Skip to main content
NCBI home page
Discover Oncology logoLink to Discover Oncology
. 2025 Sep 2;16:1668. doi: 10.1007/s12672-025-03368-4

Upregulation of EMP3 in acute myeloid leukemia: a study based on data mining, RT-qPCR and immunohistochemistry

Angui Liu 1,#, Cong Yu 2,#, Xianwei Peng 1, Jiaodi Liu 1, Yiting Zhang 1, Yubing Ma 1, Kanglai Wei 2,✉, Yinghui Lai 1,✉
PMCID: PMC12401844  PMID: 40890490

Abstract

Background

Epithelial Membrane Protein 3 (EMP3) has been associated with multiple malignancies, but its expression patterns and clinical significance in acute myeloid leukemia (AML) remain poorly characterized.

Methods

Public datasets were integrated to assess EMP3 mRNA expression levels in AML patients versus healthy donors, with validation performed using reverse transcription quantitative PCR (RT-qPCR). Protein expression was accessed through immunohistochemistry. Prognosis relevance was evaluated via survival analysis. Molecular mechanisms were investigated using weighted gene co-expression network analysis (WGCNA), single-cell RNA sequencing, immune infiltration assessment, and pathway enrichment analysis.

Results

Elevated EMP3 expression was detected in AML samples relative to healthy donors, showing a standardized mean difference (SMD) of 0.84 (95% CI: 0.63, 1.05). This upregulation was validated by RT-qPCR and immunohistochemical analyses, yielding a consistent SMD of 0.94 (95% CI: 0.57, 1.32) when RT-qPCR data were included. Prognostic assessment indicated a significant association between EMP3 levels and AML clinical outcomes. Among 829 genes co-expressed with EMP3, enrichment was observed in acute myeloid leukemia-related pathways, with BCL2A1 and ITGAM identified as hub co-expressed genes.

Conclusion

These findings suggest that EMP3 overexpression occurs in AML and potentially influences disease prognosis.

Keywords: Acute myeloid leukemia, EMP3, Gene expression omnibus, RT-qPCR

Introduction

Acute myeloid leukemia (AML) is a malignant hematologic disorder originating in the bone marrow, characterized by substantial clinical and biological heterogeneity. Notable advancements in treatment modalities, including chemotherapy, targeted therapy, and hematopoietic stem cell transplantation, have been made; however, the long-term survival prognosis for AML patients remains poor. This is especially evident in older individuals and those with relapsed or treatment-resistant disease, with a 5-year overall survival rate below 30% [1, 2]. The inherent diversity of AML is primarily attributable to its intricate molecular genetic landscape, which includes gene mutations, chromosomal rearrangements, and epigenetic modifications. These factors collectively drive disease advancement, therapeutic resistance, and recurrence [3, 4]. Thus, thoroughly investigating the molecular mechanisms behind AML and discovering new prognostic biomarkers and treatment targets are essential for enhancing the clinical management of AML patients.

As a member of the Peripheral Myelin Protein 22 (PMP22) gene family, a transmembrane protein is specified by EMP3. This protein acts as a key regulator in numerous biological mechanisms, including cell growth, programmed cell death, movement, and signaling pathways [5]. Recently, more focus has been put on the dysregulated expression of EMP3 in multiple malignancies and its functional roles in tumorigenesis. In glioblastoma (GBM), tumor cell proliferation and migration have been shown to be enhanced by EMP3, potentially serving as an independent prognostic indicator for this aggressive malignancy [6]. In renal cancer, EMP3 has been identified as a potential biomarker and an independent predictor of prognosis, with its role mechanistically linked to processes such as epithelial-mesenchymal transition (EMT), lipid accumulation, and immune infiltration. Targeted therapeutic strategies against EMP3 may thus represent a promising and effective approach for ccRCC management [7]. However, the expression pattern and clinical significance of EMP3 in AML have not been fully elucidated and its precise molecular mechanisms in AML pathogenesis necessitate further investigation.

The current research aims to systematically explore the expression profile and clinical significance of EMP3 in AML through integrative analysis of multiple AML datasets from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, supplemented with RT-qPCR validation using clinical specimens. The protein expression of EMP3 was examined using immunohistochemical analysis. WGCNA was applied to identify EMP3-associated co-expressed genes in AML, followed by functional enrichment analysis to elucidate their potential role in AML-related signaling pathways. These findings support the validation of EMP3 expression patterns in AML and may uncover novel molecular biomarkers for prognostic stratification and targeted therapy development in AML (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

The research approach of this study

Methods and materials

AML mRNA expression data selection

On the one hand, AML-related gene microarrays were screened from the GEO database [8]. The inclusion criteria were as follows: the experimental group was composed of human AML samples, whereas the control group consisted of healthy samples. Secondly, calculated mRNA expression data must be provided by all included microarrays. Thirdly, the sample sizes should exceed three. On the other hand, mRNA sequencing data of human AML samples and healthy samples was acquired from UCSC xena (https://xena.ucsc.edu/).

