Exosome RAB10 inhibits JAK1/STAT1 to hinder macrophage M1 polarization and promote tumor immune escape

Abstract

Exosomes are key mediators of communication between tumor cells and the tumor microenvironment(TME); however, the mechanisms underlying exosome-mediated crosstalk between tumor cells and macrophages remain largely unclear. This study investigated the effect of exosomal RAB10 on macrophage polarization and tumor growth. Mechanistically, RAB10 delivered by breast cancer cells binds to the interferon receptor IFNAR1 and inhibits JAK1/STAT1 pathway phosphorylation, thereby impeding M1 polarization and promoting M2 polarization. RAB10 expression was significantly upregulated in drug-resistant breast cancer cells and was correlated with poor patient prognosis. In vitro assays confirmed that RAB10 enhances cancer cell proliferation. In vivo knockdown of RAB10 suppressed tumor growth and reduced the expression of markers related to proliferation (Ki67, PCNA), invasion (MMP2), and epithelial–mesenchymal transition (Snail, Vimentin). Single-cell RNA sequencing revealed a marked decrease in the proportion of macrophages in the TME following RAB10 knockdown. This phenotypic shift increases the secretion of immunosuppressive factors such as PDL1, leading to reduced activity of CD8⁺ T cells. Animal studies further confirmed that combined targeting of RAB10 and PD-L1 produces a synergistic inhibitory effect on tumor growth. This study demonstrated that breast cancer cells can transfer RAB10 to macrophages via exosomes. RAB10 interacts with IFNAR1 to suppress the JAK1/STAT1 signaling pathway, thereby inhibiting M1 polarization and promoting M2 polarization of macrophages. Inhibition of RAB10, especially in combination with PD-L1 blockade, offers a promising strategy to enhance anti-tumor immunity and overcome therapeutic resistance in breast cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-026-02681-x.

Background

Breast cancer (BC) is one of the leading causes of cancer-related mortality among women worldwide. The emergence of drug resistance during treatment is largely driven by dynamic remodeling of the tumor microenvironment (TME) [1–3]. Although molecular subtyping based on ER, PR, and HER2 status provides a critical framework for clinical decision-making [4], the immunosuppressive TME mediated by tumor-associated macrophages continues to pose a significant barrier to overcoming therapeutic resistance [5]. TAMs exhibit substantial plasticity; their polarization toward the pro-tumoral M2 phenotype leads to the secretion of various immunosuppressive and pro-angiogenic factors, which not only impair anti-tumor T cell responses but are also strongly associated with poor prognosis and resistance to therapy [6–8]. Recent studies have highlighted the pivotal role of intercellular communication in TME-mediated regulation of TAM polarization and function through exosomal proteins and miRNAs-in regulating TAM polarization and function [9]. These findings provide new insights into drug resistance mechanisms and offer promising directions for therapeutic development.

As essential mediators of intercellular communication within the TME, exosomes actively deliver functional biomolecules that reprogram recipient cell phenotypes and behaviors. During breast cancer progression, tumor cells orchestrate a complex signaling network that drives malignancy via aberrant exosome release. These exosomes contribute to immune evasion by delivering cargo such as immune checkpoint proteins (e.g., PD-L1) and immunosuppressive factors (e.g., TGF-β), thereby directly suppressing CD8⁺ T cell activity [10, 11]. Moreover, exosomes derived from breast cancer cells have been shown to induce immunosuppressive differentiation in bone marrow-derived immune cells by promoting the expansion of myeloid-derived suppressor cells (MDSCs)—which further inhibit systemic anti-tumor immune responses [12]. In addition, specific non-coding RNAs carried by exosomes, such as circTRIM1 and miR-760, have been implicated in activating survival and metastasis-related pathways, including the PI3K/AKT pathway, thereby enhancing metastatic potential and promoting treatment resistance [13, 14].

