CD172a⁺ tumor-associated macrophages drive immunosuppression and predict poor prognosis in esophageal carcinoma

Abstract

Background

CD172a (SIRPα) is an inhibitory receptor on myeloid cells, but its role in esophageal carcinoma (EC) remains poorly characterized. This study aimed to comprehensively investigate CD172a expression patterns, functional implications, and clinical relevance in EC.

Methods

We integrated bulk RNA-seq data from TCGA (n = 181) and single-cell RNA-seq data from GEO (GSE160269, 60 tumors) to analyze CD172a expression across cellular compartments. Functional enrichment and module scoring were performed to characterize CD172a-associated pathways. A validation cohort of 33 EC patients was analyzed using multiplex immunofluorescence to assess CD172a protein expression in tumor-associated macrophages (TAMs) and its association with clinicopathological features.

Results

CD172a was predominantly expressed in macrophages within the EC tumor microenvironment, with minimal expression in tumor epithelial cells. CD172a-high macrophages exhibited an M2-like immunosuppressive phenotype, characterized by upregulation of CD276, TREM2, and MMP12, and enhanced scores for immunosuppression, phagocytosis, and M2 polarization. High CD172a expression was associated with significantly poorer overall survival in EC patients (p < 0.05), particularly in the ESCC subtype. In our validation cohort, CD172a expression in TAMs was significantly elevated in tumor tissues compared to adjacent normal tissues (p < 0.01), and all patients with distant metastases (9/9) showed CD172a-positive macrophage infiltration.

Conclusion

CD172a marks an immunosuppressive TAM subset in EC and serves as a prognostic biomarker. These findings highlight the CD172a-CD47 axis as a potential therapeutic target for EC immunotherapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04859-6.

Introduction

Esophageal cancer (EC) remains a formidable global health challenge, ranking as the eleventh most common malignancy worldwide and the seventh leading cause of cancer-related mortality [1, 2]. The disease exhibits pronounced geographical disparities, with the highest incidence rates concentrated in Eastern Asia and Eastern Africa [3, 4]. Despite advancements in diagnostic modalities and therapeutic interventions, the prognosis for esophageal cancer patients remains dismal, with 5-year survival rates stagnating at 15–25% [5–6]. This sobering statistic primarily reflects the dual challenges of late-stage diagnosis and the paucity of effective treatment options for advanced disease.

The current therapeutic paradigm for localized esophageal cancer continues to center on surgical resection, which offers the optimal prospect for long-term survival. However, a substantial proportion of patients present with locally advanced or metastatic disease at initial diagnosis, precluding curative surgical intervention. For this patient cohort, multimodal treatment strategies incorporating chemotherapy, radiotherapy, and increasingly, immunotherapy, have become standard of care [5]. The emergence of immune checkpoint inhibitors (ICIs) has particularly transformed the therapeutic landscape for advanced esophageal cancer [7]. The landmark use of pembrolizumab and nivolumab for both esophageal squamous cell carcinoma and adenocarcinoma have firmly established immunotherapy as a cornerstone of contemporary treatment algorithms [8, 9]. Nevertheless, the clinical benefits of ICIs remain heterogeneous, with complete response rates typically ranging from 20 to 25%, underscoring the imperative for more precise patient stratification and novel therapeutic targets [10].

The therapeutic efficacy of immunotherapy is fundamentally governed by the intricate dynamics of the tumor immune microenvironment (TIME), a complex ecosystem comprising diverse immune cell populations, stromal components, and soluble mediators that collectively orchestrate anti-tumor immunity [11]. While conventional immune checkpoints such as PD-1/PD-L1 and CTLA-4 have been extensively characterized in esophageal cancer, emerging evidence suggests that additional co-regulatory molecules may play pivotal roles in immune evasion. Among these, CD172a (also known as signal regulatory protein alpha, SIRPα) represents a promising but underexplored immunomodulatory target. CD172a is a transmembrane glycoprotein predominantly expressed on myeloid cells, including macrophages and dendritic cells, where it functions as a critical inhibitory receptor [12]. Through its interaction with the ubiquitously expressed ligand CD47, CD172a transmits a “don’t eat me” signal that potently suppresses phagocytic activity, thereby facilitating immune escape of malignant cells [13]. Beyond its role in phagocytosis regulation, CD172a has been implicated in modulating cytokine production, antigen presentation, and T-cell activation, suggesting multifunctional contributions to immune homeostasis and tumor immunology [14].