Clinical sample collection and RT-qPCR

Peripheral blood samples were obtained from 10 AML patients and 10 healthy donor. All normal peripheral-blood (n = 10) samples were obtained from healthy donors aged 35–74 years (median 49 years), which matches the age distribution of the patient cohort (23–79 years, median 53 years). Each healthy donor underwent complete blood count, peripheral-blood smear, and bone-marrow aspirate/biopsy examination; only individuals with (i) normal counts, (ii) no morphologic dysplasia, and (iii) < 5% blasts were included. Total RNA was isolated using the Blood Total RNA Extraction Kit (HANGZHOU SIMGEN BIOTECHNOLOGY CO., LTD, China) following the protocol of manufacturer. The cDNA synthesis was then performed using the MightyScript First-Strand cDNA Synthesis Master Mix (DNase-Free) in accordance with the reagent specifications. The Applied Biosystems QuantStudio 5 Real-Time PCR System was utilized to perform the RT-qPCR. The primer design for EMP3 gene is based on the sequence information provided in the literature and validated using Primer BLAST software. The primers were designed as below: forward primer: 5’-CCTGAATCTCTGGTACGACTGC-3’, reverse primer:5’-GCCATTCTCGCTGACATTACTG-3’. All samples were analyzed in duplicate to ensure experimental reproducibility. The cycle threshold (Ct) values were exported by the system, and the relative expression level of EMP3 was calculated using the 2(−ΔΔCt) method. GAPDH was selected as the housekeeping gene for normalization, supported by several reports identifying it as an optimal choice for normalization [9–11]. Statistical analysis was performed using GraphPad Prism 9, and the results were expressed as the mean ± standard deviation (SD).

Bone marrow biopsy specimen collection and immunohistochemical analysis

Paraffin-embedded bone marrow biopsy specimens from 10 AML patients and 10 normal subjects were obtained from our unit. All normal bone marrow (n = 10) samples were derived from the same donor cohort described above. and all IHC control specimens were confirmed negative for neoplastic involvement by two independent hematopathologists. The paraffin sections were deparaffinized by baking at 75 ℃, subjected to high-pressure antigen retrieval with citric acid buffer, and then treated with peroxidase reagent and goat serum for blocking. The primary antibody (EMP3 antibody at a dilution of 1:400) was applied and incubated overnight at 37 ℃. After rinsing with PBS, the sections were incubated with the secondary antibody at 37 ℃ for 20 min, followed by DAB visualization, hematoxylin counterstaining, dehydration, clearing and mounting with neutral gum. Two pathologists recorded EMP3 protein expression under high-power microscopy. Immunohistochemical sections were scored by multiplying staining intensity (0: none; 1: weak; 2: moderate; 3: strong) by proportion scores (0%: negative; 1: < 30% positive; 2: ≥ 30% positive). Scores of 0–2 indicated negativity, while scores of 3 or above indicated positivity.

Weighted correlation network analysis and immune infiltration analysis

EMP3-related genes were identified based on sequencing data from 173 AML patients in the TCGA dataset. Samples were divided into two groups: an EMP3 high-expression cohort (n = 87) and an EMP3 low-expression cohort (n = 86). Co-expression modules were constructed using the “WGCNA” R package. First, an optimal soft thresholding power β (0 to 20) and correlation coefficient threshold were selected, followed by generation of a topological overlap matrix (TOM). Hierarchical clustering was then performed based on TOM-based dissimilarity (dissTOM), producing a clustering tree. The dynamic tree cut method was then applied to this tree to identify gene co-expression modules. The relationship between modules and EMP3 was assessed using Pearson’s correlation coefficient analysis. The immune infiltration analysis utilized the R package CIBERSORT.

Enrichment analysis

Gene Ontology (GO) analysis, encompassing molecular function (MF), cellular components (CC), and biological processes (BP), was executed using the ClusterProfiler package. In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was carried out to investigate the role of genes in the onset and progression of AML.

Single cell sequencing analysis

The GSE154109 dataset containing 10X single-cell transcriptome data from 15 AML samples was analyzed using single-cell sequencing. Data processing and analysis were performed with the “Seurat” and “SingleR” R packages. For quality control, genes detected in fewer than three cells were excluded during preprocessing. Cells with less than 200 genes were discarded, as were cells with more than 5% mitochondrial gene content. Cell clustering was performed using the “FindNeighbors” and “FindClusters” functions from the “Seurat” package were employed. Subsequently, t-SNE analysis was carried out with the “RunTSNE” function, and clusters were visualized in two-dimensional t-SNE space (t-SNE-1 and t-SNE-2). Differential gene expression analysis between clusters was conducted through Wilcoxon–Mann–Whitney tests implemented in “FindAllMarkers” function in the “Seurat” package.

Statistical analysis

Independent sample t-tests and paired sample t-tests were performed using SPSS19.0. Stata 12.0 was employed to integrate results from RT-qPCR, GEO microarray data, and sequencing data. ROC curve analysis was conducted with these datasets to evaluate the diagnostic potential of EMP3 for distinguishing AML from normal samples.

