Abstract

Background

   Tumor-associated macrophages (TAMs), as key immune components of the
   TME, play a pivotal role in tumor progression by fostering an
   immunosuppressive environment. Programmed death 1 (PD-1), a critical
   immune checkpoint molecule predominantly expressed on T cells, mediates
   immune suppression by binding to programmed death-ligand 1 (PD-L1) on
   tumor cells within the tumor microenvironment (TME). Emerging evidence
   reveals that TAMs also express PD-1, however, the expression and
   functional regulatory mechanisms of PD-1 on TAM remain poorly
   understood.

Methods

   In this study, we combined bulk RNA sequencing and single-cell RNA
   sequencing (scRNA-seq) data to investigate the association between PD-1
   expression on macrophages and patient prognosis, while also uncovering
   the molecular mechanisms by which PD-1 regulates macrophage function.
   To further investigate the role of PD-1 in macrophage activity, we
   established a fluorescence-labeled tumor-bearing mouse model using CT26
   cells, a murine colorectal cancer cell line, to evaluate the
   relationship between PD-1 expression on TAMs and their phagocytic
   activity as well as other functions. Additionally, to mimic the TME in
   vitro, we cultured bone marrow-derived macrophages (BMDMs) with
   CT26-conditioned medium (CT26-CM).

Results

   Our results suggest that PD-1 expression on TAMs drives macrophage
   polarization toward an M2-like phenotype, suppresses their phagocytic
   activity, inhibits the synthesis of interferon-γ (IFN-γ) signaling
   molecules, and ultimately promotes tumor progression. Mechanistically,
   we demonstrated that PD-1 regulates the synthesis of IFN-γ signaling
   molecules and the polarization of M2-type macrophages in BMDMs through
   the JAK2-STAT3 signaling pathway. Overall, our study demonstrates that
   PD-1 expression on TAMs facilitates the formation of an
   immunosuppressive microenvironment, ultimately accelerating tumor
   progression.

Clinical trial number

   Not applicable.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12967-025-06469-4.

   Keywords: Programmed death-1, Tumor-associated macrophages, Anti-tumor,
   JAK2-STAT3 signaling pathway, Bone marrow-derived macrophages

Introduction

   Colorectal cancer (CRC) is the third most common cancer worldwide
   [[42]1]. In recent years, immune checkpoint inhibitors have gained
   increasing attention as a promising therapeutic strategy for CRC
   patients [[43]2, [44]3]. In 2015, Dung T. Le et al. discovered that
   monoclonal antibodies targeting programmed cell death protein 1 (PD-1,
   encoded by PDCD1) exhibited significant efficacy in patients with high
   microsatellite instability/deficient DNA mismatch repair (MSI-H/dMMR)
   CRC, whereas resistance was observed in patients with microsatellite
   stable/proficient DNA mismatch repair (MSS/pMMR) CRC [[45]4]. PD-1, an
   apoptosis-related inhibitory receptor, is primarily expressed on
   activated T cells [[46]5]. By interacting with programmed death-ligand
   1 (PD-L1) on tumor cells, PD-1 suppresses T cell-mediated anti-tumor
   immune responses, thereby promoting immune evasion [[47]6–[48]8]. The
   role of the PD-1/PD-L1 axis in adaptive immunity has been extensively
   studied [[49]9, [50]10]. Increasing evidence suggests that interactions
   between tumor cells and macrophages may contribute to resistance to
   anti-PD-1 therapy [[51]11]. Therefore, to overcome the variability in
   PD-1 immune therapy sensitivity and resistance across different CRC
   subtypes, there is an urgent need to gain a comprehensive understanding
   of the mechanisms of PD-1 action from a multi-cellular perspective.

   Tumor-associated macrophages (TAMs) are key components of the tumor
   microenvironment (TME) [[52]12], classified into pro-inflammatory M1
   and anti-inflammatory M2 types [[53]13–[54]16]. TAMs promote tumor
   progression by establishing an immunosuppressive microenvironment that
   suppresses the host’s anti-tumor immune response [[55]17–[56]24].
   Research by Strauss L et al. has shown that PD-1 expression on the
   surface of macrophages plays a crucial role in anti-tumor immune
   responses. Compared to T cell-specific PD-1 knockout, conditional
   deletion of PD-1 in myeloid cells significantly suppresses tumor growth
   [[57]25]. This finding underscores the need to shift our research focus
   toward macrophages to further investigate the specific mechanisms of
   PD-1. Existing evidence suggests that high macrophage infiltration is
   closely associated with resistance to PD-1/PD-L1 immune checkpoint
   inhibitors. Therefore, TAMs are considered potential targets for
   overcoming PD-1 therapy resistance [[58]26–[59]29]. Further
   investigation into the expression of PD-1 on macrophages and its
   regulatory mechanisms has become a key focus of current research.

   Sydney R. Gordon’s research on PD-1 expression in macrophages has laid
   the foundation for this field, initiating new explorations into the
   role of PD-1 in immune regulation. It has been confirmed that PD-1 is
   predominantly expressed on the surface of M2 macrophages. Moreover,
   Giemsa staining of TAMs and in vitro phagocytosis assays using
   Staphylococcus aureus bio-particles revealed that macrophages
   expressing PD-1 exhibit relatively lower phagocytic capacity [[60]30].
   Recent studies on the mechanisms of intracellular PD-1 expression in
   macrophages have also been explored in chronic inflammation models
   [[61]31, [62]32]. The high expression of PD-1 in macrophages has been
   shown to depend on transforming growth factor-beta1, which sustains
   PD-1 expression through the SMAD3/STAT3 signaling pathways [[63]33].
   However, comprehensive studies on the regulation of macrophage function
   by PD-1 are still lacking, and the specific intracellular regulatory
   mechanisms of PD-1 in macrophages remain underexplored.

   This study investigates the impact of PD-1 expression on the surface of
   macrophages within the CRC tumor microenvironment on their anti-tumor
   immune functions. By combining bulk RNA sequencing and single-cell RNA
   sequencing (scRNA-seq) data, we provide a comprehensive analysis of the
   regulatory effects and underlying mechanisms of PD-1 expression in
   macrophages. To validate our findings, we employed in vivo and in vitro
   models, including Enhanced Green Fluorescent Protein (EGFP)^+CT26
   BALB/c (a murine colorectal cancer cell line) tumor-bearing mice, CT26
   cell models and bone marrow-derived macrophage (BMDM) models. These
   experiments confirmed the broad impact of PD-1 expression in the TME on
   macrophage function and further elucidated the mechanisms by which PD-1
   regulates macrophage activity. Overall, this study aims to advance our
   understanding of how PD-1 influences macrophage-mediated anti-tumor
   immunity in CRC and to uncover its potential regulatory mechanisms.