The Rab GTPase family, a key regulator of intracellular vesicle transport, modulates various pathophysiological processes by mediating membrane trafficking [15]. Among them, RAB10 is a critical member. Although previous studies have primarily investigated its role in Alzheimer’s disease, its function in BC remains inadequately explored [16]. Recent research has demonstrated that RAB10 is upregulated in BC tissues and significantly enhances the malignant phenotype of tumor cells, including their proliferation, migration, and invasiveness [17]. Clinically, elevated levels of RAB10 in the peripheral blood of BC patients are notably correlated with the co-expression of key metastasis-promoting signaling molecules, TLR4 and NF-κB. This co-expression is associated with an increased risk of disease recurrence, greater propensity for metastasis, and poor clinical prognosis [18]. Nevertheless, the precise molecular mechanisms by which RAB10 facilitates metastasis in breast cancer, particularly its role in modulating the dynamic interactions between tumor cells and immune cells within the TME—remain a significant scientific question warranting further investigation.

This study is the first to demonstrate that RAB10 is directly transferred from tumor cells to macrophages via tumor-derived exosomes, reprogramming their functions to promote tumor proliferation and invasion. This finding not only elucidates a novel mechanism by which RAB10 mediates tumor–immune cell crosstalk through exosomal pathways but also identifies RAB10 as a critical molecular target for modulating tumor immune evasion.

Method

Cell culture

The human breast cancer cell line SKBR3, monocyte cell line THP-1, and mouse breast cancer cell line 4T-1 were all obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai) and authenticated via short tandem repeat (STR) DNA profiling. The taxol-resistant SKBR3 variant (SKBR3/PR) was developed in-house using a stepwise dose escalation method. All the cell lines were routinely tested for mycoplasma contamination. The cells were cultured in a humidified incubator at 37 °C and 5% CO₂. SKBR3, SKBR3/PR, and 4T-1 cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. THP-1 cells were cultured in RPMI 1640 medium supplemented with identical concentrations of FBS and antibiotics. THP-1 monocytes were seeded at a density of 1 × 10⁶ cells per well and stimulated with 100 ng/mL PMA (Sigma-Aldrich) for 24–48 h to induce differentiation into adherent macrophages, which were subsequently used in co-culture experiments.

Exosome isolation and identification

Exosomes were isolated using a commercial kit (Umibio, UR52120) following the manufacturer’s instructions. The cell culture supernatant (≥ 20 mL) was centrifuged at 4 °C for 10 min at 3000 × g. This step was repeated until no visible cell debris remained. The resulting supernatant was centrifuged at 4 °C for 10 min at 10,000 × g to remove residual impurities. The Exosome Concentration Solution (ECS reagent) was added to the clarified supernatant at a volume ratio of 5:20, mixed by vortexing for 1 min, and incubated at 4 °C for 8–24 h. Following incubation, the mixture was centrifuged at 4 °C for 60 min at 10,000 × g, and the supernatant was discarded. The pellet was resuspended in 1× PBS buffer (200 µL per 20 mL of the initial supernatant) and centrifuged again at 4 °C for 2 min at 12,000 × g to collect the crude exosome-containing supernatant. This was passed through an Exosome Purification Filter (EPF column) via centrifugation at 4 °C for 10 min at 3000 × g. Purified exosomes were collected from the bottom of the tube and stored at − 80 °C for subsequent use. Transmission electron microscopy (TEM) and nanoparticle tracking analysis were employed for exosome characterization and imaging.

Exosome labeling and tracing

Macrophages were induced to the M0 phenotype by PMA treatment for 24 h. Purified exosomes were labeled using an exosome fluorescent dye (Umibio, UR52302), according to the manufacturer’s protocol. The final dye concentration was 5 µM, and the exosome concentration was 1 µg/µL. After a 10 min incubation at room temperature, the labeled exosomes were added to M0 macrophages. After 24 h, fluorescence signals were visualized and imaged using a laser confocal microscope.