Results

Expression of EMP3 in AML based on chips data

The expression profile data was retrieved from 16 GEO datasets. EMP3 expression levels in each chip dataset were analyzed through independent sample t-tests. Among the GEO chips, 8 of them (GSE65409, GSE9476, GSE37307, GSE103223, GSE97485, GSE84881, GSE35010, and GSE14924) revealed a significantly elevated EMP3 expression in AML (Table 1; Fig. 2). To improve accuracy, t-test results from the 16 GEO chips were integrated through a continuous variable meta-analysis. The random effects model revealed significant EMP3 upregulation in AML, showing a standardized mean difference (SMD) of 0.84 (95% CI:0.63, 1.05). Publication bias was absent according to funnel plot analysis (Fig. 3a, b). ROC analysis was subsequently conducted across all 16 chips (Table 2; Fig. 4). The summary ROC curve was constructed based on 16 eligible GEO chips, yielding an AUC of 0.88 (0.84, 0.90), with pooled sensitivity and specificity values of 0.73 (0.63, 0.81) and 0.92 (0.79, 0.97), respectively (Fig. 3c, d).

Table 1.

The results of independent sample t-test based on EMP3 expression data from 16 GEO chip data and RT-qPCR data

ID Healthy donor AML T p value
n Mean SD n Mean SD
GSE245305 3 8.3167 0.43751 19 7.5332 1.0563 -1.247 0.227
GSE138883 6 8.7993 0.26272 6 9.3797 0.69168 1.921 0.084
GSE103223 3 8.5577 0.32568 6 10.1333 0.62673 3.996 0.005
GSE97485 10 11.7066 0.09969 20 11.8268 0.11276 2.856 0.008
GSE92778 6 11.0534 0.13284 6 11.1746 0.05475 2.066 0.08
GSE84881 4 6.7698 0.59617 19 7.5439 0.49437 2.758 0.012
GSE65409 8 11.0913 0.23276 30 12.466 0.78267 4.867 < 0.0001
GSE37307 17 9.7504 0.61582 30 10.4545 0.76327 3.247 0.002
GSE35010 16 7.5383 0.93004 15 8.3005 0.80499 2.433 0.021
GSE26294 8 12.7223 0.38218 8 12.7212 0.57774 -0.004 0.997
GSE24395 5 -1.4404 2.35233 12 -0.1736 1.91421 1.166 0.262
GSE24006 31 9.7629 0.97463 23 10.3859 1.38675 1.94 0.058
GSE17054 4 7.3207 1.3032 9 8.1138 2.54699 0.58 0.574
GSE14924 21 9.8104 0.5743 20 10.1676 0.3668 2.36 0.023
GSE9476 38 10.1257 0.54755 26 10.8449 0.86032 3.772 0.001
GSE983 6 11.6416 1.03308 3 11.0967 1.90979 -0.574 0.584
RT-qPCR 10 0.9669 1.20662 10 8.7235 2.0419 10.342 < 0.0001
Open in a new tab

Fig. 2.

Fig. 2

Open in a new tab

Evaluation of mRNA expression levels of EMP3 in AML based on 16 GEO datasets. a GSE983. b GSE9475. c GSE14924. d GSE17054. e GSE24006. f GSE24395. g GSE26294. h GES35010. i GSE37307. j GSE65409. k GSE84881. l GSE92778. m GSE97485. n GSE103223. o GSE138883. p GSE245305

Fig. 3.

Fig. 3

Open in a new tab

Comprehensive evaluation of EMP3 expression in AML via a continuous variable meta-analysis. a Forest plot. b Funnel plot. c Summary ROC curve. d Sensitivity and specificity. e Forest plot after adding RT-qPCR. f Summary ROC curve after adding RT-qPCR

Table 2.

Diagnostic test based on GEO chip data, RT-qPCR data and IHC data

ID Sensitivity Specificity True positive False positive False negative True negative
GSE245305 53% 100% 10 0 9 3
GSE138883 83% 67% 5 2 1 4
GSE103223 100% 100% 6 0 0 3
GSE97485 90% 70% 18 3 2 7
GSE92778 67% 100% 4 0 2 6
GSE84881 95% 75% 18 1 1 3
GSE65409 90% 100% 27 0 3 8
GSE37307 50% 100% 15 0 15 17
GSE35010 87% 69% 13 5 2 11
GSE26294 50% 88% 4 1 4 7
GSE24395 67% 100% 8 0 4 5
GSE24006 70% 87% 16 4 7 27
GSE17054 67% 100% 6 0 3 4
GSE14924 70% 62% 14 8 6 13
GSE9476 54% 100% 14 0 12 38
GSE983 33% 100% 1 0 2 6
RT-qPCR 100% 100% 10 0 0 10
IHC 90% 70% 9 3 1 7
Open in a new tab

Fig. 4.

Fig. 4

Open in a new tab

Receiver operating characteristic curve based on 16 GEO datasets. a GSE983. b GSE9475. c GSE14924. d GSE17054. e GSE24006. f GSE24395. g GSE26294. h GES35010. i GSE37307. j GSE65409. k GSE84881. l GSE92778. m GSE97485. n GSE103223. o GSE138883. p GSE245305

Expression of EMP3 in AML based on chips and RT-qPCR data

Upregulated EMP3 can be found in 10 AML samples compared with 10 control samples (Fig. 5a, b). Integration of GEO data confirmed significant EMP3 upregulation in AML, with the corresponding SMD being 0.94 (95% CI: 0.57, 1.32) (Fig. 3e). The summary sROC curve was constructed based on RT-qPCR data and 16 eligible GEO chips, yielding an AUC of 0.89 (0.86, 0.92), with pooled sensitivity and specificity values of 0.75 (0.65, 0.83) and 0.93 (0.81, 0.98), respectively (Fig. 3f).