Methods

Cell culture and construction of EGFP^+ CT26 cell line

   CT26 and Raw264.7 cells were obtained from the National Collection of
   Authenticated Cell Cultures, China, and cultured at the Northern
   Translational Medicine Center, Harbin Medical University. CT26 cells
   were maintained in Roswell Park Memorial Institute 1640 medium (RPMI
   1640, Thermo Fisher Scientific, USA) supplemented with 10% fetal bovine
   serum (FBS, Thermo Fisher Scientific, USA) and 1%
   penicillin-streptomycin. Raw264.7 cells were cultured in Dulbecco’s
   Modified Eagle’s Medium (DMEM, Thermo Fisher Scientific, USA) under the
   same supplementation conditions.

   HEK 293T cells were used as host cells for lentivirus production. The
   lentiviral plasmid PLV (VectorBuilder, China), carrying an EGFP
   fluorescent tag and a neomycin resistance gene, was utilized. Plasmids
   PxpaX2, PMD2g and PLV were mixed at a mass ratio of 3:2:1 and
   transfected into 293T cells using Lipofectamine 3000 (Thermo Fisher
   Scientific, USA) to form plasmid-reagent complexes, which were then
   internalized by the cells. Viral supernatants were collected 48–72 h
   post-transfection and subsequently used to infect CT26 cells in the
   presence of 8 µg/ml Polybrene (Univ, China) to enhance infection
   efficiency. Finally, stable cell lines were established by selecting
   resistant clones with 800 µg/ml G418 (Thermo Fisher Scientific, USA).

Bone marrow-derived macrophages

   Cervical dislocation was performed to euthanize 6- to 8-week-old
   C57BL/6 mice. After femurs and tibias were harvested, bone marrow was
   collected and cultured in DMEM containing 10% FBS and 10 ng/ml
   recombinant mouse macrophage colony-stimulating factor (M-CSF). After 3
   days, half of the medium was replaced and BMDMs were harvested on day
   7. To detach the cells, BMDMs were incubated with Accutase
   (Sigma-Aldrich, USA) in a CO[2] incubator for 25 min, facilitating cell
   detachment from the culture dish.

Mice and tumor establishment

   SPF-grade BALB/c mice (Vitalriver, China) were subcutaneously injected
   with 100 µl of CT26 cell suspension (5 × 10⁵ cells per mouse) into the
   left flank. A nodular elevation appeared immediately after injection,
   with palpable granules detectable on day 5. Tumor size was measured on
   day 7 using the formula: length × width^2/2. If multiple nodules were
   present, their measurements were summed. During the modeling period,
   temperature, humidity, airflow, and lighting were strictly controlled
   to simulate day/night cycles in the animal facilities. Mice were
   provided with adequate feed and sterilized water, with regular
   replacement of high-pressure processed bedding and cages. Continuous
   access to drinking water and food was ensured. All animal procedures
   were approved by the Ethics Committee of Harbin Medical University
   (Harbin, China).

Flow cytometry

   For macrophage detection, the CD16/CD32 (Biolegend, USA) blocking
   antibody was used. Macrophage markers and polarization-related
   antibodies included PerCP-anti-F4/80, PE-anti-CD86, and FITC-anti-CD206
   (BioLegend, USA). Additionally, anti-mouse PD-1 (BIO X Cell, USA),
   anti-mouse PD-L1, PerCP-anti-Interferon-γ (IFN-γ) (BioLegend, USA) and
   PE-anti-Rat IgG (Abcam, UK) were employed. For intracellular staining,
   the Fix/Permeabilization Kit (BD Biosciences, USA) was utilized.

Macrophage phagocytosis

   Bone marrow-derived macrophages (BMDMs) were seeded at 200,000 cells
   per well in a 12-well plate. After the cells had fully adhered and
   spread, 1 μm latex beads (Sigma-Aldrich, USA) were added at a 1:250
   dilution and incubated for 1 h. Non-phagocytosed fluorescent
   microspheres were then removed by washing with PBS. Subsequently, cells
   were stained with DiD membrane dye (Absin, China) at a 1:10 dilution,
   followed by DAPI (Beyotime, China) for nuclear staining.

SiPD-1

   To knock down PD-1 on the surface of BMDMs, siPD-1 (5’-3’:
   CACUUCUAGGGACUUGAGATT; 3’-5’: UCUCAAGUCCCUAGAAGUGTT) and simock (5’-3’:
   TTCTCCGAACGTGTCACGT; 3’-5’: ACGTGACACGTTCGGAGAA) were transfected using
   INTERFERin^® transfection reagent (Polyplus-transfection^®, France).
   For siRNA-mediated knockdown in a 24-well plate, 8.4 ng of siRNA was
   first diluted in 100 µl of serum-free medium, followed by the addition
   of 2 µl INTERFERin^®. The mixture was gently vortexed and incubated at
   room temperature for 10 min to allow the formation of transfection
   complexes. Subsequently, the transfection complex was added to each
   well containing cells at 50% confluency in 500 µl of fresh medium.
   After a 72-hour incubation, protein levels were assessed to confirm
   knockdown efficiency.

Migration and invasion

   For the migration assay, the lower chamber of the 24-well transwell
   plate was filled with 600 µl RPMI 1640 containing 20% FBS and 1%
   penicillin-streptomycin. The upper chamber was seeded with 20,000 cells
   in RPMI 1640. After 24 h of incubation, cells were fixed with 4%
   formaldehyde and stained with 0.2% crystal violet to assess migration.

   For the invasion assay, the transwell chamber was coated with Matrigel
   (Corning, USA) to mimic the extracellular matrix. The lower chamber was
   filled with 600 µl RPMI 1640 containing 20% FBS and 1%
   penicillin-streptomycin, while the upper chamber contained 40,000 cells
   in RPMI 1640. After 24 h of incubation, cells were fixed with 4%
   formaldehyde and stained with 0.2% crystal violet to evaluate invasion.