Western blotting

Extracellular vesicles or cells were lysed in RIPA buffer (Beyotime, P0013C) supplemented with PMSF (Beyotime, ST506). For the selected samples, a phosphatase inhibitor cocktail (Beyotime, P1081) was also added. Protein lysates were separated by SDS-PAGE for 1.5 h and transferred to a PVDF membrane by electrophoresis for 2 h. The membrane was blocked for 30 min using a rapid blocking solution, followed by overnight incubation with the primary antibody at 4 °C and then 2 h of incubation with the HRP-conjugated secondary antibody at 37 °C. Protein signals were detected by chemiluminescence. To detect additional targets, the membranes were stripped using Western Blot Fast Strip Buffer (Epizyme, PS107S), re-blocked, and re-probed with different primary antibodies. The following antibodies were used: CD63 (Abcam, ab134045), TSG101 (Abcam, ab125011), RAB10 (zenbio, R25515), CD80 (zenbio, R381650), CD86 (zenbio, R380350), ARG1 (zenbio, 680036), IFNAR1 (zenbio, R381718), IgG (zenbio, 550124), EEA1 (zenbio, R30011), p-JAK1 (Affinity, AF2012), JAK1 (Affinity, AF5012), p-STAT1 (Affinity, AF3300), STAT1 (Affinity, AF6300), and β-actin (Biosharp, BOAM-00010-A).

RNA extraction and quantitative real-time PCR

Total RNA was extracted using TRIzol reagent (Invitrogen, 15596018CN), and cDNA was synthesized using the HiScript II Q Select RT SuperMix kit. Quantitative real-time PCR was performed using ChamQ SYBR qPCR premix (Vazyme) according to the manufacturer’s protocol.

Cell function experiment

A total of 500 breast cancer cells were seeded in 6-well plates and cultured under standard conditions for 10 days. The cells were then gently washed twice with PBS, fixed in 4% paraformaldehyde for 30 min, and stained with 0.1% crystal violet for 20 min. Excess stain was removed by rinsing with running water, and the plates were air-dried at room temperature. The colony formation rate was determined by counting stained colonies. To evaluate the effect of macrophage-conditioned medium on cancer cell migration and invasion, 24-well Transwell chambers (8 μm pore size, Corning) were used. For the invasion assay, chambers were pre-coated with Matrigel to simulate the extracellular matrix; for the migration assay, uncoated chambers were used. After 24 h of incubation, cells that had migrated or invaded to the lower chamber were fixed with 4% paraformaldehyde and stained with 1% crystal violet. Three random fields from each insert were selected and counted to quantify cell migration and invasion.

Immunofluorescence (IF) staining

THP-1 monocytes were seeded at a density of 1 × 10⁶ cells per well in 24-well plates and incubated overnight with exosomes. Cells were then washed with PBS, fixed with 4% paraformaldehyde for 30 min, permeabilized with 1% Triton X-100 for 20 min, and blocked with 5% bovine serum albumin for 60 min. Cells were incubated overnight at 4 °C with an anti-RAB10 primary antibody, followed by incubation with HRP-conjugated goat anti-mouse or anti-rabbit secondary antibodies at room temperature for 60 min. After washing, cells were treated with TSA-520 dye for 15 min. Antibodies were then removed using antibody elution buffer, and cells were re-incubated with anti-IFNAR1 antibody and TSA-570 dye (zenbio, 18002). Nuclear staining was performed using DAPI for 15 min at room temperature. Fluorescence signals were captured using a laser confocal microscope.

Immunoprecipitation (Co-IP)

The cells were collected by centrifugation and washed twice with pre-cooled PBS (Elabscience, EA-IP-K007M). Add 1 mL of L1 lysate containing PMSF to every 0.5-1 × 10⁷ cells and incubate on ice for 10–20 min. After vortexing the lysis buffer until it becomes transparent, continue to incubate on ice for 30 min, centrifuge at 4 °C and 12,000 rpm for 10 min, and collect the supernatant. Take 40 µL of Protein A/G magnetic bead suspension (containing 10 µL of magnetic beads), gently wash twice with 500 µL of PBS, then add the specific antibody diluted with 500 µL of PBS, and incubate on a shaker at room temperature for 30 min. After washing the magnetic bead-antibody complex four times, add 400 µL of sample lysis buffer and incubate on a shaker at room temperature for another 30 min. Wash the magnetic beads four times with PBST, add a mixture of 40 µL of PBS and 10 µL of 5× loading buffer, heat at 95 °C for 5 min, and collect the supernatant by magnetic separation. Endogenous RAB10 co-immunoprecipitation was performed using magnetic beads conjugated with anti-RAB10 antibody, with IgG conjugated magnetic beads serving as the negative control.