Fig. 5.

Fig. 5

Open in a new tab

Evaluation of expression levels of EMP3 in AML based on RT-qPCR and Immunohistochemical. a Evaluation of EMP3 expression in AML and healthy donor bone marrow based on RT-qPCR. b ROC based on RT-qPCR data. c Evaluation of EMP3 expression in AML and healthy donor bone marrow tissue based on IHC. d ROC based on IHC data. e EMP3 immunohistochemical staining on the cell membrane in AML tissues (20X). f EMP3 immunohistochemical staining on cell membrane in normal bone marrow tissue (20X)

Protein expression of EMP3 in AML based on immunohistochemical

Immunohistochemical (IHC) analysis revealed EMP3 protein localization as brownish-yellow granular deposits on cell membranes in AML tissues. An independent samples t-test revealed that EMP3 protein was significantly overexpressed in AML tissues compared to normal bone marrow tissues (Fig. 5c, d). Figure 5e, f illustrates representative EMP3 protein expression patterns (negative and positive) in both normal bone marrow and AML tissues. Among 10 AML tissues, 9 (90%) showed high EMP3 protein expression, and 1 (10%) showed low expression. Of the 10 normal bone marrow tissues, 3 (30%) showed high expression, and 7 (70%) showed low expression.

Relationship of EMP3 expression and prognosis of AML patients

Survival analysis of TCGA data indicates that AML patients with elevated EMP3 levels exhibit poorer prognosis (Fig. 6a, OS HR = 1.90, P = 0.028). Similar results are validated using data from the Gene Set Cancer Analysis (GSCA) database (https://guolab.wchscu.cn/GSCA/#/) (Fig. 6b). A significant positive correlation was identified between EMP3 expression and both Afatinib and Gefitinib within the GSCA database. Conversely, EMP3 expression is marked negatively correlated with (5Z)-7-Oxozeaenol, Dabrafenib, and Bortezomib (Fig. 6c).

Fig. 6.

Fig. 6

Open in a new tab

Survival analysis and drug sensitivity anlysis based on TCGA and GSCA database. a Survival analysis based on TCGA. b Survival analysis based on GSCA database. c Drug sensitivity anlysis based on SCA database

Relationship of EMP3 expression and immune cell infiltration

Distinct immune cell infiltration patterns between AML samples and healthy donors are shown in Fig. 7a, b. Monocytes infiltration was significantly elevated in AML samples. Conversely, five immune cells types presented a substantially reduction in infiltration levels, including naive B cells, plasma cells, resting CD4 memory T cells, activated CD4 memory T cells, and resting NK cells. Correlation analysis between EMP3 expression and immune cell infiltration was performed across 22 types of immune cells subtypes. As a result, EMP3 was found to positively associate with Monocytes, while exhibiting negative associations with three kinds of immune cells, namely plasma cells, resting CD4 memory T cells, and activated dendritic cells (Fig. 7c–f).

Fig. 7.

Fig. 7

Open in a new tab

Relationship of EMP3 expression and immune cells. a The difference of immune cell infiltration in AML and healthy donors. b The proportions of 22 types of immune cells. c activated dendritic cells. d Monocytes. e plasma cells. f resting CD4 memory T cells

Expression of EMP3 in single cells

After comprehensive quality control and rigorous filtering, ten high-quality 10X single-cell transcriptomic datasets were selected for analysis. This investigation characterizes EMP3 expression patterns across bone marrow immune cell populations. Cross-dataset validation consistently detected EMP3 expression across multiple hematopoietic lineages, including monocytes, B lymphocytes, and natural killer (NK) cells (Fig. 8).

Fig. 8.

Fig. 8

Open in a new tab

single-cell transcriptomic analysis based on GSE154109. a GSM4664013. b GSM4664015. c GSM4664016. d GSM4664017. e GSM4664018. f GSM4664019. g GSM4664020. h GSM4664024. i GSM4664026. j GSM4664027

EMP3 co-expressed genes and enrichment analysis

WGCNA analysis of TCGA data identified the blue module as most strongly correlated with EMP3 expression, with a correlation coefficient of 0.56. This module contained 2662 EMP3-associated genes (Fig. 9a–c). To enhance the disease specificity, differentially expressed genes (DEGs) in acute myeloid leukemia were analyzed using the GEPIA2.0 database(TCGA-based). A total of 6,629 upregulated DEGs were detected. Ultimately, intersection analysis between these DEGs and the WGCNA-derived gene set revealed 829 EMP3-related genes (Fig. 9d), representing potential EMP3 co-expressed genes in AML. These genes are considered to be co-expressed with EMP3 in AML. Notably, significant terms of GO-BP were primarily associated with leukocyte activation involved in immune response-activating signaling pathway. GO-MF pathways were mainly linked to immune receptor activity. GO-CC analysis revealed that the 829 genes were notably enriched in the tertiary granule and the secretory granule membrane (Fig. 9e). In the KEGG pathway enrichment analysis, EMP3 and its associated genes were primarily enriched in the pathways of “Acute myeloid leukemia” (Fig. 9f). The genes enriched in the “Acute Myeloid Leukemia” signaling pathway include BCL2A1, ITGAM, KIT, PIK3CD, ZBTB16, CSF1R, SPI1, NFKB1, CD14, FCGR1A, and DUSP6.

