Bork P STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets Nucleic Acids Res 2019 47 D1 D607 13 10.1093/nar/gky1131 30476243
Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–13.30476243
22. Matre P Velez J Jacamo R Qi Y Su X Cai T Chan SM Lodi A Sweeney SR Ma H Inhibiting glutaminase in acute myeloid leukemia: metabolic dependency of selected AML subtypes Oncotarget 2016 7 48 79722 35 10.18632/oncotarget.12944 27806325
Matre P, Velez J, Jacamo R, Qi Y, Su X, Cai T, Chan SM, Lodi A, Sweeney SR, Ma H, et al. Inhibiting glutaminase in acute myeloid leukemia: metabolic dependency of selected AML subtypes. Oncotarget. 2016;7(48):79722–35.27806325
23. Weng H Huang F Yu Z Chen Z Prince E Kang Y Zhou K Li W Hu J Fu C The m(6)a reader IGF2BP2 regulates glutamine metabolism and represents a therapeutic target in acute myeloid leukemia Cancer Cell 2022 40 12 1566 e15821510 10.1016/j.ccell.2022.10.004 36306790
Weng H, Huang F, Yu Z, Chen Z, Prince E, Kang Y, Zhou K, Li W, Hu J, Fu C, et al. The m(6)a reader IGF2BP2 regulates glutamine metabolism and represents a therapeutic target in acute myeloid leukemia. Cancer Cell. 2022;40(12):1566–e15821510.36306790
24. Reed DR Elsarrag RZ Morris AL Keng MK Enasidenib in acute myeloid leukemia: clinical development and perspectives on treatment Cancer Manage Res 2019 11 8073 80 10.2147/CMAR.S162784
Reed DR, Elsarrag RZ, Morris AL, Keng MK. Enasidenib in acute myeloid leukemia: clinical development and perspectives on treatment. Cancer Manage Res. 2019;11:8073–80.
25. Shin DY. TP53 mutation in Acute myeloid leukemia: An Old Foe Revisited. Cancers 2023, 15(19).
26. Smith SM Le Beau MM Huo D Karrison T Sobecks RM Anastasi J Vardiman JW Rowley JD Larson RA Clinical-cytogenetic associations in 306 patients with therapy-related myelodysplasia and myeloid leukemia: the University of Chicago series Blood 2003 102 1 43 52 10.1182/blood-2002-11-3343 12623843
Smith SM, Le Beau MM, Huo D, Karrison T, Sobecks RM, Anastasi J, Vardiman JW, Rowley JD, Larson RA. Clinical-cytogenetic associations in 306 patients with therapy-related myelodysplasia and myeloid leukemia: the University of Chicago series. Blood. 2003;102(1):43–52.12623843
27. Yao Y Chai X Gong C Zou L WT1 inhibits AML cell proliferation in a p53-dependent manner Cell Cycle (Georgetown Tex) 2021 20 16 1552 60 10.1080/15384101.2021.1951938 34288813
Yao Y, Chai X, Gong C, Zou L. WT1 inhibits AML cell proliferation in a p53-dependent manner. Cell Cycle (Georgetown Tex). 2021;20(16):1552–60.34288813
28. Welch JS Ley TJ Link DC Miller CA Larson DE Koboldt DC Wartman LD Lamprecht TL Liu F Xia J The origin and evolution of mutations in acute myeloid leukemia Cell 2012 150 2 264 78 10.1016/j.cell.2012.06.023 22817890
Welch JS, Ley TJ, Link DC, Miller CA, Larson DE, Koboldt DC, Wartman LD, Lamprecht TL, Liu F, Xia J, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264–78.22817890
29. Ley TJ Miller C Ding L Raphael BJ Mungall AJ Robertson A Hoadley K Triche TJ Jr Laird PW Baty JD Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia N Engl J Med 2013 368 22 2059 74 10.1056/NEJMoa1301689 23634996
Ley TJ, Miller C, Ding L, Raphael BJ, Mungall AJ, Robertson A, Hoadley K, Triche TJ Jr., Laird PW, Baty JD, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059–74.23634996
30. Hosen N Shirakata T Nishida S Yanagihara M Tsuboi A Kawakami M Oji Y Oka Y Okabe M Tan B The Wilms’ tumor gene WT1-GFP knock-in mouse reveals the dynamic regulation of WT1 expression in normal and leukemic hematopoiesis Leukemia 2007 21 8 1783 91 10.1038/sj.leu.2404752 17525726
Hosen N, Shirakata T, Nishida S, Yanagihara M, Tsuboi A, Kawakami M, Oji Y, Oka Y, Okabe M, Tan B, et al. The Wilms’ tumor gene WT1-GFP knock-in mouse reveals the dynamic regulation of WT1 expression in normal and leukemic hematopoiesis. Leukemia. 2007;21(8):1783–91.17525726