Immunohistochemistry

Tissue samples were fixed in formalin for 24 h, embedded in paraffin, and sectioned at 4–5 μm thickness onto positively charged slides. Slides were baked at 60 °C for 40 min, deparaffinized in xylene (3 × 5 min), and rehydrated through a graded ethanol series. Antigen retrieval was performed using sodium citrate buffer (pH 6.0) in a pressure cooker for 2.5 min. After cooling, the sections were rinsed with PBS and blocked with 5% homologous serum at room temperature for 30 min.

For conventional immunohistochemistry, primary antibodies were applied and incubated overnight. Subsequently, the sections were incubated with species-specific enzyme-conjugated secondary antibodies, and the signal was developed using a DAB substrate kit. Staining was visualized under a bright-field microscope. The immunohistochemical staining was evaluated by an experienced pathologist using the Immunoreactivity Score (IRS). This system integrates the staining intensity (scored 0–3: negative, weak, moderate, strong) and the proportion of positive cells (scored 0–4: 0%, 1–25%, 26–50%, 51–75%, 76–100%), resulting in a final score ranging from 0 to 12.

For multiplex fluorescence immunohistochemistry, primary antibodies against CD4 and CD8a (for T cells) and iNOS and ARG1 (for macrophages) were applied and incubated overnight. Corresponding species-specific secondary antibodies were then added. After three washes with PBS, nuclei were counterstained with DAPI for 5 min. Slides were mounted using an anti-fade mounting medium. Fluorescence signals were imaged using a laser confocal microscope. The mean optical density (AOD) across different channels was measured using ImageJ software.

Single-cell RNA sequencing experimental method

Single-cell RNA sequencing data were analyzed using the R package Seurat. To ensure data quality, only genes expressed in at least three single cells were retained. Cells were excluded if they contained fewer than 200 or more than 8000 detected genes, or if over 25% of their transcripts were mitochondrial or ribosomal. Following these filtering criteria, 18,884 high-quality cells were retained from an initial total of 31,416. Normalization was performed using the ‘NormalizeData’ function. The normalized data were converted into a Seurat object, and the top 2,000 highly variable genes were identified using the ‘FindVariableFeatures’ function. Principal component analysis (PCA) was conducted via the ‘RunPCA’ function, and significant principal components (PCs) were selected based on JackStraw and variance ratios. For clustering, the ‘FindNeighbors’ and ‘FindClusters’ functions were employed, and cell distributions were visualized using UMAP or t-SNE. Cluster-specific gene expression was determined using Wilcoxon rank-sum tests with the ‘FindAllMarkers’ and ‘FindMarkers’ functions from the ‘scran’ R package. The ‘FeaturePlot’ function was used to visualize specific gene expression patterns, and pathway enrichment analysis was conducted using the ‘clusterProfiler’ function.

Flow cytometry

To isolate tumor samples containing macrophages, cells were first washed with pre-cooled PBS and then resuspended in approximately 100 µL PBS per 1 × 10⁶ cells. Pre-mixed macrophage phenotype-specific antibodies were added, including APC-labeled anti-CD86 (BioLegend, 159216), PE-labeled anti-CD206 (BioLegend, 141706), and PerCP-Cy5.5-labeled anti-CD11b (BioLegend, 101228). The cell suspension was thoroughly mixed and incubated at 4 °C in the dark for 30 min. Following incubation, 2 mL of pre-cooled PBS was added, and the samples were centrifuged at 1500 g for 10 min at 4 °C. The supernatant was carefully discarded, and the washing step was repeated once. The cell pellet was resuspended in PBS containing over 2% formaldehyde and fixed at 4 °C in the dark for 24 h. M1 and M2 macrophage subpopulations were subsequently sorted based on CD86 and CD206 expression.

To prepare T cell samples, tumor-derived cells were washed with pre-cooled PBS and resuspended in approximately 100 µL PBS per 1 × 10⁶ cells. Pre-mixed T cell-specific antibodies were added, including PE-Cy7-labeled anti-CD3 (BioLegend, 100220) and APC-labeled anti-CD8 (BioLegend, 200610). Samples were incubated at 4 °C in the dark for 30 min, followed by the addition of 2 mL pre-cooled PBS and centrifugation at 1500 g for 10 min at 4 °C. After discarding the supernatant, the washing step was repeated. Cells were then resuspended in PBS containing more than 2% formaldehyde and fixed at 4 °C in the dark for 24 h. Flow cytometric analysis was subsequently performed.