Western blot

   JAK2 Antibody (Abmart, China), Phospho-JAK2 (Tyr931) Antibody (Abmart,
   China), STAT3 Antibody (Abmart, China), and Phospho-STAT3 (Y705)
   Antibody (Abmart, China) were used for Western blot analysis to detect
   signaling proteins in macrophages. Colivelin (MCE, USA), a
   cell-permeable neuroprotective peptide, was utilized in this study as a
   potent activator of STAT3.

   The Western blotting procedure was performed according to standard
   protocols described in previous literature [[64]34]. Briefly, treated
   cells were collected and thoroughly lysed, followed by centrifugation
   to obtain the supernatant. Protein samples were prepared by boiling the
   collected supernatant. After protein quantification, samples were
   subjected to electrophoresis using 10% and 12.5% SDS-polyacrylamide
   gels, followed by wet transfer (300 mA, 90 min) onto PVDF membranes.
   The membranes were then blocked and sequentially incubated with primary
   and secondary antibodies. For signal detection, ECL substrate solution
   was freshly prepared by mixing solution A and B at recommended ratios.
   The PVDF membranes were evenly coated with the ECL mixture and
   immediately analyzed using a chemiluminescence imaging system.

Transcriptomic data collection

   The bulk RNA-seq data of colorectal cancer (TCGA-COAD) used in this
   study were obtained from the UCSC Xena database
   ([65]http://xenabrowser.net/hub). The scRNA-seq data from CRC murine
   models were sourced from the Gene Expression Omnibus (GEO) database
   ([66]GSE150970; [67]http://www.ncbi.nlm.nih.gov/geo/).

Bulk RNA-Seq data processing and normalization

   By integrating clinical information with gene expression profiles, a
   total of 379 samples were included for subsequent analysis. To
   normalize the gene expression data, we converted the raw count data to
   transcripts per million (TPM) and applied a log[2](TPM + 1)
   transformation to the TPM values.

Quantification of PDCD1 High-Activity macrophage

   To quantify the PDCD1-related activity level of macrophages in the
   samples, we developed an activity score (PDCD1Score) based on the
   expression levels of the PDCD1 and CD64 genes. The calculation formula
   is as follows:
   graphic file with name d33e435.gif

   where exp represents the gene expression level. This score reflects the
   intensity of PDCD1-related activity in macrophages. Based on the third
   quartile (75th percentile) of the activity score, samples were
   categorized into two groups: the high-activity group (PD-1^high TAM)
   and the low-activity group (PD-1^low TAM).

Prognostic analysis

   To evaluate the survival difference between PD-1^high TAM and PD-1^low
   TAM, we used Kaplan-Meier analysis to generate survival curves and used
   the log-rank test to assess statistical significance between the two
   groups. Additionally, we integrated multiple clinical indicators and
   assessed the independent prognostic value of the PDCD1Score through
   multivariate Cox regression analysis. Forest plots were constructed to
   visually present the results of the multivariate analysis.

Single-cell RNA-seq data processing and analysis

   The Seurat package in R software was used for the basic analysis
   workflow. After preprocessing the data, a linear transformation was
   performed using ScaleData, followed by principal component analysis
   with RunPCA to identify the significant dimensions in the dataset.
   Clustering analysis was performed with FindClusters using the Louvain
   algorithm after constructing the shared nearest neighbor graph.
   Visualization was generated on a two-dimensional map using UMAP. Cell
   annotation files provided by the contributors and well-known marker
   genes of mouse cell lineage (Cd79a, Cd79b, Ms4a1, Fcmr, and Igkc for B
   cells; Col1a1, Dcn, Col1a2, and Tm4sf1 for CD45^− cells; Flt3, Cd80,
   and Cd209a for dendritic cells (DC); Cd14, Cd68, Adgre1, Apoe, and Lyz2
   for macrophages; Gzma, Gzmb, Ncr1, Klrb1c, and Nkg7 for natural killer
   cells (NK); Cd2, Cd3e, Cd3d, and Cd3g for T cells) [[68]35] were used
   to identify cell types in CRC. Differentially expressed genes (DEGs)
   with high discriminatory power across distinct cell types were
   identified using FindAllMarkers and visualized with scRNAtoolVis.

Functional and pathway enrichment analysis

   Functional enrichment analysis was performed to investigate the
   biological significance of the top 100 upregulated DEGs, which were
   selected based on their log fold change (logFC) values and filtered
   using an adjusted p-value threshold of < 0.05. The analysis was
   conducted with the clusterProfiler package in R.

   Gene Ontology (GO) enrichment analysis was carried out using the
   enrichGO function, covering three major categories: Biological Process
   (BP), Cellular Component (CC), and Molecular Function (MF).
   Additionally, pathway enrichment analysis was performed using the
   enrichKEGG function to identify significantly enriched pathways from
   the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. For both
   GO and KEGG analyses, statistical significance was determined by a
   p-value threshold of < 0.01 and a q-value threshold of < 0.05. The
   significantly enriched GO terms and KEGG pathways were further refined
   and analyzed to elucidate the potential biological functions and
   regulatory pathways associated with the upregulated DEGs.

Statistical analysis

   Experimental data are expressed as mean ± SD. Data from two
   experimental groups were analyzed using the t-test, while data from
   multiple groups were analyzed using one-way ANOVA. P-value of < 0.05
   was considered statistically significant (*P < 0.05, **P < 0.01,
   ***P < 0.001). Statistical graphs were created using GraphPad Prism
   10.1.2. Bioinformatics data were analyzed with RStudio 4.3.1, and data
   normality was assessed using SPSSAU.

Results

Single-cell and bulk transcriptomics reveals the anti-tumor effects of PD-1
expression on macrophages in colorectal cancer

   Given the potential impact of macrophage PD-1 expression on patient
   prognosis, we constructed a PDCD1Score based on the expression levels
   of PDCD1 and CD64, with Cd64 validated as a macrophage marker at the
   single-cell level (Fig. [69]1E). Patients were then stratified into
   PDCD1^high TAM and PDCD1^low TAM groups, and the results demonstrated
   that PDCD1Score effectively distinguished prognostic outcomes in COAD
   patients (Fig. [70]1A). Furthermore, multivariate Cox regression
   analysis, incorporating clinical variables such as sex, age, and
   pathological stage, identified PDCD1^+ macrophage activity score as an
   independent prognostic factor for COAD patients, as demonstrated in the
   forest plot (Fig. [71]1B).

Fig. 1.