Despite its established role in other malignancies, the expression patterns, functional significance, and clinical relevance of CD172a in esophageal cancer remain largely enigmatic. Comprehensive investigations integrating high-resolution molecular profiling with clinical outcomes are conspicuously absent from the current literature.

The advent of single-cell RNA sequencing (scRNA-seq) technology has revolutionized our capacity to deconvolute the cellular architecture and transcriptional programs of the TIME with unprecedented resolution. When combined with bulk transcriptomic data from large-scale cohorts such as The Cancer Genome Atlas (TCGA), this approach enables multi-dimensional characterization of immune checkpoints across different cellular compartments and disease states. Such integrative analyses are particularly valuable for identifying context-dependent expression patterns and elucidating the cellular networks through which immune regulators exert their functions.

In this study, we employ an integrative multi-omics approach combining scRNA-seq analysis of esophageal cancer specimens with bulk RNA-seq data from TCGA to comprehensively characterize the immunobiological role of CD172a in esophageal cancer. Our specific aims are threefold: (1) to delineate the cellular expression patterns of CD172a within the esophageal cancer TIME at single-cell resolution; (2) to investigate the association between CD172a expression and clinical-pathological features, including survival outcomes and response to immunotherapy; and (3) to elucidate the molecular networks and functional pathways associated with CD172a expression, with particular emphasis on its relationship with myeloid cell polarization. Through these investigations, we aim to investigate whether CD172a can be used as a novel biomarker and potential therapeutic target in esophageal cancer, thereby contributing to the development of more effective immunotherapeutic strategies for this devastating disease.

Materials and methods

Patient enrollment and follow-up

A total of 33 esophageal cancer patients who underwent surgical resection at Peking University International Hospital from April 2025 to June 2025 were enrolled in this study. Paired tumor and adjacent normal tissue samples were collected from surgical specimens or biopsy samples obtained during the diagnostic workup. All patients had not received any prior treatment. All patients were diagnosed with esophageal carcinoma according to the Guidelines for the Diagnosis and Treatment of Esophageal Cancer (2022 Edition). The observation period commenced from the date of surgery until patient death or the last follow-up. Postoperative imaging data, including computed tomography (CT) and/or magnetic resonance imaging (MRI), were collected for all patients at 4–6 weeks after surgery and subsequently every 3 months. The study protocol was approved by the Institutional Review Board of Peking University International Hospital, and written informed consent was obtained from each participant prior to inclusion.

Data acquisition

Transcriptomic data for this study were obtained from two complementary sources. Bulk RNA-sequencing data and corresponding clinical annotations for 181 esophageal cancer patients were retrieved from TCGA database (https://portal.gdc.cancer.gov/), comprising Illumina HiSeq-derived RNA-seq profiles processed through the standard TCGA pipeline, with both raw count data and normalized FPKM values downloaded for subsequent analyses. Single-cell RNA-sequencing data were acquired from the Gene Expression Omnibus (GEO) repository under accession number GSE160269, which includes scRNA-seq profiles from 60 esophageal squamous cell carcinoma tumor samples and 4 matched adjacent normal tissues obtained from 60 individuals. All data processing and analytical workflows were implemented in R (version 4.2.0) with appropriate computational biology packages.

Single-cell RNA sequencing analysis

Single-cell RNA sequencing data analysis was conducted using Seurat (version 4.3.0) and complementary R packages including dplyr (1.1.0), patchwork (1.1.2), and tidyverse (2.0.0). The analytical pipeline incorporated quality control based on mitochondrial gene content (<20%) and unique molecular identifier counts (500-6000 per cell), followed by data normalization using the "LogNormalize" method. Highly variable genes were identified through the variance-stabilizing transformation approach, and principal component analysis (PCA) was performed for dimensional reduction. Cell clustering was achieved using the Find Neighbors and Find Clusters functions with a resolution parameter of 0.5, and results were visualized through distributed stochastic neighbor embedding (TSNE). Cell type annotation was performed according to established lineage-specific markers: Macrophage (CD168, CD63), Mast cell (CPA3, MS4A2), endothelial cells (CCL19,CLEC9A). Macrophage subtyping into M1 and M2-like subsets was performed based on expression signatures of canonical marker genes. M1 macrophages were defined by elevated expression of pro-inflammatory markers ( NOS2, CD80, IL1B, TNF), while M2 macrophages were identified by high expression of anti-inflammatory markers (CD163, MRC1, ARG1, IL10). Subtype assignment was determined using the AddModuleScore function in Seurat, with cells assigned to the M1 or M2 category if their corresponding signature score exceeded 0.5 and was greater than the score of the opposing subset.