Animal experiments

Four-week-old female BALB/c mice were used for in vivo experiments. Parental 4T1 cells, 4T1 cells transfected with an empty vector (NC), and shRAB10-transfected 4T1 cells (4T1-KD) were prepared. Subsequently, 3 × 10⁵ cells from each of these three preparations were subcutaneously injected into mice to assess the impact of RAB10 knockdown on tumor growth. To further evaluate the combined effects of PD-L1 inhibition, an additional cohort was inoculated with 3 × 10⁵ 4T1 or 4T1-KD cells, with some cells co-expressing PD-L1. The experimental groups included: Control, inhibitor, RAB10-KD, and RAB10-KD+inhibitor. Mice in the inhibitor-treated groups received intraperitoneal injections of anti-PD-L1 antibody (20 mg/kg) every three days, starting from day 7 post-inoculation.Starting one week after inoculation, tumor length, width, and body weight were measured every two days. All mice were euthanized 21 days post-inoculation. Tumor volume was calculated using the formula: volume = width × length × (width + length)/2. Tumors were excised for histological analysis. The animal experiments have been reviewed and approved by the Animal Ethics Committee of Bengbu Medical College (Ethics Approval Number: [2023] 650).

Human monocytes and T cells isolation

PBMC Isolation: Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized human blood using Ficoll-Paque PLUS (density 1.077 g/mL) according to a standard protocol. Briefly, whole blood diluted 1:1 with PBS was layered onto Ficoll and centrifuged; buffy coat cells were collected, washed twice with PBS containing 2% FBS, and resuspended in separation buffer (PBS + 0.1% BSA + 2mM EDTA).

T Cell Isolation: T cells were negatively isolated from PBMCs using the Dynabeads™ Untouched™ Human T Cells Kit (Invitrogen, 11344D). Briefly, PBMCs were incubated with the antibody cocktail against non-T cells, followed by the addition of pre-washed magnetic beads to bind non-target cells; the non-T cell-bead complexes were removed by two sequential magnetic separations, yielding T cells of > 95% purity.

Monocyte Isolation and Stimulation: Monocytes were negatively isolated from PBMCs using the Monocyte Negative Isolation Kit (Dynal Biotech). Monocyte purity (> 95%) was confirmed by flow cytometry (CD14 expression). Isolated monocytes were resuspended in RPMI-1640 medium supplemented with 10% FBS. For M1 polarization, monocytes were stimulated with LPS(20 µg/mL). Cells were incubated at 37 °C under 5% CO₂ for 6–48 h. The experiment was reviewed and approved by the Ethics Committee of Bengbu Medical College (Ethics Approval Number: [2023] 236).

Statistical analysis

All data are presented as mean ± SEM. Statistical analyses were performed using GraphPad Prism 10. One-way analysis of variance (ANOVA) was employed to compare continuous variables across subgroups. A p-value of < 0.05 was considered statistically significant.

Results

RAB10 is highly expressed in drug-resistant breast cancer cells and promotes migration