   [72]Fig. 1
   [73]Open in a new tab

   CRC single cell and bulk transcriptome data reveal the effect of PD-1
   expression on macrophages in patient prognosis and function. (A)
   Kaplan-Meier survival curves for PDCD1^low TAM and PDCD1^high TAM
   patients. (B) Clinical characteristics of CRC patients and PDCD1score
   multivariate Cox regression analysis results. (C) UMAP visualization of
   all single cells, with colors representing different samples. (D) UMAP
   visualization of all single cells, with colors indicating annotated
   cell types. (E) UMAP visualization of Cd64 expression in macrophage
   subclusters. (F) Dot plot depicting marker gene expression across cell
   types, with point size indicating the proportion of cells expressing
   each gene and color gradient reflecting average expression levels. (G)
   Volcano plot illustrating significant DEGs across cell types, with
   upregulated genes in red and downregulated in blue; the top 5 for each
   category are labeled. (H) Stacked bar charts showing the proportion of
   cell components in all samples. (I) Enrichment bar plot of upregulated
   DEGs linked to biological functions in Pdcd1^+ macrophages, showcasing
   significant GO terms from BP, MF and CC, with validated pathways
   highlighted in red. CRC: Colorectal Cancer; DEGs: Differentially
   Expressed Genes; GO: Gene Ontology; BP: Biological Processes; MF:
   Molecular Functions; CC: Cellular Components

   To comprehensively investigate the expression of the immune checkpoint
   PD-1 on non-T lymphocytes and its impact on CRC progression, we
   performed an in-depth analysis of the publicly available single-cell
   sequencing dataset [74]GSE150970. After quality control and filtering
   steps, a total of 143,680 cells were retained for subsequent analysis.
   Due to the large sample size and potential batch effects, we examined
   the clustering pattern of the samples using a two-dimensional UMAP
   plot. The results showed minimal heterogeneity between the samples,
   with a good mixture of samples in the plot (Fig. [75]1C). We then
   annotated the cell types based on the expression of marker genes and
   the metadata provided by the dataset. The UMAP plot demonstrated the
   distribution of different cell types, including DC, B cells,
   macrophages, T cells, NK and CD45^− cells (Fig. [76]1D). A bubble plot
   showed the expression of typical marker genes for these cell types
   (Fig. [77]1F). Furthermore, we identified upregulated and downregulated
   DEGs in these six cell types (Fig. [78]1G), with most of the DEGs
   closely related to cell type-specific genes.

   We then used a stacked bar chart to display the cellular composition
   across different samples. The results showed significant differences in
   the proportions of cell types among the 30 CT26 tumor-bearing mouse
   model samples, reflecting the heterogeneity in cellular composition.
   Myeloid cells, particularly macrophages, exhibited a high infiltration
   proportion, with most samples showing more than 40% macrophage
   infiltration, while only five samples had less than 30% (Fig. [79]1H).
   This phenomenon suggests that excessive macrophage infiltration may
   have a potential impact on the TME and tumor progression. To delineate
   the impact of PD-1 expression on TAM functionality, we isolated 64,297
   cells from macrophage subsets and stratified them into Pdcd1^+ (high
   Pdcd1 expression) and Pdcd1^− (low/undetectable Pdcd1 expression)
   groups based on Pdcd1 levels (PD-1 protein, encoded by the Pdcd1 gene).
   Differential expression analysis was then performed between the two
   groups, followed by functional enrichment analysis of the upregulated
   genes to explore their potential biological significance. The
   enrichment analysis results indicated that Pdcd1 levels in macrophages
   were associated with their anti-tumor functions, primarily involving
   phagocytosis, polarization, type II interferon production, and
   regulation of monocyte proliferation (Fig. [80]1I, Fig. [81]S1A and B).
   These findings suggest that PD-1 may play a key role in the anti-tumor
   functions of macrophages.

The correlation between macrophage PD-1 expression and tumor progression as
well as anti-tumor function in EGFP^+CT26 tumor-bearing mouse model

   Based on the functional enrichment analysis results from single-cell
   sequencing in Fig. [82]1, we further validated the impact of PD-1
   expression levels in macrophages on their anti-tumor functionality
   through in vivo experiments. To facilitate subsequent detection, we
   developed an EGFP-overexpressing CT26 colorectal cancer cell line and
   implanted it subcutaneously into BALB/c mice (n = 19) to establish a
   tumor-bearing model (Fig. S2A). Punctate tumor nodules were first
   observed on day 7 post-transplantation, with continuous monitoring of
   tumor volume progression. On day 22 post-implantation, tumor tissues
   were harvested and photographed for documentation (Fig. [83]2A C).
   Macrophages were identified using F4/80 as a specific marker [[84]36].
   Flow cytometric analysis was then conducted to assess the expression
   profiles and co-expression patterns of PD-1, CD86, and EGFP (Fig. S2B
   and S2C).

Fig. 2.

   [85]Fig. 2
   [86]Open in a new tab

   The expression of PD-1 on TAMs correlates positively with tumor growth
   and infiltration, but negatively with phagocytosis and M1 polarization
   in tumor-bearing mice. (A) Construction of CT26 subcutaneous tumor
   mouse model; tumor volume measured on days 7, 13, 16, 19, and 22
   (n = 19). (B) SPSS test for normal distribution characteristics of
   F4/80^+, F4/80^+PD-1^+, F4/80^+CD86^+, and F4/80^+EGFP^+ flow results.
   (C) Changes of tumor volume in mice (n = 19). (D) Correlation analysis
   of TAM PD-1 expression and tumor volume, with sample order from low to
   high PD-1 expression (n = 19). (E) PD-1 expression in individual mice
   and its relation to tumor growth (n = 19). (F) Correlation analysis of
   TAM PD-1 expression (F4/80^+PD-1^+) and TAM infiltration (F4/80^+)
   (n = 19). (G) Correlation analysis of TAM PD-1 expression
   (F4/80^+PD-1^+) and TAM phagocytosis (F4/80^+EGFP^+) (n = 19). (H)
   Correlation analysis of TAM PD-1 expression (F4/80^+PD-1^+) and M1
   macrophage polarization (F4/80^+CD86^+) (n = 19). PD-1: Programmed
   Death-1; TAMs: Tumor-Associated Macrophages

   Flow cytometry data from 19 samples were first tested for normality
   using the Shapiro-Wilk test to validate the assumptions for statistical
   analysis [[87]37]. Results indicated that all four datasets followed a
   normal distribution (Fig. [88]2B). Pearson correlation coefficients (r)
   were then calculated between PD-1 expression (F4/80^+PD-1^+) and tumor
   volume, TAM infiltration (F4/80^+), M1-like polarization
   (F4/80^+CD86^+), and phagocytic activity of TAM (F4/80^+EGFP^+). In the
   CT26-driven colorectal cancer model, PD-1 expression on macrophages was
   positively correlated with tumor volume and TAM infiltration, while
   negatively correlated with TAM phagocytic function and M1 polarization
   status (Fig. [89]2D-G). Mice were divided into PD-1^high and PD-1^low
   groups based on PD-1 expression levels on TAM. Compared to the PD-1^low
   group, the PD-1^high group exhibited higher tumor growth rates and TAM
   infiltration, along with relatively lower phagocytic activity and M1
   polarization (Fig. [90]S3). These findings suggest that surface PD-1
   expression on TAMs is closely associated with tumor progression and
   impaired anti-tumor macrophage function.