Differential gene expression analysis between cell clusters or experimental conditions was performed using the Wilcoxon rank-sum test with thresholds of minimum percentage (min.pct = 0.1) and log-fold change (logfc.threshold = 0.25).

Functional enrichment analysis

Functional enrichment analysis was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID, version 2023q4). A set of genes or cell clusters most significantly associated with CD172a was submitted to DAVID for analysis. Enrichment results from Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were extracted, and the top eight significantly enriched terms (ranked by ascending P-value, with P < 0.05 as the significance threshold) are reported.

Bioinformatic analysis of CD172a functional associations in macrophages

Bioinformatic analysis of CD172a functional associations in macrophage populations was performed to characterize its role in macrophage biology. Macrophage subsets were identified from single-cell RNA sequencing data using established marker genes (CD68, CD163), and CD172a expression levels were stratified into high and low groups based on median expression value. Functional characterization was conducted through gene set enrichment analysis using predefined signature gene sets representing key macrophage functions, including M2 polarization signature (CD163, MRC1, MS4A4A, TGFB1, IL10, ARG1), immunosuppressive function (PDCD1LG2, CD274, IDO1, IL10, TGFB1, ARG1), phagocytic capacity (FCGR2A, FCGR2B, ITGAM, ITGAX, CD36), and metabolic pathways (LDHA, HK2, SLC2A1, MTOR, PPARG). Module scores for each functional category were calculated using the AddModuleScore function in Seurat with parameters set at min.pct = 0.1, logfc.threshold = 0.25, and 100 control genes matched for expression level. The correlation between CD172a expression and functional signature scores was assessed using Spearman’s rank correlation coefficient, while differential activity of functional pathways between CD172a-high and CD172a-low macrophage populations was evaluated using Wilcoxon rank-sum test with significance set at p < 0.05. All analyses were performed in R (version 4.2.0) using Seurat (4.3.0), ggplot2 (3.4.0), and dplyr (1.1.0) packages, with multiple testing correction applied using Benjamini-Hochberg method where appropriate.

Multiple immunofluorescence analysis​

The multiplex immunofluorescence (mIF) assay was conducted in accordance with the manufacturer’s protocol (Servicebio, Wuhan, China). Tissue sections were scanned and digitized using a 3D HISTECH Panoramic Scanner (Panoramic DESK, P-MIDI, P250; Hungary). Quantitative analysis of positively stained cells was performed independently by two experienced investigators to ensure objective and reproducible assessment.

Macrophage infiltration and CD172a expression were quantified by calculating the percentage of CD172a-positive macrophages relative to total CD163-positive cells. For each tissue section, a minimum of five representative high-power fields were analyzed; when fewer than five qualified fields were available, all macrophages in the entire section were enumerated.

Statistical analysis

Statistical analyses were performed using SPSS version 19.0 (IBM, Armonk, NY, USA) and GraphPad Prism version 9.5.0 (GraphPad Software, San Diego, CA, USA). Intergroup differences were assessed with two-tailed unpaired Student’s t-test, Pearson’s χ² test, Mann-Whitney U test, two-way ANOVA, or log-rank test, as appropriate for the data distribution and experimental design. A two-tailed P value < 0.05 was considered statistically significant.

CD172a exhibits pan-cancer differential expression and subtype-specific prognostic significance in esophageal carcinoma