The RAB protein family plays a critical role in exosome formation, and RAB10 has been implicated in the malignant transformation of epithelial cells in various cancers. This study focused on RAB10 due to its potential relevance in breast cancer. Kaplan–Meier survival analysis revealed that high RAB10 expression correlates with reduced overall survival in breast cancer patients (Fig. 1A). As drug resistance is a major contributor to treatment failure, we employed the previously established paclitaxel-resistant breast cancer cell line SKBR3/PR [19], which overexpresses drug resistance-associated proteins ABCG2 and PGP. CCK-8 assays confirmed that SKBR3/PR exhibits significantly higher drug resistance compared to the parental SKBR3 line [20]. Both ABCG2 and PGP belong to the ATP-binding cassette (ABC) transporter superfamily and facilitate drug efflux through ATP hydrolysis [21]. Western blotting was first conducted to verify that SKBR3/PR retains resistance to paclitaxel by assessing ABCG2 and PGP levels. In SKBR3/PR cells, the expression of ABCG2 and PGP was markedly elevated compared to SKBR3 (Fig. 1B). Subsequent analysis revealed higher RAB10 expression in SKBR3/PR compared to SKBR3 (Fig. 1C). Therefore, to study the function of RAB10, we knocked down RAB10 in SKBR3/PR cells and overexpressed RAB10 in SKBR3 cells. Western blotting confirmed the expected modulation of RAB10 expression, while transfection with empty plasmids had no effect (Fig. 1D–E). To evaluate the functional consequences of RAB10 expression, a colony formation assay was performed. RAB10 overexpression significantly enhanced the proliferative capacity of SKBR3 cells, whereas RAB10 knockdown in SKBR3/PR cells suppressed this phenotype (Fig. 1F). To validate these findings in vivo, we utilized 4T-1, a mouse triple-negative breast cancer cell line, and generated a RAB10-knockdown variant (Fig. 1G). A subcutaneous xenograft model was established, and tumor growth and body weight were monitored over three weeks. Mice injected with shRAB10 cells exhibited significantly reduced tumor volumes and weights compared to controls (Fig. 1H). Immunohistochemical analysis of tumor sections demonstrated that knockdown of RAB10 led to reduced expression of proliferation markers (Ki67, PCNA), invasion-related protein (MMP2), and epithelial-mesenchymal transition (EMT) markers (Snail, Vimentin) (Fig. 1I). Collectively, these in vitro and in vivo results demonstrate that RAB10 knockdown significantly inhibits breast cancer cell proliferation—as shown by colony formation, in vivo tumor growth, and proliferation marker expression—and downregulates invasion- and EMT-associated markers (MMP2, Snail, and Vimentin).

Fig. 1.

RAB10 is highly expressed in drug-resistant breast cancer cells and promotes their proliferation. A Kaplan–Meier survival analysis of RAB10 was performed using the Kaplan–Meier Plotter. B Western blotting was used to detect the expression of ABCG2 and P-glycoprotein (PGP) in SKBR3 and SKBR3/PR cells. C RAB10 expression levels in SKBR3 and SKBR3/PR cells were analyzed via Western blotting. D RAB10 was knocked down in SKBR3/PR cells, and the knockdown efficiency was validated by Western blotting. E RAB10 was overexpressed in SKBR3 cells, and the expression level was verified by Western blotting. F The proliferative ability of cells overexpressing or silenced for RAB10 was evaluated by a colony formation assay. G A 4T-1 cell line with RAB10 knockdown was constructed, and RAB10 expression was confirmed via Western blotting. H Images of dissected tumors from euthanized mice, along with tumor volume and weight measurements for each experimental group. I Immunohistochemistry (IHC) was conducted to evaluate the expression of Ki67, MMP2, PCNA, Snail, and Vimentin in the control and sh-RAB10 groups (scale bar = 100 μm). Statistical significance was analyzed by one-way ANOVA with Tukey’s test (n = 5 for mice studies; n = 3 for others). *p < 0.05, **p < 0.01, ***p < 0.001

Single-cell sequencing analysis reveals altered cell composition and changes in tumor tissue following RAB10 knockdown

To explore the tumor-suppressive effects mediated by RAB10, single-cell RNA sequencing was performed on tumor-bearing models as previously described. In the initial stage of the scRNA-seq analysis, we assessed cell characteristics, including total cell number, mitochondrial gene content, and ribosomal gene content. Following rigorous quality control to eliminate low-quality cells, 18,884 cells were retained for further analysis (Fig. 2A). After data normalization and correction for batch effects, the top 2,000 highly variable genes (HVGs) were selected (Fig. 2B). We calculated 20 principal components (PCS) to reduce the dimensionality of single-cell data. According to the Elbow Plot, the variance contribution curve tends to flatten after PC12, and the module discrimination of the PC Heatmap significantly decreases after PC12. Therefore, the first 12 PCS are selected to construct a low-dimensional space for downstream analysis (Fig. 2C and D). Subsequently, a K-nearest neighbor (kNN) graph was constructed based on these 12 PCS and visualized through uniform manifold approximation and projection (UMAP), ultimately identifying 22 cell clusters (Fig. 2E). Based on the CellMarker 2.0 database [22], these 22 clusters were annotated and merged into six major cell types according to the expression of classical marker genes: epithelial cells, endothelial cells, fibroblasts, T and NK cells, macrophages, and neutrophils. The reliability of this classification was confirmed by visualizing the expression distribution of typical marker genes using stacked violin plots (Fig. 2F). The UMP diagram shows the distribution patterns of six types of cells (Fig. 2G). We identified differentially expressed genes (|log₂FC| > 0.5, adj.p < 0.05) and performed functional enrichment, revealing significant activation in pathways critical to immune regulation: Th1 and Th2 cell differentiation, TNF signaling, Herpes simplex virus 1 infection, NOD-like receptor signaling, Epstein-Barr virus infection, T-cell leukemia virus 1 infection, and cytokine-cytokine receptor interaction (FDR < 0.05) (Fig. 2H-I). Macrophages emerged as the most abundant immune cells in the tumor microenvironment, with notable compositional shifts observed in the cell proportion plot (Fig. 2J). Furthermore, cell–cell communication analysis indicated a high degree of interaction between macrophages and cancer cells, identifying cancer cells as the primary signal senders and macrophages as major recipients (Fig. 2K). Consequently, macrophages were identified as the primary potential recipient cells for cancer cell-derived exosomes.