The expression of PD-1 on the surface of BMDMs inhibits its phagocytic
function

   To investigate the impact of the TME on PD-1 expression in macrophages,
   BMDMs isolated from the femur and tibia of C57BL/6 mice were utilized
   as an in vitro experimental model (Fig. [91]3A). BMDMs were stimulated
   with CT26-conditioned medium (CT26-CM), and PD-1 expression was
   analyzed by flow cytometry. The results demonstrated that CT26-CM
   significantly upregulated PD-1 expression on the surface of BMDMs in a
   time-dependent manner, with expression levels progressively increasing
   over time (Fig. [92]3B). Consistent findings were observed in Raw264.7
   cells (Fig. [93]S4A). These data confirm that the TME induces PD-1
   upregulation on macrophages.

Fig. 3.

   [94]Fig. 3
   [95]Open in a new tab

   The effect of PD-1 expression on morphological changes of macrophages
   and phagocytic functions were verified. (A) Examination of
   morphological changes in primary BMDMs cultures at day 3 and day 5. (B)
   Flow cytometry analysis of PD-1 expression on BMDMs induced by CT26-CM.
   (C) Microscopic assessment of morphological differences among BMDMs
   across various groups (100x). (D, E) Flow cytometry and
   immunofluorescence assess BMDMs phagocytosis of green fluorescent
   microspheres under varying treatments. PD-1: Programmed Death-1; BMDMs:
   Bone Marrow-Derived Macrophages; CT26-CM: CT26-Conditioned Medium.
   *P < 0.05 by unpaired t-test

   To further explore the functional role of PD-1 in macrophages, PD-1
   expression was knocked down in BMDMs using siRNA (Fig. [96]S4B).
   Microscopic analysis revealed that PD-1-silenced BMDMs exhibited a
   significant increase in pseudopodia formation compared to the siRNA
   control (simock) group (Fig. [97]3C). To directly assess the impact of
   PD-1 knockdown on macrophage phagocytic function, BMDMs were
   transfected with PD-1-specific siRNA in the presence or absence of
   CT26-CM. Phagocytic activity was evaluated using green fluorescent
   microspheres, quantified by flow cytometry and visualization via
   immunofluorescence microscopy. Results demonstrated that PD-1 knockdown
   significantly enhanced the phagocytic capacity of BMDMs toward
   fluorescent microspheres in both untreated and CT26-CM-stimulated
   groups compared to their respective controls (Fig. [98]3D, E, and
   [99]S4C). These findings indicate that high PD-1 expression on BMDMs
   critically suppresses their phagocytic function.

Knockdown of PD-1 in BMDMs promotes their polarization toward the M1
phenotype and suppresses tumor invasion and migration

   Macrophage polarization is often closely associated with tumor
   progression [[100]38, [101]39]. In this study, CD86 and CD206 were
   selected as surface markers for M1- and M2-polarized macrophages,
   respectively. The polarization state was evaluated by the CD86/CD206
   ratio. Results showed that CT26-CM stimulation significantly reduced
   CD86 expression, increased CD206 expression, and markedly decreased the
   CD86/CD206 ratio (Fig. [102]4A), indicating that TME promotes
   macrophage polarization toward an M2 phenotype in vitro.

Fig. 4.

   [103]Fig. 4
   [104]Open in a new tab

   The expression of PD-1 on BMDMs promotes the polarization of
   macrophages M2 and ultimately promotes the invasion and migration of
   CT26 cells. (A) After CT26-CM stimulation, flow cytometry measures CD86
   and CD206 expression in BMDMs, calculating the CD86/CD206 ratio to
   indicate macrophage polarization. (B) In NC and CM groups, PD-1
   knockdown on BMDMs is followed by flow cytometry to measure CD86 and
   CD206 expression and calculate the CD86/CD206 ratio. (C) Collection of
   supernatants from NC simock, NC siPD-1, CM simock, and CM siPD-1 to
   stimulate CT26, followed by transwell assessment of CT26 invasion and
   migration in the four groups. PD-1: Programmed Death-1; CT26-CM:
   CT26-Conditioned Medium; BMDMs: Bone Marrow-Derived Macrophages. Data
   represent mean with SD. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001 by ordinary one-way ANOVA. *P < 0.05, **P < 0.01 by
   unpaired t-test

   Subsequently, PD-1 was knocked down in BMDMs. Flow cytometry analysis
   revealed that PD-1 silencing in macrophages promoted macrophage
   polarization toward the M1 phenotype in both untreated (NC) and
   CT26-CM-stimulated (CM) groups (Fig. [105]4B). In cancer research,
   reprogramming of M2-polarized macrophages toward the M1 phenotype holds
   significant antitumor potential [[106]40, [107]41]. These findings
   confirm that the TME upregulates PD-1 on macrophages, thereby driving
   M2 polarization, while PD-1 knockdown reverses this effect.

   To further validate the functional impact of PD-1 expression in BMDMs
   on tumor cells, supernatants from control (simock) and PD-1-silenced
   (siPD-1) BMDMs were collected and co-cultured with CT26 cells. Results
   demonstrated that supernatants from siPD-1 BMDMs significantly
   suppressed CT26 cells migration and invasion compared to control
   supernatants (Fig. [108]4C). Collectively, these findings suggest that
   PD-1 is a potential therapeutic target for reprogramming M2-polarized
   macrophages and inhibiting tumor progression.