Fig. 9.

Fig. 9

Open in a new tab

WGCNA and pathway analysis. a Scale free fit index under different soft thresholds and mean connectivity under different soft thresholds. b Dendrogram of all genes clustered based on optimal soft thresholdand topological overlap matrix. c The correlation and significance between modules and EMP3. d intersection analysis between AML DEGs from GEPIA2.0 database and the WGCNA. e GO analysis. f KEGG analysis

Validation of the clinical values of several hub co-expressed genes of EMP3 in AML

Additional expression validation was conducted for the hub co-expressed genes. Ten genes (BCL2A1, ITGAM, KIT, PIK3CD, CSF1R, SPI1, NFKB1, CD14, FCGR1A, DUSP6) showed upregulation in AML patients (Fig. 10). EMP3 expression correlated with these genes (Fig. 11). The prognostic values of these co-expressed genes were also performed. As a result, up-regulated BCL2A1 and ITGAM were related to poor prognosis of AML patients (Fig. 12).

Fig. 10.

Fig. 10

Open in a new tab

Evaluation of EMP3 hub co-expressed gene expression in AML based on TCGA and GTEx data. a BCL2A1. b ITGAM. c KIT. d PIK3CD. e CSF1R. f SPI1. g NFKB1. h CD14. i FCGR1A. j DUSP6

Fig. 11.

Fig. 11

Open in a new tab

Correlation analysis between EMP3 and its hub co-expressed genes. a BCL2A1. b ITGAM. c KIT. d PIK3CD. e CSF1R. f SPI1. g NFKB1. h CD14. i FCGR1A. j DUSP6

Fig. 12.

Fig. 12

Open in a new tab

prognosis value of EMP3 co-expressed genes. a BCL2A1. b ITGAM. c KIT. d PIK3CD. e CSF1R. f SPI1. g NFKB1. h CD14. i FCGR1A. j DUSP6

Discussion

In this study, multiple GEO microarrays, RT-qPCR, and immunohistochemistry were integrated to systematically evaluate the expression levels of EMP3 in AML. Our research revealed that EMP3 is markedly upregulated in AML patients, and a higher level of EMP3 expression is closely linked to a reduced overall survival rate. Although the meta-analysis revealed a significant overall up-regulation of EMP3 in AML, scrutiny of individual GEO datasets uncovered pronounced heterogeneity. Such variability likely reflects disparities in sample processing (e.g., mononuclear-cell enrichment versus whole-bone-marrow), microarray platforms, or the relative abundance of high-versus low-blast-count specimens. These findings underscore the necessity of a meta-analytic strategy and indicate that EMP3 dysregulation is a consistent hallmark of AML, albeit with magnitude fluctuations across cohorts. These findings are consistent with recent publications in the field [12]. It was also shown that EMP3 was highly expressed in the CD34 + CD117dim AML population [13]. The findings imply that EMP3 could function as a promising prognostic marker for patients with AML.

Previous studies have shown aberrant EMP3 expression across multiple malignancies. In non-small-cell lung cancer, EMP3 levels are markedly reduced in tumour tissue compared with matched normal lung, and patients with lower EMP3 expression exhibit significantly shorter post-operative recurrence-free survival. Conversely, EMP3 is markedly up-regulated in oral squamous-cell carcinoma, breast, gastric, glioblastoma and hepatocellular cancers, where its over-expression correlates with advanced disease and poorer prognosis [5, 6].

Currently, the biological pathways involved in the activity of EMP3 in AML remain incompletely investigated. In contrast, the molecular mechanisms of EMP3 in various other human cancers have been extensively studied. For instance, Zhou et al. reported that EMP3 prevents the entry of S-phage into breast cancer cells and impairs DNA replication, DNA damage repair, and the stem-like properties of these cells. Additionally, EMP3 suppresses the activation of the Akt-mTOR signaling pathway and induces autophagy. EMP3 exhibits an inverse correlation with several genes linked to the S phase of the cell cycle, such as CCNE2, CDK2, PCNA, RFC4, MCM4, and GINS1. Research has demonstrated a significant association between EMP3 expression and the cell cycle of tumor cells [14]. In glioblastoma, EMP3 sustains the oncogenic EGFR/CDK2 signaling pathway by limiting receptor degradation in glioblastoma [15]. Ma et al. revealed that the suppression of EMP3 triggers the MAPK/ERK signaling pathway, thereby influencing the development of gallbladder cancer. Specifically, EMP3, which is regulated by miR-663a, modulates the MAPK/ERK signaling pathway to impede the progression of gallbladder cancer [16]. EMP3 was identified as a core component of the CD13/EMP3/Akt/NF-kB regulatory pathway and may reverse drug resistance by functioning as a downstream oncogene of TWIST1/2. Elevated EMP3 expression correlated with worse overall survival and shorter progression-free survival in gastric cancer patients [17, 18]. These studies indicate that EMP3 may affect the biological behavior of tumor cells through various mechanisms.