Fig. 2.

Macrophages are the most abundant immune cells in the tumor microenvironment and closely interact with tumor cells. A Single-cell RNA sequencing (scRNA-seq) and quality control identified 18,884 high-quality core cells. B The 2,000 most highly variable genes were selected for downstream analysis. C The top 20 principal components (PCs) from principal component analysis (PCA) are shown. D The heat map based on the top 20 principal components (PCA PCs) was included. The results showed that after PC12, the module discrimination decreased significantly. E Uniform Manifold Approximation and Projection (UMAP) identified 22 distinct cell clusters. F Marker gene expression levels in each cell cluster are displayed. G These 22 clusters were merged into six major cell types. H Multiple sets of volcano diagrams show the differentially expressed genes in each cell cluster. I Enrichment analysis was conducted using differentially expressed genes. J Stacked bar charts showing the components of each group of cell types. K String plots represent the intensity of ligand-receptor interactions among different cell populations

RAB10 is enriched in breast cancer cell-derived exosomes and transferred to macrophages

Exosomes were purified from the conditioned media of SKBR3 and SKBR3/PR breast cancer cell lines using a commercial exosome extraction and purification kit. Transmission electron microscopy (TEM) revealed that the isolated vesicles displayed the characteristic cup-shaped morphology of exosomes, with particle sizes predominantly ranging from 50 to 100 nm (Fig. 3A). Nano-flow cytometry (NanoFCM) further confirmed the particle size distribution, with exosomes from both cell lines measuring slightly above 50 nm (Fig. 3B). Western blot analysis validated the presence of exosomal markers CD63 and TSG101 in both SKBR3 and SKBR3/PR cell-derived exosomes (Fig. 3C, D). To determine whether RAB10 is present in breast cancer cell-derived exosomes, we assessed RAB10 expression levels in exosomes isolated from SKBR3 and SKBR3/PR cells. Following transfection of SKBR3/PR cells with RAB10-targeting shRNA, the level of RAB10 in exosomes was significantly reduced compared to the control and negative control groups. Conversely, RAB10 overexpression in SKBR3 cells significantly increased its level in exosomes relative to the control (Fig. 3E). These findings demonstrate that intracellular RAB10 levels directly regulate the quantity of RAB10 packaged into exosomes. Exosomes can transmit biological information to adjacent or distant cells after secretion, thereby facilitating intercellular communication within the tumor microenvironment [23]. Macrophages, which are the most abundant immune-related stromal cells infiltrating tumors, play a critical role in shaping the tumor microenvironment. Previous studies have demonstrated that exosomes can mediate protein transfer between cells [24]. To investigate whether RAB10 is transferred to recipient cells via exosomes, THP-1 macrophages were co-cultured with PKH67-labeled exosomes derived from SKBR3 and SKBR3/PR cells. Higher concentrations of exosomal RAB10 led to greater internalization by macrophages (Fig. 3F–G). Additional uptake assays confirmed that PKH67-labeled exosomes from both cell lines could be efficiently internalized by breast cancer cells (Fig. 3H–I).

Fig. 3.