The expression of IFN-γ signaling protein in BMDMs is directly regulated by
PD-1

   Functional enrichment analysis of Result 1 suggested that Pdcd1
   expression correlates with type II interferon production in
   macrophages. IFN-γ, a key member of the type II interferon family, is
   primarily produced by lymphocytes but can also be secreted by
   macrophages [[109]42, [110]43]. In this study, we investigated the
   changes in IFN-γ expression in BMDMs after CT26-CM treatment. Flow
   cytometry revealed that CT26-CM significantly reduced IFN-γ expression
   in BMDMs, with levels progressively decreasing over time
   (Fig. [111]5A). Strikingly, siRNA-mediated PD-1 silencing reversed this
   suppression of IFN-γ (Fig. [112]5B). Notably, PD-1 knockdown increased
   the frequency of CD86^+IFN-γ^+ and CD206^−IFN-γ^+ cell populations,
   highlighting stronger co-expression of IFN-γ with CD86 (Fig. [113]5C).
   These results demonstrate that PD-1 directly modulates IFN-γ signaling
   protein production in BMDMs, with IFN-γ exhibiting preferential synergy
   with CD86.

Fig. 5.

   [114]Fig. 5
   [115]Open in a new tab

   PD-1 expression of BMDMs affects IFN-γ signaling protein expression.
   (A) Flow cytometry analysis of IFN-γ protein expression in BMDMs
   stimulated by CT26-CM over time. (B) Knockdown of PD-1 in BMDMs within
   NC and CM groups, followed by flow cytometry measurement of IFN-γ
   protein expression. (C) Flow cytometry assessed the proportion of IFN-γ
   co-expressed with CD86 and CD206 in BMDMs before and after PD-1
   knockdown. IFN-γ: Interferon-γ; PD-1: Programmed Death-1; BMDMs: Bone
   Marrow-Derived Macrophages; CT26-CM: CT26-Conditioned Medium; DEGs:
   Differentially Expressed Genes. Data represent mean with SD. *P < 0.05,
   **P < 0.01, ***P < 0.001 by ordinary one-way ANOVA. *P < 0.05,
   ***P < 0.001 by unpaired t-test

PD-1 regulates macrophage polarization and anti-tumor function through
JAK2-STAT3 signaling pathway

   To elucidate the molecular mechanisms underlying PD-1
   expression-mediated regulation of macrophage function, we performed
   KEGG pathway enrichment analysis on the top 100 DEGs. The analysis
   revealed a significant association between PD-1 up-regulation in
   macrophages and the JAK-STAT3 signaling pathway (Fig. [116]6A).
   Previous studies have established the critical role of JAK2-STAT3
   signaling in maintaining macrophage homeostasis and regulating
   polarization [[117]44, [118]45]. Phosphorylation is a key
   post-translational modification that regulates protein activity. In the
   JAK-STAT3 pathway, STAT3 phosphorylation is essential for its
   activation and nuclear translocation, driving downstream
   transcriptional programs. Thus, the phosphorylation levels of JAK2 and
   STAT3 (p-JAK2 and p-STAT3) are commonly used as direct indicators of
   JAK-STAT pathway activation [[119]46]. Western blot analysis
   demonstrated that PD-1 knockdown in BMDMs significantly decreased
   p-JAK2/JAK2 and p-STAT3/STAT3 ratios, suggesting inhibition of
   JAK2-STAT3 pathway activation (Fig. [120]6B and Fig. [121]S5A).

Fig. 6.

   [122]Fig. 6
   [123]Open in a new tab

   PD-1 of BMDMs affects macrophage polarization, IFN-γ signaling protein
   production and tumor promotion through JAK2-STAT3 pathway. (A) KEGG
   pathway enrichment bubble chart of DEGs in Pdcd1^+ macrophages. (B)
   Western blot detected the protein levels of STAT3, p-STAT3, JAK2, and
   p-JAK2 after PD-1 knockdown of BMDMs. (C) Flow cytometry was used to
   detect the expression of macrophage polarization-related proteins CD86
   and CD206. (D) Flow cytometry was used to detect IFN-γ signaling
   protein production in macrophages in the three groups. (E) Collect
   supernatants from BMDMs treated with CM siPD-1, and CM
   siPD-1 + colivelin to stimulate CT26 and evaluate CT26 migration via
   transwell assays. (F) Collect supernatants from BMDMs treated with CM
   siPD-1, and CM siPD-1 + colivelin to stimulate CT26 and assess CT26
   invasion through transwell assays. IFN-γ: Interferon-γ; PD-1:
   Programmed Death-1; Kyoto Encyclopedia of Genes and Genomes (KEGG);
   BMDMs: Bone Marrow-Derived Macrophages; DEGs: Differentially Expressed
   Genes. Data represent mean with SD. **P < 0.01 by unpaired t-test.
   *P < 0.05, **P < 0.01 by ordinary one-way ANOVA

   To investigate whether JAK2-STAT3 signaling mediates the regulatory
   effects of PD-1 on macrophage function, Colivelin, a potent STAT3
   activator, was added to PD-1-silenced BMDMs treated with CT26-CM. Flow
   cytometry analysis revealed that colivelin treatment reversed the M1
   polarization shift induced by PD-1 knockdown, redirecting macrophages
   toward M2 polarization (Fig. [124]6C). These findings demonstrate that
   PD-1 regulates macrophage polarization through the JAK2-STAT3 pathway.

   While PD-1 silencing enhanced macrophage phagocytic capacity, colivelin
   was unable to reverse this effect (Fig. [125]S5B), suggesting that
   alternative pathways may be involved in PD-1-mediated regulation of
   phagocytosis. However, the up-regulation of IFN-γ induced by PD-1
   knockdown was fully suppressed by colivelin treatment (Fig. [126]6D and
   Fig. [127]S5C), supporting the STAT3-dependent regulation of IFN-γ
   production. We further evaluated the functional consequences by
   collecting BMDM-CM from differently treated macrophages
   (CT26-CM + siPD-1 vs. CT26-CM + siPD-1 + colivelin) to stimulate CT26
   cell lines. Notably, colivelin reversed the suppression of tumor cell
   migration and invasion induced by PD-1-silenced macrophages
   (Fig. [128]6E and F).

   Collectively, these results demonstrate that PD-1 modulates macrophage
   polarization, IFN-γ secretion, and ultimately governs the invasive and
   migratory capacities of CT26 cells through the JAK2-STAT3 signaling
   pathway.