Single-cell sequencing analysis and immune infiltration analysis have suggested that Diverse immune cell subsets exhibit EMP3 expression, including monocytes, B cells, and NK cells, and may affect the functions of these immune cells. Prior research has demonstrated a link between EMP3 and a variety of immune cell types, highlighting its potential role in immune-related mechanisms. For instance, Kusumoto et al. demonstrated that EMP3 enhances TNF-α secretion in macrophages, reducing IL-2R expression on CD8 + T cells and inhibiting allogeneic reactive cytotoxic T lymphocytes induction [19]. TCGA analysis by Chen et al. revealed EMP3 expression association with M2 tumor-associated macrophage markers, connecting it to tumor immune suppression mechanisms involving M2 macrophage recruitment and T cell infiltration inhibition [20]. In AML, EMP3 expression across multiple immune cell types implies these cells may represent potential targets for EMP3 activity, though specific mechanisms require further in vivo and in vitro validation.

In this study, BCL2A1 and ITGAM were identified as hub co-expression genes with EMP3 in AML through co-expression analysis. BCL2A1, a member of the BCL-2 family, has anti-apoptotic effects and is highly expressed in various hematologic malignancies [21, 22]. Recent studies have indicated that elevated BCL2A1 expression in AML is strongly associated with poor patient outcomes and resistance to chemotherapy [23, 24]. Yamatani et al. found that the inactivation of STAT5 inhibits BCL2A1, overcoming the chemoresistance of FLT3-ITD/D835 mutated AML [25]. Zhou et al. found that high expression of ITGAM indicates poor prognosis in AML patients. Silencing ITGAM inhibits AML cell viability and induces apoptosis by blocking cell cycle progression [26]. A retrospective study by Wang et al. on 179 clinical cases showed that the positive expression of ITGAM is closely related to AML classification, higher white blood cell count, and poor chemotherapy outcomes [27].

KEGG pathway enrichment analysis revealed close associations between EMP3 and its co-expressed genes with AML pathway, the NF-κB signaling pathway, and the PI3K-Akt signaling pathway. These results indicate potential involvement of EMP3 in AML pathogenesis through these molecular pathways. Previous studies have established connections between these pathways and AML progression, as demonstrated by PHF6-mediated regulation of NF-κB signaling in AML maintenance. Treating myeloid leukemia cells with overexpressed PHF6 with NF-κB inhibitors significantly increased their apoptosis and reduced their proliferation [28]. Shao et al. found that COMMD7 regulated by ZNF460 promotes the proliferation of AML through the NF-κB signaling pathway [29]. Wang et al. found that activation of the PI3K-AKT pathway may promote both proliferation and apoptosis in AML [30]. In addition, these pathways include a number of genes closely related to the development of AML, such as FLT3, KIT, and RAS [31–33].

Although this study revealed increased EMP3 expression in AML and underscored its prognostic significance, as well as its association with co-expressed genes and implicated signaling pathways, several limitations remain. The conclusions are predominantly based on bioinformatics analysis, and we acknowledge that the data are correlative. Definitive functional evidence showing that EMP3 promotes proliferation, inhibits apoptosis, or facilitates immune evasion in AML cells is currently unavailable. Accordingly, future research should incorporate CRISPR/Cas9-mediated gene knockout or inducible knockdown of EMP3 in well-characterized AML cell lines and patient-derived xenograft models. Follow-up in vitro and in vivo assays are essential to clarify the causal role of EMP3 in AML progression and its therapeutic relevance. Additionally, the sample sizes used for immunohistochemistry and RT-qPCR were extremely limited, and additional validation of EMP3 expression in AML using larger cohorts is necessary. This study also did not quantify EMP3 expression in AML cell lines or assess its relationship with chemotherapy sensitivity or resistance in these models. Furthermore, because CIBERSORT relies on linear deconvolution of bulk RNA-seq data against reference gene-expression signatures, its sensitivity to detect subtle or overlapping immune populations is inherently limited, particularly in leukaemia samples. Consequently, minor shifts in rare immune subsets may be underestimated, and orthogonal single-cell profiling will be essential for definitive validation.

Conclusion

The present study validated EMP3 upregulation in AML and its prognostic relevance while identifying associated co-expressed genes and signaling pathways. These results contribute novel molecular insights into AML pathogenesis and suggest potential therapeutic targets for intervention.

Acknowledgements

The authors express their gratitude to GEO and TCGA database for their invaluable raw data.

Author contributions

AGL, CY, YTZ and XWP were responsible for collecting and analyzing data from public datasets, as well as performing statistical analysis. JDL and YBM ware responsible for collecting clinical samples. YHL and KLW contributed to the conception and design of the study.

Funding

The present study was supported by the Guangxi natural science foundation of China(2023JJA140185). Guangxi Zhuang Autonomous Region Health Commission Key Laboratory of Medical Genetics and Genomics Research Open Project(GXKMGG202304). Guangxi Zhuang Autonomous Region Health Commission self-funded scientific research projects(Z-A20240639).

Data availability

The datasets analyzed in this study are available in the following databases: UCSC xena, https://xena.ucsc.edu. GEO, https://www.ncbi.nlm.nih.gov/geo/.(GEO accession numbers: GSE245305, GSE138883, GSE103223, GSE97485, GSE92778, GSE84881,GSE65409, GSE37307, GSE35010, GSE26294, GSE24395, GSE24006, GSE17054,GSE14924, GSE9476, and GSE983). The RT-qPCR and immunohistochemistry data from the present study can be acquired from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This research program was approved by the Medical Ethics Committee of the Second Affiliated Hospital of Guangxi Medical University(Approval Number:2025-KY0255). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional. All participants signed informed consent forms.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

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

Angui Liu and Cong Yu are contributed equally to this work.