Comprehensive analysis of CD172a expression patterns across multiple cancer types revealed distinct tumor-versus-normal differential expression profiles. As illustrated in Fig. 1A, CD172a was significantly upregulated in tumor tissues compared to matched adjacent normal tissues in several malignancies, including cholangiocarcinoma (CHOL), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), pancreatic adenocarcinoma (PAAD), and skin cutaneous melanoma (SKCM). In contrast, cervical squamous cell carcinoma (CESC) and kidney chromophobe (KICH) exhibited lower CD172a expression in tumor tissues. Notably, esophageal carcinoma demonstrated markedly elevated CD172a levels in tumor samples, suggesting its potential involvement in esophageal carcinogenesis. Analysis of clinical associations in esophageal cancer (Fig. 1B and C) indicated that CD172a expression was significantly higher in esophageal squamous cell carcinoma (ESCC) than in esophageal adenocarcinoma (EAC), though no strong correlations were observed with other clinicopathological parameters. Critically, survival analysis revealed that high CD172a expression was associated with significantly poorer overall survival in esophageal cancer patients (Fig. 1D). This prognostic significance was particularly pronounced in the ESCC subtype, where CD172a expression effectively stratified patient outcomes (log-rank P = 0.027), while no significant association was observed in EAC (log-rank P = 0.237) (Supplementary Fig. 1), underscoring the differential clinical relevance of CD172a across esophageal carcinoma histological subtypes.

Fig. 1.

CD172a exhibits pan-cancer differential expression and subtype-specific prognostic significance in esophageal carcinoma. A (*p<0.05, p<0.01,p<0.001,***p<0.0001) or "ns" (not significant). ESCA shows significantly elevated CD172a expression in tumors. B Correlation heatmap between CD172a expression and clinicopathological parameters in esophageal carcinoma, including tumor type, age, gender, stage, and metastatic status. C Dot plots comparing CD172a expression between different clinicopathological parameters. D Kaplan-Meier survival curves showing overall survival of esophageal cancer patients stratified by high (red) and low (blue) CD172a expression levels. Log-rank p-value is indicated

Functional enrichment and protein localization analysis of CD172a in esophageal carcinoma

Functional enrichment analysis of genes co-expressed with CD172a in esophageal carcinoma revealed distinct biological associations. GO analysis demonstrated significant enrichment in biological processes related to cell division and DNA repair, molecular functions involving protein binding and protein domain-specific binding, and cellular components associated with cytosol and nucleoplasm. KEGG pathway analysis identified significant enrichment in ubiquitin-mediated proteolysis and choline metabolism in cancer, indicating CD172a’s potential role in cellular proliferation and metabolic reprogramming within the tumor microenvironment (Fig. 2A). Immunohistochemical analysis of CD172a protein expression in four independent esophageal carcinoma specimens obtained from The Human Protein Atlas demonstrated minimal to absent CD172a staining in tumor epithelial cells across all examined cases. At varying magnification levels, the neoplastic esophageal epithelial compartments showed negligible CD172a immunoreactivity, while occasional positive staining was observed in scattered stromal immune cells, confirming the transcriptional data indicating limited CD172a expression in esophageal carcinoma parenchymal cells and suggesting its predominant localization within the tumor stroma (Fig. 2B). Analysis of single-cell RNA sequencing data from esophageal carcinoma specimens available through The Human Protein Atlas revealed cell-type-specific expression patterns of CD172a, with predominant expression in macrophage populations and minimal expression detected in epithelial tumor cells, fibroblasts, endothelial cells, or T lymphocytes. Quantitative assessment showed that over 85% of CD172a-expressing cells belonged to the macrophage lineage, establishing CD172a as a macrophage-enriched marker in the esophageal carcinoma microenvironment. This cellular specificity, combined with the observed enrichment in leukocyte migration pathways, supports CD172a’s involvement in myeloid cell biology within the esophageal carcinoma tumor microenvironment (Fig. 2C).

Fig. 2.

Functional enrichment and cellular localization of CD172a in esophageal carcinoma. A Dot plots of Gene Ontology (GO) and KEGG pathway enrichment analysis for genes co-expressed with CD172a. Blue dots represent BP, purple for MF, green for CC, and red for KEGG pathways. Dot size indicates gene count, and color intensity represents -log10 (p-value). B Representative immunohistochemical staining of CD172a in esophageal carcinoma specimens from The Human Protein Atlas. Low- and high-magnification images show minimal staining in tumor epithelial cells and occasional positivity in stromal immune cells. C Single-cell RNA sequencing data demonstrating cell-type-specific expression of CD172a. Over 85% of CD172a-expressing cells are macrophages, with minimal expression in epithelial cells, fibroblasts, endothelial cells, or T lymphocytes

Single-cell resolution characterization of CD172a expression in esophageal carcinoma myeloid compartment