31. DiNardo CD Stein EM de Botton S Roboz GJ Altman JK Mims AS Swords R Collins RH Mannis GN Pollyea DA Durable remissions with Ivosidenib in IDH1-Mutated relapsed or refractory AML N Engl J Med 2018 378 25 2386 98 10.1056/NEJMoa1716984 29860938
DiNardo CD, Stein EM, de Botton S, Roboz GJ, Altman JK, Mims AS, Swords R, Collins RH, Mannis GN, Pollyea DA, et al. Durable remissions with Ivosidenib in IDH1-Mutated relapsed or refractory AML. N Engl J Med. 2018;378(25):2386–98.29860938
32. Gil-Sierra MD Briceño-Casado MP Sierra-Sanchez JF Ivosidenib and Azacitidine in IDH1-Mutated AML N Engl J Med 2022 386 26 2535 6 10.1056/NEJMc2206489 35767448
Gil-Sierra MD, Briceño-Casado MP, Sierra-Sanchez JF. Ivosidenib and Azacitidine in IDH1-Mutated AML. N Engl J Med. 2022;386(26):2535–6.35767448
33. Mustafa Ali MK Williams MT Corley EM AlKaabba F Niyongere S Impact of KRAS and NRAS mutations on outcomes in acute myeloid leukemia Leuk Lymphoma 2023 64 5 962 71 10.1080/10428194.2023.2190432 37042657
Mustafa Ali MK, Williams MT, Corley EM, AlKaabba F, Niyongere S. Impact of KRAS and NRAS mutations on outcomes in acute myeloid leukemia. Leuk Lymphoma. 2023;64(5):962–71.37042657
34. Zhang H Nakauchi Y Köhnke T Stafford M Bottomly D Thomas R Wilmot B McWeeney SK Majeti R Tyner JW Integrated analysis of patient samples identifies biomarkers for venetoclax efficacy and combination strategies in acute myeloid leukemia Nat cancer 2020 1 8 826 39 10.1038/s43018-020-0103-x 33123685
Zhang H, Nakauchi Y, Köhnke T, Stafford M, Bottomly D, Thomas R, Wilmot B, McWeeney SK, Majeti R, Tyner JW. Integrated analysis of patient samples identifies biomarkers for venetoclax efficacy and combination strategies in acute myeloid leukemia. Nat cancer. 2020;1(8):826–39.33123685
35. Nie D Ma P Chen Y Zhao H Liu L Xin D Cao W Wang F Meng X Liu L MiR-204 suppresses the progression of acute myeloid leukemia through HGF/c-Met pathway Hematol (Amsterdam Netherlands) 2021 26 1 931 9
Nie D, Ma P, Chen Y, Zhao H, Liu L, Xin D, Cao W, Wang F, Meng X, Liu L, et al. MiR-204 suppresses the progression of acute myeloid leukemia through HGF/c-Met pathway. Hematol (Amsterdam Netherlands). 2021;26(1):931–9.
36. Langabeer SE Haslam K Elhassadi E The mutant CALR allele burden in essential thrombocythemia at transformation to acute myeloid leukemia Blood Cells Mol Dis 2017 65 66 7 10.1016/j.bcmd.2017.05.004 28552475
Langabeer SE, Haslam K, Elhassadi E. The mutant CALR allele burden in essential thrombocythemia at transformation to acute myeloid leukemia. Blood Cells Mol Dis. 2017;65:66–7.28552475
37. Jiang Y Wu SY Chen YL Zhang ZM Tao YF Xie Y Liao XM Li XL Li G Wu D CEBPG promotes acute myeloid leukemia progression by enhancing EIF4EBP1 Cancer Cell Int 2021 21 1 598 10.1186/s12935-021-02305-z 34743716
Jiang Y, Wu SY, Chen YL, Zhang ZM, Tao YF, Xie Y, Liao XM, Li XL, Li G, Wu D, et al. CEBPG promotes acute myeloid leukemia progression by enhancing EIF4EBP1. Cancer Cell Int. 2021;21(1):598.34743716
38. Cheng H Huang C Tang G Qiu H Gao L Zhang W Wang J Yang J Chen L Emerging role of EPHX1 in chemoresistance of acute myeloid leukemia by regurlating drug-metabolizing enzymes and apoptotic signaling Mol Carcinog 2019 58 5 808 19 10.1002/mc.22973 30644597
Cheng H, Huang C, Tang G, Qiu H, Gao L, Zhang W, Wang J, Yang J, Chen L. Emerging role of EPHX1 in chemoresistance of acute myeloid leukemia by regurlating drug-metabolizing enzymes and apoptotic signaling. Mol Carcinog. 2019;58(5):808–19.30644597
39. Jiang S Qiu GH Zhu N Hu ZY Liao DF Qin L ANGPTL3: a novel biomarker and promising therapeutic target J Drug Target 2019 27 8 876 84 10.1080/1061186X.2019.1566342 30615486
Jiang S, Qiu GH, Zhu N, Hu ZY, Liao DF, Qin L. ANGPTL3: a novel biomarker and promising therapeutic target. J Drug Target. 2019;27(8):876–84.30615486
40. Deng L Jiang A Zeng H Peng X Song L Comprehensive analyses of PDHA1 that serves as a predictive biomarker for immunotherapy response in cancer Front Pharmacol 2022 13 947372 10.3389/fphar.2022.947372 36003495