Contributor Information

Kanglai Wei, Email: yxwwkl@163.com.

Yinghui Lai, Email: yinghuilai@sina.com.

References

  • 1.Dohner H, DiNardo CD, Appelbaum FR, Craddock C, Dombret H, Ebert BL, Fenaux P, Godley LA, Hasserjian RP, Larson RA, et al. Genetic risk classification for adults with AML receiving less-intensive therapies: the 2024 ELN recommendations. Blood. 2024;144(21):2169–73. [DOI] [PubMed] [Google Scholar]
  • 2.Dohner H, Wei AH, Appelbaum FR, Craddock C, DiNardo CD, Dombret H, Ebert BL, Fenaux P, Godley LA, Hasserjian RP, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345–77. [DOI] [PubMed] [Google Scholar]
  • 3.Papaemmanuil E, Gerstung M, Bullinger L, Gaidzik VI, Paschka P, Roberts ND, Potter NE, Heuser M, Thol F, Bolli N, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Wachter F, Pikman Y. Pathophysiology of acute myeloid leukemia. Acta Haematol. 2024;147(2):229–46. [DOI] [PubMed] [Google Scholar]
  • 5.Zhu W, Song S, Xu Y, Sheng H, Wang S. EMP3: a promising biomarker for tumor prognosis and targeted cancer therapy. Cancer Biomark. 2024;40(3–4):227–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Li L, Xia S, Zhao Z, Deng L, Wang H, Yang D, Hu Y, Ji J, Huang D, Xin T. EMP3 as a prognostic biomarker correlates with EMT in GBM. BMC Cancer. 2024;24(1):89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Lv Q, Xiao W, Xiong Z, Shi J, Miao D, Meng X, Yuan H, Yang H, Zhang X. Identification of candidate biomarker EMP3 and its prognostic potential in clear cell renal cell carcinoma. Front Biosci (Landmark Ed). 2021;26(11):1176–90. [DOI] [PubMed] [Google Scholar]
  • 8.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Gupta DG, Varma N, Abdulkadir SA, Singh P, Sachdeva MUS, Naseem S, Siddiqui MR, Bose P, Binota J, Malhotra P, Khadwal A, Varma S. Identification and validation of the optimal reference genes for standardizing the gene expression profiling diagnostic panel of Ph-like B-lineage acute lymphoblastic leukemia. Clin Exp Med. 2023;23(8):4539–51. [DOI] [PubMed] [Google Scholar]
  • 10.Gupta DG, Varma N, Kumar A, Naseem S, Sachdeva MUS, Bose P, Binota J, Gupta M, Sonam P, Rana P, Malhotra P, Varma S. Identification and validation of suitable housekeeping genes for gene expression studies in BCR-ABL1 positive B-lineage acute lymphoblastic leukemia. Mol Biol Rep. 2022;49(6):4841–8. [DOI] [PubMed] [Google Scholar]
  • 11.Gupta DG, Varma N, Sreedharanunni S, Abdulkadir SA, Naseem S, Sachdeva MUS, Binota J, Bose P, Malhotra P, Khadwal A, Varma S. Evaluation of adverse prognostic gene alterations & MRD positivity in BCR::ABL1-like B-lineage acute lymphoblastic leukaemia patients, in a resource-constrained setting. Br J Cancer. 2023;129(1):143–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Jiang L, Li XP, Dai YT, Chen B, Weng XQ, Xiong SM, Zhang M, Huang JY, Chen Z, Chen SJ. Multidimensional study of the heterogeneity of leukemia cells in t(8;21) acute myelogenous leukemia identifies the subtype with poor outcome. Proc Natl Acad Sci U S A. 2020;117(33):20117–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Li X, Dai Y, Chen B, Huang J, Chen S, Jiang L. Clinical significance of CD34(+)CD117(dim)/CD34(+)CD117(bri) myeloblast-associated gene expression in t(8;21) acute myeloid leukemia. Front Med. 2021;15(4):608–20. [DOI] [PubMed] [Google Scholar]
  • 14.Zhou K, Sun Y, Dong D, Zhao C, Wang W. EMP3 negatively modulates breast cancer cell DNA replication, DNA damage repair, and stem-like properties. Cell Death Dis. 2021;12(9):844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Martija AA, Krauss A, Bachle N, Doth L, Christians A, Krunic D, Schneider M, Helm D, Will R, Hartmann C, et al. EMP3 sustains oncogenic EGFR/CDK2 signaling by restricting receptor degradation in glioblastoma. Acta Neuropathol Commun. 2023;11(1):177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ma Q, Zhang Y, Liang H, Zhang F, Liu F, Chen S, Hu Y, Jiang L, Hao Y, Li M, et al. EMP3 as a key downstream target of miR-663a regulation interferes with MAPK/ERK signaling pathway to inhibit gallbladder cancer progression. Cancer Lett. 2023;575:216398. [DOI] [PubMed] [Google Scholar]
  • 17.Guo Q, Jing FJ, Xu W, Li X, Li X, Sun JL, Xing XM, Zhou CK, Jing FB. Ubenimex induces autophagy Inhibition and EMT suppression to overcome cisplatin resistance in GC cells by perturbing the CD. Aging. 2019;12(1):80–105. 13/EMP3/PI3K/AKT/NF-kappaB axis. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Han M, Xu W. EMP3 is induced by TWIST1/2 and regulates epithelial-to-mesenchymal transition of gastric cancer cells. Tumour Biol. 2017;39(7):1010428317718404. [DOI] [PubMed] [Google Scholar]