To further delineate the cellular distribution and functional implications of CD172a in the esophageal carcinoma microenvironment, we performed comprehensive single-cell RNA sequencing analysis of myeloid cells from 60 tumor tissues and 4 adjacent normal tissues obtained from the GEO Series GSE160269 dataset. TSNE visualization of the integrated single-cell data (Fig. 3A) revealed distinct clusters corresponding to major myeloid cell populations, including macrophages, mast cells, dendritic cells (DCs), and a minor subset of T cells. Cell type annotation was validated by canonical marker expression patterns (Fig. 3B), with macrophages exhibiting high expression of CD68 and CD163, mast cells expressing CPA3 and MS4A2, and DCs showing enrichment of CCL19 and CLEC9A. Examination of CD172a expression across these cell types (Fig. 3C) demonstrated predominant localization in macrophages, with minimal expression in DCs and negligible expression in mast cells, consistent with our previous findings in Fig. 3C. Comparative analysis of CD172a expression between tumor and adjacent normal tissues (Fig. 3D) revealed significantly elevated CD172a levels in tumor-infiltrating myeloid cells, with statistically significant differences observed in macrophages, mast cells, and DCs. To investigate the functional consequences of CD172a expression heterogeneity, we stratified macrophages into CD172a-high and CD172a-low groups and performed differential gene expression analysis (Fig. 3E). CD172a-high macrophages exhibited significant upregulation of genes associated with immune regulation and tumor progression, including CD276, TREM2, PLA2G7, LGALS3BP, NR1H3, MMP12, and OLFML3, with strong positive correlations between CD172a expression and these markers. Further characterization of macrophage polarization states (Fig. 3F) revealed that CD172a expression was significantly enriched in M2-like macrophages compared to M1-like macrophages. To quantify the functional states associated with CD172a expression, we computed four gene module scores representing immunosuppression, metabolism, phagocytosis, and M2 polarization (Fig. 3G). CD172a-high macrophages demonstrated significantly higher scores across all four functional modules compared to CD172a-low macrophages. Correlation analysis (Fig. 3H) confirmed strong positive associations between CD172a expression and both M2 score (r = 0.38) and phagocytosis score (r = 0.6), suggesting that CD172a expression in tumor-associated macrophages is closely linked to immunosuppressive and phagocytic functional states in the esophageal carcinoma microenvironment.

Fig. 3.

Single-cell resolution characterization of CD172a expression in esophageal carcinoma myeloid compartment. A t-SNE projection of integrated single-cell data from 60 tumor and 4 adjacent normal tissues, colored by major myeloid cell types. B Heatmap showing expression of canonical lineage markers for each cell cluster. C t-SNE plot overlaid with CD172a expression levels, demonstrating predominant expression in macrophages. D Violin plots comparing CD172a expression between tumor and normal tissues across myeloid cell subsets. E Heatmap of differentially expressed genes between CD172a-high and CD172a-low macrophages. F Violin plots comparing CD172a expression between M1-like and M2-like macrophage subsets. G Module scores for immunosuppression, metabolism, phagocytosis, and M2 polarization in CD172a-high versus CD172a-low macrophages. H Correlation heatmap showing associations between CD172a expression and functional module scores (Spearman's r values indicated)

CD172a expression in tumor-associated macrophages may correlate with metastatic progression in esophageal carcinoma

A total of 33 esophageal cancer patients were enrolled in this study, with detailed clinical and pathological characteristics summarized in Table 1. The cohort consisted of 28 males and 5 females with a median age of 64 years (range: 59–68 years). Tumor distribution across esophageal segments was as follows: cervical esophagus (Ce, n = 4), upper thoracic esophagus (Ut, n = 6), middle thoracic esophagus (Mt, n = 20), and lower thoracic esophagus (Lt, n = 8). According to TNM staging, the majority of patients presented with advanced disease, including 19 cases of stage III and 9 cases of stage IV disease. Lymph node metastasis was observed in 32 patients, while 9 patients had distant metastases.

Table 1.