Discussion

   Emerging evidence has expanded our understanding of PD-1, highlighting
   its role not only in T cells but also in macrophages [[129]25, [130]47,
   [131]48]. However, the role of PD-1^+ macrophages in regulating
   anti-tumor immunity remains underexplored. Given that current studies
   often focus on a single aspect of PD-1 function on macrophage surfaces
   [[132]49–[133]51], this study takes the lead in using a multi-omics
   strategy for a systematic analysis. Bulk transcriptomic analysis
   revealed significant correlations between PD-1 hyperexpression in
   macrophages and poor patient prognosis. Subsequent high-resolution
   single-cell transcriptomic analysis, complemented by in vitro and in
   vivo functional validation, demonstrated that PD-1 inhibits key
   antitumor immune functions of macrophages through the JAK2-STAT3
   signaling pathway, ultimately promoting tumor progression.

   To comprehensively elucidate the role of PD-1 in macrophage-mediated
   anti-tumor immunity, we conducted further validation in tumor-bearing
   models. For subsequent evaluation of the relationship between
   macrophage phagocytic activity and PD-1 expression levels, we generated
   a stable EGFP^+CT26 cell line and established EGFP^+CT26-BALB/c murine
   tumor models. Analysis revealed that PD-1 expression on macrophages
   within the CT26 tumor microenvironment positively correlated with tumor
   progression and TAM infiltration, but negatively associated with both
   phagocytic capacity and M1-polarization status of TAMs. To further
   decipher the regulatory mechanism of PD-1 on macrophage functionality,
   we performed in vitro investigations using CT26-CM to simulate the TME.

   Clinical data have emphasized the significant correlation between a
   decreased M1/M2 TAM ratio and poor prognosis in certain cancers
   [[134]18, [135]52–[136]54]. Targeting the polarization signaling
   pathways to reprogram TAM from pro-tumorigenic M2 phenotype to
   anti-tumor M1 phenotype represents a promising therapeutic strategy in
   cancer treatment. This phenotypic switching has been demonstrated to
   effectively remodel the TME and augment anti-tumor immunity. Therefore,
   identification of key mediators governing macrophage polarization may
   provide novel insights for immunotherapy. In the current study,
   CT26-CM-stimulated BMDMs exhibited elongated morphological features.
   McWhorter et al. have established that macrophage polarization status
   correlates with morphological alterations, wherein M2-polarized cells
   display more elongated morphology compared to their M1 counterparts
   [[137]55]. We therefore systematically investigated CM-induced
   polarization dynamics and their functional implications within the TME.
   Flow cytometric analysis confirmed that such conditioning promoted M2
   polarization concomitant with enhanced CT26 cell migration and
   invasion. Notably, genetic ablation of PD-1 in BMDMs effectively
   reversed M2 polarization and attenuated tumor cell motility. These
   findings collectively demonstrate that TME-driven PD-1 upregulation
   orchestrates macrophage M2 polarization, thereby facilitating tumor
   progression and metastasis.

   Intriguingly, parallel studies have revealed that adult worm
   excretory/secretory products-treated macrophages undergo M2
   polarization with concomitant PD-1 elevation, a correlation consistent
   with our observations [[138]56]. Gordon’s team further corroborated
   that PD-1 expression on TAMs inversely correlates with phagocytic
   capacity, demonstrating that PD-1/PD-L1 blockade enhances
   macrophage-mediated tumor clearance and suppresses neoplastic growth
   [[139]30]. Mechanistically, PD-1 upregulation in BMDMs activates tumor
   immune evasion programs through JAK2-STAT3 signaling axis.
   Pharmacological inhibition of this pathway successfully reversed
   polarization commitment. Our study advances the understanding of
   macrophage-TME crosstalk, establishing PD-1 as a critical regulator of
   JAK2-STAT3-mediated polarization switch that ultimately governs tumor
   cell dissemination.

   Macrophages can directly eliminate tumor cells through phagocytosis or
   induce apoptosis, thereby suppressing tumor progression at early
   stages. Sydney R. Gordon observed through electron microscopy that
   PD-1^+ TAMs exhibited a “foamy” appearance, mainly composed of
   undigested phagosomes and lysosomes, which filled the cytosol with
   lipid-rich content. These findings suggest that PD-1 expression in
   macrophages inhibits phagocytosis [[140]30]. In this study, from
   single-cell transcriptomics to flow cytometry and immunofluorescence,
   we have qualitatively and quantitatively verified the effect of PD-1 on
   macrophage phagocytosis, providing a more comprehensive and intuitive
   comparison of how PD-1 expression affects macrophage phagocytic
   ability. Previous studies have shown that macrophages can use filopodia
   to pull particles toward the cell body for phagocytosis [[141]57].
   Therefore, we hypothesized that the extension of pseudopodia in
   macrophages with downregulated PD-1 might be associated with changes in
   their phagocytic capacity.

   Our in vitro experiments demonstrated that PD-1, an inhibitory immune
   checkpoint protein expressed on macrophages, suppresses their
   phagocytic capacity. Interestingly, CT26-CM appears to activate the
   phagocytic function of BMDMs, suggesting that macrophage phagocytosis
   within the actual TME may undergo a transition from activation to
   exhaustion. In the complex TME, macrophages interact with diverse
   cytokines and factors, ultimately leading to inhibition of their
   phagocytic activity. Although two-dimensional in vitro models cannot
   fully replicate the intricate cellular interactions and signaling
   pathways of the TME, our study robustly highlights the impact of the
   single factor PD-1 on macrophage function. These individual factors
   collectively contribute to the complexity of the TME, and unraveling
   their interplay remains a primary objective of our research. In
   mechanistic investigations, we explored whether PD-1 influences
   macrophage phagocytosis via the JAK2-STAT3 signaling pathway; however,
   no statistically significant changes in phagocytosis were observed
   following modulation of this pathway. KEGG pathway enrichment analysis
   revealed that PD-1 upregulates the Fcγ receptor-mediated phagocytic
   pathway in macrophages. Thus, future studies will focus on elucidating
   the regulatory role of Fcγ receptors in PD-1-mediated modulation of
   macrophage phagocytosis.