Deng L, Jiang A, Zeng H, Peng X, Song L. Comprehensive analyses of PDHA1 that serves as a predictive biomarker for immunotherapy response in cancer. Front Pharmacol. 2022;13:947372.36003495
41. Abulimiti M Jia ZY Wu Y Yu J Gong YH Guan N Xiong DQ Ding N Uddin N Wang J Exploring and clinical validation of prognostic significance and therapeutic implications of copper homeostasis-related gene dysregulation in acute myeloid leukemia Ann Hematol 2024 103 8 2797 826 10.1007/s00277-024-05841-6 38879648
Abulimiti M, Jia ZY, Wu Y, Yu J, Gong YH, Guan N, Xiong DQ, Ding N, Uddin N, Wang J. Exploring and clinical validation of prognostic significance and therapeutic implications of copper homeostasis-related gene dysregulation in acute myeloid leukemia. Ann Hematol. 2024;103(8):2797–826.38879648
42. Langer F Quick H Beitzen-Heineke A Janjetovic S Mäder J Lehr C Bokemeyer C Kuta P Renné T Fiedler W Regulation of coagulation activation in newly diagnosed AML by the heme enzyme myeloperoxidase Thromb Res 2023 229 155 63 10.1016/j.thromres.2023.07.006 37473552
Langer F, Quick H, Beitzen-Heineke A, Janjetovic S, Mäder J, Lehr C, Bokemeyer C, Kuta P, Renné T, Fiedler W, et al. Regulation of coagulation activation in newly diagnosed AML by the heme enzyme myeloperoxidase. Thromb Res. 2023;229:155–63.37473552
43. Specchia G Buquicchio C Pansini N Di Serio F Liso V Pastore D Greco G Ciuffreda L Mestice A Liso A Monitoring of cardiac function on the basis of serum troponin I levels in patients with acute leukemia treated with anthracyclines J Lab Clin Med 2005 145 4 212 20 10.1016/j.lab.2005.02.003 15962840
Specchia G, Buquicchio C, Pansini N, Di Serio F, Liso V, Pastore D, Greco G, Ciuffreda L, Mestice A, Liso A. Monitoring of cardiac function on the basis of serum troponin I levels in patients with acute leukemia treated with anthracyclines. J Lab Clin Med. 2005;145(4):212–20.15962840
44. Tang L Wu J Li CG Jiang HW Xu M Du M Yin Z Mei H Hu Y Characterization of Immune Dysfunction and Identification of Prognostic Immune-related risk factors in Acute myeloid leukemia Clin cancer Research: Official J Am Association Cancer Res 2020 26 7 1763 72 10.1158/1078-0432.CCR-19-3003
Tang L, Wu J, Li CG, Jiang HW, Xu M, Du M, Yin Z, Mei H, Hu Y. Characterization of Immune Dysfunction and Identification of Prognostic Immune-related risk factors in Acute myeloid leukemia. Clin cancer Research: Official J Am Association Cancer Res. 2020;26(7):1763–72.
45. Reddel CJ, Tan CW, Chen VM. Thrombin Generation and Cancer: contributors and consequences. Cancers 2019, 11(1).
46. Wang H Lin SY Hu FF Guo AY Hu H The expression and regulation of HOX genes and membrane proteins among different cytogenetic groups of acute myeloid leukemia Mol Genet Genom Med 2020 8 9 e1365 10.1002/mgg3.1365
Wang H, Lin SY, Hu FF, Guo AY, Hu H. The expression and regulation of HOX genes and membrane proteins among different cytogenetic groups of acute myeloid leukemia. Mol Genet Genom Med. 2020;8(9):e1365.
47. Juul-Dam KL Shukla NN Cooper TM Cuglievan B Heidenreich O Kolb EA Rasouli M Hasle H Zwaan CM Therapeutic targeting in pediatric acute myeloid leukemia with aberrant HOX/MEIS1 expression Eur J Med Genet 2023 66 12 104869 10.1016/j.ejmg.2023.104869 38174649
Juul-Dam KL, Shukla NN, Cooper TM, Cuglievan B, Heidenreich O, Kolb EA, Rasouli M, Hasle H, Zwaan CM. Therapeutic targeting in pediatric acute myeloid leukemia with aberrant HOX/MEIS1 expression. Eur J Med Genet. 2023;66(12):104869.38174649
48. Xu P Zhou D Yan G Ouyang J Chen B Correlation of miR-181a and three HOXA genes as useful biomarkers in acute myeloid leukemia Int J Lab Hematol 2020 42 1 16 22 10.1111/ijlh.13116 31670914
Xu P, Zhou D, Yan G, Ouyang J, Chen B. Correlation of miR-181a and three HOXA genes as useful biomarkers in acute myeloid leukemia. Int J Lab Hematol. 2020;42(1):16–22.31670914