Clinical and pathological characteristics of esophageal cancer patients enrolled in the research

Factor	Category		CD172a high	CD172a low
Age, years		64 (59, 68)	64 (59, 70)	64 (59,71)
Sex	Male	28	19	9
	Female	5	3	2
Tumor location	Ce	4	3	1
	Ut	6	3	3
	Mt	20	13	7
	Lt	8	6	2
Metastatic status	No	1	1	0
	Lymph node metastasis​	32	21	11
	Gastric metastasis	1	1	0
	lung metastasis	3	3	0
	heart metastasis	1	1	0
Histopathology	well differentiated	3	2	1
	moderately differentiated	25	17	8
	poorly differentiated	5	3	2
T	T0	0	0	0
	T1	3	1	2
	T2	7	5	2
	T3	15	10	5
	T4	6	4	2
	Tx	2	2	0
N	N0	1	1	0
	N1	23	13	10
	N2	5	4	1
	N3	2	2	0
	NX	2	2	0
M	M0	23	16	7
	M1	9	5	4
	MX	1	1	0
Stage	I	1	1	0
	II	4	2	2
	III	19	12	7
	IV	9	7	2

Ce Cervical esophagus, Ut Upper thoracic esophagus, Mt Middle thoracic esophagus, Lt Lower thoracic esophagus

Paired tumor and adjacent normal tissue samples were collected from all patients and subjected to multiplex immunofluorescence staining using antibodies against CD163 (macrophage marker) and CD172a. Patients were stratified into CD172a-high (≥ 50% CD172a-positive macrophages) and CD172a-low (< 50% CD172a-positive macrophages) groups based on quantitative assessment.

Immunofluorescence analysis of CD172a expression revealed distinct localization patterns within the esophageal carcinoma microenvironment. In Fig. 4A, staining for DAPI (nuclei), CD163 (macrophage marker), and CD172a demonstrated that CD172a immunoreactivity in macrophages was significantly stronger in tumor regions compared to matched adjacent normal (paratumoral) tissues. Quantitative analysis confirmed significantly higher CD172a expression in tumor tissues compared to matched adjacent normal tissues (p < 0.01), consistent with our previous transcriptomic findings. However, when patients were stratified according to clinicopathological parameters (Fig. 4B), higher infiltration of CD172a⁺macrophages did not show a clear association with advanced tumor (T) stage (T2-T4 vs. T0-T1), nodal (N) metastasis (N2-N3 vs. N0-N1), or distant metastasis (M) status. Nevertheless, all patients with documented distant metastases (9/9) exhibited CD172a-positive macrophage infiltration, suggesting a potential association between CD172a expression and metastatic progression.

Fig. 4.

Clinical validation of CD172a expression in tumor-associated macrophages correlates with disease progression. A Multiplex immunofluorescence staining of paired tumor and adjacent normal tissues. DAPI (blue) stains nuclei, CD163 (green) marks macrophages, and CD172a (red/orange) shows co-localization with macrophages. Magnified insets highlight increased CD172a expression in tumor-associated macrophages compared to para-tumoral regions. B Stacked bar charts showing the proportion of CD172a-high (red) versus CD172a-low (blue) macrophages across clinicopathological parameters: (i) tumor versus para-tumor tissues, (ii) T classification (T0-T1 vs. T2-T4), (iii) N classification (N0-N1 vs. N2-N3), and (iv) M classification (M0 vs. M1). Percentages indicate the proportion of CD172a-high macrophages within each subgroup. All patients with distant metastases (M1) showed CD172a-positive macrophage infiltration

Although the current sample size remains relatively limited, these clinical observations support the hypothesis that CD172a may serve as a relevant biomarker for tumor aggressiveness and metastatic potential in esophageal carcinoma.

Discussion

This study presents a comprehensive investigation into the role of CD172a (SIRPα) within the tumor immune microenvironment of esophageal carcinoma, integrating multi-omics data from public repositories with clinical validation from our patient cohort. Our principal findings consistently demonstrate that CD172a is predominantly expressed by tumor-associated macrophages (TAMs), specifically an M2-polarized subset, and its elevated expression is robustly associated with immunosuppressive functional states, advanced disease progression, and poorer patient survival. These results position CD172a as a compelling biomarker and a potential immunotherapeutic target in esophageal cancer.

Our pan-cancer analysis confirmed that CD172a expression is significantly upregulated in ESCA tumor tissues compared to adjacent normal tissues, a pattern shared by several other malignancies. More importantly, through single-cell RNA sequencing, we precisely localized CD172a expression predominantly to myeloid cells, with over 85% of CD172a-expressing cells identified as macrophages. This myeloid-specific expression pattern is critical, as it distinguishes CD172a from more ubiquitously expressed checkpoints like PD-1/PD-L1. The minimal expression in tumor epithelial cells, confirmed by both transcriptomic data and IHC from The Human Protein Atlas, suggests that CD172a’s role is primarily mediated through the regulation of innate immune responses rather than direct tumor cell-intrinsic mechanisms.