   Integrative bioinformatics analysis further validated the involvement
   of the JAK2-STAT3 pathway in PD-1-regulated macrophage functions. We
   confirmed that PD-1 modulates the production of IFN-γ signaling
   proteins through this pathway. IFN-γ signaling is essential for
   pro-inflammatory activation of macrophages [[142]58]. Our study also
   demonstrated stronger co-expression of IFN-γ with CD86, an M1
   polarization marker. Additionally, evidence indicates that IFN-γ
   promotes tumor cell apoptosis and suppresses tumor growth [[143]59]. By
   modulating macrophage polarization states and enhancing IFN-γ signaling
   activity, we may develop more effective therapeutic strategies to
   counteract tumor immune evasion and improve patient outcomes.

   The clinical efficacy of PD-1/PD-L1 inhibitors as monotherapy in CRC
   remains to be improved, and enhancing patient response rates represents
   a persistent research priority. The immunosuppressive TME, particularly
   the potent immunosuppressive effects mediated by TAMs, has been
   critically linked to suboptimal responses to immunotherapy. Identifying
   key molecular targets and elucidating underlying mechanisms governing
   TAM-mediated immunosuppression may provide therapeutic opportunities to
   counteract their protumorigenic functions and improve the efficacy of
   PD-1/PD-L1 blockade. Current research has identified several potential
   regulatory targets across multiple domains, including metabolic
   reprogramming (e.g., glycolysis and lipid metabolism) [[144]60],
   signaling pathway modulation (e.g., CSF1R and STAT3 inhibition)
   [[145]61, [146]62], and phenotypic regulation (e.g., IRG1) [[147]63].
   Notably, CSF1R-a critical regulator of TAM survival and
   differentiation-has been targeted in clinical trials using inhibitors
   such as PLX3397 and BLZ945, which exert antitumor effects by depleting
   TAM populations or suppressing their protumorigenic functions
   [[148]64]. In this study, we specifically investigate PD-1 expression
   on TAMs as a potential therapeutic target to overcome immunosuppression
   in CRC. Unlike previous studies that primarily focused on PD-1 in T
   cells, our research, for the first time, comprehensively elucidates the
   diverse functions of PD-1 in macrophages, particularly its effects on
   phagocytic capacity, polarization states, and IFN-γ signaling pathway
   activity, ultimately shaping the tumor immune microenvironment and
   promoting tumor progression. Given the pivotal role of PD-1 in
   macrophage regulation, targeting TAM (such as JAK2-STAT3 pathway
   inhibitors, TAM reprogramming agents, or IFN-γ pathway activators) in
   combination with immune checkpoint inhibitors holds promise for
   enhancing the anti-tumor effects of ICIs, improving response rates to
   immunotherapy, and overcoming resistance to PD-1/PD-L1 monotherapy in
   certain patients. Specifically, in refractory CRC patients, where
   blocking PD-1/PD-L1 in T cells alone has shown limited efficacy,
   interventions targeting PD-1 on macrophages may achieve more sustained
   immune reprogramming, thereby improving clinical outcomes.

   In conclusion, this study integrates single-cell and bulk
   transcriptomic data, in vivo mouse models, and in vitro primary cell
   stimulation experiments to systematically reveal the crucial role of
   PD-1 on TAM in driving CRC progression and its underlying molecular
   mechanisms (Fig. [149]7A).

Fig. 7.

   [150]Fig. 7
   [151]Open in a new tab

   Macrophages with high expression of PD-1 inhibit their anti-tumor
   immune function. (A) Tumor progression is closely related to the
   expression of PD-1 in macrophages, and macrophages with high PD-1
   expression generally show M2 polarization, phagocytosis damage, and
   IFN-γ signaling protein synthesis block. Meanwhile, JAK2-STAT3
   signaling pathway is considered to be a key signaling pathway for
   macrophage dysfunction. Macrophages can regulate macrophage
   polarization and signal protein synthesis through this pathway to
   promote tumor progression. The images are drawn by Adobe Illustrator
   software. IFN-γ: Interferon-γ; PD-1: Programmed Death-1

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [152]12967_2025_6469_MOESM1_ESM.jpg^ (478.2KB, jpg)

   Supplementary Material 1: Supplement Fig.1 Differential expression
   analysis and functional enrichment network analysis of Pdcd1^+
   macrophages. (A) The volcano plot of DEGs between Pdcd1^+ macrophages
   and Pdcd1^- macrophages. (B) Gene-function network composed of 5 key GO
   terms. In the network, gray points represent genes, yellow points
   represent the enriched GO terms, and the size of GO nodes corresponds
   to the number of genes enriched in each term. DEGs: Differentially
   Expressed Genes; GO: Gene Ontology.
   [153]12967_2025_6469_MOESM2_ESM.jpg^ (285.4KB, jpg)

   Supplementary Material 2: Supplement Fig. 3 Grouping mice by median
   PD-1 expression from flow cytometry, comparing tumor growth curves for
   PD-1high and PD-1low groups, as well as macrophage PD-1 expression,
   phagocytosis, polarization, and infiltration within each group. PD-1:
   Programmed Death-1. Data represent mean with SD. *P<0.05, **P<0.01,
   ***P<0.001 by unpaired t-test.
   [154]12967_2025_6469_MOESM3_ESM.jpg^ (1.1MB, jpg)

   Supplementary Material 3: Supplement Fig.4 Validation of macrophage
   knockdown efficiency and phagocytic function. (A) Flow cytometry
   measurement of PD-1 levels in Raw264.7 cells stimulated with CT26-CM at
   0h, 12h, and 24h. (B) Flow cytometry, Real-time PCR and Western
   Blotanalyses of PD-1 protein and RNA expression in BMDMs post-siPD-1
   treatment. (C) Flow cytometry monitoring and statistical results of the
   phagocytic function of macrophages in four groups. PD-1: Programmed
   Death-1; CT26-CM: CT26-Conditioned Medium; BMDMs: Bone Marrow-Derived
   Macrophages. Data represent mean with SD. *P<0.05, **P<0.01, ***P<0.001
   by unpaired t-test.
   [155]12967_2025_6469_MOESM4_ESM.jpg^ (919.2KB, jpg)

   Supplementary Material 4: Supplement Fig.5 Phagocytosis of BMDMs was
   detected after STAT3 activation. (A) Western Blot analysis of
   JAK2-STAT3 pathway related protein expression after siPD-1. (B) The
   effect of STAT3 agonist on phagocytosis of BMDMs was detected by flow
   cytometry. (C) Statistical results for IFN-γ. BMDMs: Bone
   Marrow-Derived Macrophages; IFN-γ: Interferon-γ. Data represent mean
   with SD. ***P<0.001 by unpaired t-test.

Acknowledgements