  • 19.Kusumoto Y, Okuyama H, Shibata T, Konno K, Takemoto Y, Maekawa D, Kononaga T, Ishii T, Akashi-Takamura S, Saitoh SI, et al. Epithelial membrane protein 3 (Emp3) downregulates induction and function of cytotoxic T lymphocytes by macrophages via TNF-alpha production. Cell Immunol. 2018;324:33–41. [DOI] [PubMed] [Google Scholar]
  • 20.Chen Q, Jin J, Huang X, Wu F, Huang H, Zhan R. EMP3 mediates glioblastoma-associated macrophage infiltration to drive T cell exclusion. J Exp Clin Cancer Res. 2021;40(1):160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Decombis S, Bellanger C, Le Bris Y, Madiot C, Jardine J, Santos JC, Boulet D, Dousset C, Menard A, Kervoelen C, et al. CARD11 gain of function upregulates BCL2A1 expression and promotes resistance to targeted therapies combination in B-cell lymphoma. Blood. 2023;142(18):1543–55. [DOI] [PubMed] [Google Scholar]
  • 22.Burley TA, Kennedy E, Broad G, Boyd M, Li D, Woo T, West C, Ladikou EE, Ashworth I, Fegan C et al. Targeting the Non-Canonical NF-kappaB pathway in chronic lymphocytic leukemia and multiple myeloma. Cancers (Basel) 2022;14(6). [DOI] [PMC free article] [PubMed]
  • 23.Zhang H, Nakauchi Y, Kohnke 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Alharbi H, Ahmad M, Cui Z, Meng D, Xin Y, Yan X. Identification of potential biomarkers and pathways in acute myeloid leukemia: correlation between the calcineurin signaling pathway and vascular brittleness in acute myeloid leukemia. Int J Lab Hematol. 2025;47(2):288–96. [DOI] [PubMed] [Google Scholar]
  • 25.Yamatani K, Ai T, Saito K, Suzuki K, Hori A, Kinjo S, Ikeo K, Ruvolo V, Zhang W, Mak PY, et al. Inhibition of BCL2A1 by STAT5 inactivation overcomes resistance to targeted therapies of FLT3-ITD/D835 mutant AML. Transl Oncol. 2022;18:101354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zhou C, Yang L, Zhao K, Jiang L. ITGAM sustains MAPK signaling and serves as an adverse prognostic factor and therapeutic target in acute myeloid leukemia. Transl Cancer Res. 2024;13(6):3062–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Wang J, Hao JP, Uddin MN, Wu Y, Chen R, Li DF, Xiong DQ, Ding N, Yang JH, Ding XS. Identification and validation of inferior prognostic genes associated with immune signatures and chemotherapy outcome in acute myeloid leukemia. Aging. 2021;13(12):16445–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Hou S, Wang X, Guo T, Lan Y, Yuan S, Yang S, Zhao F, Fang A, Liu N, Yang W, et al. PHF6 maintains acute myeloid leukemia via regulating NF-kappaB signaling pathway. Leukemia. 2023;37(8):1626–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Shao X, Zhong L, Chu X, Wan P, Chen S, Zhou Z, Zhang H, Wang M, Liu B. ZNF460-regulated COMMD7 promotes acute myeloid leukemia proliferation via the NF-kappaB signaling pathway. Int J Med Sci. 2023;20(4):520–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Wang F, Yang L, Xiao M, Zhang Z, Shen J, Anuchapreeda S, Tima S, Chiampanichayakul S, Xiao Z. PD-L1 regulates cell proliferation and apoptosis in acute myeloid leukemia by activating PI3K-AKT signaling pathway. Sci Rep. 2022;12(1):11444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Bystrom R, Levis MJ. An update on FLT3 in acute myeloid leukemia: pathophysiology and therapeutic landscape. Curr Oncol Rep. 2023;25(4):369–78. [DOI] [PubMed] [Google Scholar]
  • 32.Ishikawa Y, Kawashima N, Atsuta Y, Sugiura I, Sawa M, Dobashi N, Yokoyama H, Doki N, Tomita A, Kiguchi T, et al. Prospective evaluation of prognostic impact of KIT mutations on acute myeloid leukemia with RUNX1-RUNX1T1 and CBFB-MYH11. Blood Adv. 2020;4(1):66–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 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. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The datasets analyzed in this study are available in the following databases: UCSC xena, https://xena.ucsc.edu. GEO, https://www.ncbi.nlm.nih.gov/geo/.(GEO accession numbers: GSE245305, GSE138883, GSE103223, GSE97485, GSE92778, GSE84881,GSE65409, GSE37307, GSE35010, GSE26294, GSE24395, GSE24006, GSE17054,GSE14924, GSE9476, and GSE983). The RT-qPCR and immunohistochemistry data from the present study can be acquired from the corresponding author upon reasonable request.

Articles from Discover Oncology are provided here courtesy of Springer

RESOURCES