The functional enrichment analysis provided further mechanistic insights. The significant enrichment of CD172a-co-expressed genes in biological processes related to cell division and DNA repair, as well as the KEGG pathway “ubiquitin-mediated proteolysis”, indicates CD172a’s potential role in cellular proliferation and protein degradation processes within the tumor microenvironment. The enrichment in “choline metabolism in cancer” is particularly intriguing, as altered choline metabolism is increasingly recognized as a hallmark of immunosuppressive TAMs, linking CD172a expression to metabolic reprogramming within the tumor microenvironment (TME). Previous studies have reported that tumor cells compete with M1-like TAMs for choline, thereby inhibiting the cGAS/STING signaling pathway and promoting the repolarization of M1-like macrophages towards an M2-like phenotype, which helps maintain the immunosuppressive state of the tumor microenvironment (TME) [15]. This provides a plausible mechanistic link, indicating that CD172a is not merely a marker but potentially an active participant in the metabolic reprogramming that sustains the immunosuppressive function of TAMs in esophageal carcinoma.

Macrophages polarized toward the M2 phenotype within the tumor microenvironment exhibit immunosuppressive functions. Existing studies have indicated that alterations in multiple genes within macrophages can influence their polarization [16–18]. Notably, a key finding of our study is the strong association between high CD172a expression and an M2-like, immunosuppressive tumor-associated macrophage phenotype. This was evidenced by multiple lines of investigation: firstly, the direct comparison showed significantly higher CD172a expression in M2-like macrophages compared to their M1-like counterparts. Secondly, CD172a-high macrophages exhibited a distinct transcriptional signature characterized by upregulation of well-established immunosuppressive and pro-tumoral genes such as CD276(B7-H3), TREM2, and MMP12. CD276 is an emerging immune checkpoint that inhibits T-cell function [19], while TREM2 is a marker of lipid-associated macrophages with potent immunosuppressive capacity [20]. The co-expression of these genes with CD172a suggests a synergistic role in fostering an immune-evasive niche. Furthermore, the module scoring analysis provided quantitative evidence that CD172a-high TAMs possess enhanced functional capacities in immunosuppression, phagocytosis, metabolism, and M2 polarization. The strong positive correlation (r = 0.6) between CD172a expression and the phagocytosis score is paradoxical yet fascinating. While CD172a’s interaction with CD47 is classically known to deliver an inhibitory “don’t eat me” signal [21], our data suggest that CD172a-high TAMs are not functionally anergic but are rather actively engaged in a phagocytic program that may be skewed towards the clearance of apoptotic cells or other non-activating substrates, thereby contributing to tissue remodeling and resolution of inflammation without eliciting anti-tumor immunity. This aligns with the concept of “non-canonical” phagocytosis in tumor promotion.

The clinical relevance of our findings is substantial. The significant association between high CD172a expression and reduced overall survival, particularly in the ESCC subtype, underscores its prognostic value. The stratification of patients based on CD172a-positive macrophage infiltration in our validation cohort further strengthens its potential as a practical biomarker. Most notably, the observation that all patients with distant metastases (9/9) exhibited CD172a-positive macrophage infiltration, while preliminary due to sample size, suggests that CD172a may be implicated in the metastatic cascade. This could be mediated through the mechanisms: M2-like TAMs are known to promote angiogenesis, degrade extracellular matrix to facilitate invasion, and create pre-metastatic niches [22–25].

From a therapeutic perspective, the CD172a-CD47 axis represents a promising target. Several anti-CD47 antibodies are currently in clinical development. Our data suggest that esophageal cancer patients, especially those with ESCC and high CD172a expression in TAMs, could be ideal candidates for such therapies. Blocking the CD47-“don’t eat me” signal could potentially “re-educate” CD172a-high TAMs, unleashing their phagocytic potential against tumor cells. Moreover, the subtype-specific difference (significant in ESCC but not EAC) highlights the necessity for biomarker-driven patient selection in future clinical trials, reflecting the growing understanding of ESCC and EAC as distinct disease entities with unique TMEs.

Despite the compelling evidence, our study has several limitations that must be acknowledged. First, the sample size of our clinical validation cohort (n = 33) is relatively modest, which limits the statistical power for robust multivariate analysis and subgroup analyses. The intriguing link to metastasis, based on 9 patients, requires validation in a larger, independent cohort. Second, the correlative nature of our findings, while strongly suggestive, does not establish causality. The lack of functional experiments, such as in vitro macrophage cultures or in vivo animal models, means we cannot definitively conclude that CD172a expression drives the immunosuppressive and pro-tumoral functions of TAMs. It is plausible that CD172a is merely a marker of a specific TAM subset whose phenotype is determined by other factors. Future research should prioritize several areas: (1) Functional validation:​ Employ genetic approaches (e.g., CRISPR/Cas9) to knock out or overexpress CD172a in primary human macrophages co-cultured with esophageal cancer cells to directly assess its impact on phagocytosis, cytokine secretion, and T-cell suppression. (2) Mechanistic elucidation: (1) Whether CD172a signaling synergizes with or compensates for other checkpoint axes (e.g., PD‑1/PD‑L1, LAG‑3, TIM‑3) in TAM‑mediated immunosuppression; (2) how CD172a regulates downstream metabolic or epigenetic reprogramming (e.g., via arginase‑1, IDO, or histone modifications) that sustains the M2‑like phenotype; and (3) whether CD172a interacts with stromal or tumor‑derived factors (e.g., IL‑10, TGF‑β, extracellular vesicles) to establish an integrated immunosuppressive network. 3. Spatial context: Utilize technologies like imaging mass cytometry to understand the spatial relationship between CD172a⁺ TAMs, CD8⁺ T cells, and tumor cells within the esophageal TME, which could reveal critical interactions for combination therapy design. 4. Larger clinical studies: Validate the prognostic and predictive value of CD172a in large, multi-center prospective cohorts, and explore its potential role in predicting response to immunotherapy.

Conclusion

In conclusion, our integrated multi-omics and clinical validation study establishes CD172a as a macrophage-enriched immune regulator that is intricately linked to an immunosuppressive TAM phenotype, disease progression, and adverse outcomes in esophageal carcinoma. By pinpointing CD172a’s expression to a specific, functionally relevant myeloid subset, we provide a strong rationale for targeting the CD172a-CD47 axis as a novel therapeutic strategy, particularly for ESCC patients. While further functional studies are warranted to establish causality, our findings significantly advance the understanding of myeloid cell biology in esophageal cancer and highlight CD172a as a promising candidate for future biomarker-driven immunotherapy development.

Abbreviations

Ce

Cervical esophagus

CESC

Cervical squamous cell carcinoma

CHOL

Cholangiocarcinoma

EC

Esophageal cancer

ESCA

Esophageal carcinoma

GO

Gene Ontology

HNSC

Head and neck squamous cell carcinoma

ICIs

Immune checkpoint inhibitors

KEGG

Kyoto Encyclopedia of Genes and Genomes

KICH

Kidney chromophobebu da hui

KIRC

Kidney renal clear cell carcinoma

Lt

Lower thoracic esophagus

Mt

Middle thoracic esophagus

PAAD

Pancreatic adenocarcinoma

scRNA-seq

Single-cell RNA sequencing

SKCM

Skin cutaneous melanoma

TAMs

Tumor-associated macrophages

TCGA

The Cancer Genome Atlas

TIME

Tumor immune microenvironment

TME

Tumor microenvironment

Tsne

T-distributed stochastic neighbor embedding

Ut

Upper thoracic esophagus

Authors’ contributions

XZ, XG and WS are responsible for project design; XZ and WG are responsible for data collection and processing; XZ are responsible for experiment data verification and analysis; XZ and WG are responsible for paper writing; XZ, XG, WS and WG are responsible for the modification and improvement of the paper. All authors read and approved the final manuscript.

Funding

No funds, grants, or other support was received.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The research protocol and execution adhere to the ethical standards outlined in the 1975 Declaration of Helsinki. All participants or their legal guardians were apprised of the study’s goals and potential outcomes, and data confidentiality was assured. The study’s approval came from the Faculty of Medicine’s ethical committee at Peking University International Hospital (Ethical Approval No:202504-09).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.