RAB10 protein from breast cancer cells is transferred to macrophages via exosomes. A Transmission electron microscopy (TEM) images of exosomes derived from SKBR3 and SKBR3/PR cells (scale bar = 200 nm). B Nanoparticle characterization of purified exosomes from SKBR3 and SKBR3/PR cells using NanoFCM. C, D Western blot analysis detecting exosomal markers CD63 and TSG101 in cell lysates and purified exosomes from SKBR3 and SKBR3/PR cells. E Western blot analysis of RAB10 expression in exosomes and breast cancer cells treated with RAB10 short hairpin RNA (shRNA) or RAB10 overexpression plasmid. F–G THP-1 macrophages were co-cultured with exosomes at indicated doses. H–I Representative confocal microscopy images showing internalization of PKH67-labeled exosomes (green) derived from SKBR3 and SKBR3/PR cells by macrophages. Green puncta indicate internalized exosomes

Exosomes derived from RAB10-expressing breast cancer cells inhibit M1 macrophage polarization

To examine whether exosomes from breast cancer cells influence macrophage polarization, the human monocyte cell line THP-1 was utilized as a model. THP-1 cells were first differentiated into unprimed M0 macrophages using PMA. Subsequently, M0 macrophages were stimulated with LPS or IL-4 to induce polarization into classical pro-inflammatory M1 or anti-inflammatory M2 macrophages, respectively. The validity of the polarization model was confirmed through morphological analysis (Fig. 4A) and Western blot detection of the M1 marker iNOS (Fig. 4B). We then investigated the effect of exosomes from SKBR3 cells overexpressing RAB10 on macrophage polarization. Quantitative real-time PCR (qRT-PCR) revealed that macrophages co-cultured with exosomes from RAB10-overexpressing SKBR3 cells exhibited significantly increased mRNA levels of M2 markers (CD206, ARG-1, IL-10), alongside a marked reduction in M1 marker expression (INOS, CD86, TNF-α), compared to those co-cultured with exosomes from wild-type SKBR3 cells (Fig. 4C). In contrast, macrophages exposed to exosomes derived from SKBR3/PR cells with RAB10 knockdown displayed reduced M2 marker expression and elevated M1 marker expression (Fig. 4D). Western blotting was conducted to validate these findings at the protein level. Consistent with the qRT-PCR results, macrophages incubated with RAB10-overexpressing exosomes showed a significant increase in the M2 marker ARG-1 and a concurrent decrease in the M1 markers CD86 and CD80, relative to controls. Conversely, exosomes from RAB10-silenced SKBR3/PR cells induced the opposite expression pattern (Fig. 4E–F). Finally, in vivo immunohistochemistry further substantiated the in vitro observations. In tumor tissues derived from the RAB10 knockdown group (shRAB10), there was a decrease in M2-type macrophage infiltration and a corresponding increase in M1-type macrophage infiltration. Moreover, the expression of immune checkpoint molecules PD-L1 and LAG3 was significantly reduced in the shRAB10 group, indicating that RAB10 modulates the tumor immune microenvironment by promoting immunosuppressive features (Fig. 4G).

Fig. 4.

Exosomal RAB10 promotes macrophage polarization towards the M2 phenotype and suppresses M1 polarization. A Representative images of Mφ, M1, and M2 after induction. B Western blotting confirmed the expression of M1 and M2 polarization markers, validating the model. C qRT-PCR results showed that THP-1 cells co-incubated with exosomes from RAB10-overexpressing SKBR3 cells exhibited upregulation of M2 markers. D qRT-PCR results showed that THP-1 cells co-incubated with exosomes from RAB10-silenced SKBR3/PR cells displayed increased M1 marker expression. E, F Western blot analysis of polarization markers in macrophages treated with exosomes from SKBR3 or SKBR3/PR cells. G Immunohistochemistry analysis of tumor tissues from mouse models. Statistical significance was analyzed by one-way ANOVA with Tukey’s test (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001

Exosomal RAB10 inhibits the JAK1/STAT1 signaling pathway by binding to IFNAR1

To further elucidate the mechanisms by which tumor-derived exosomal RAB10 regulates M1 macrophage polarization, we analyzed sequencing data from macrophages. The interferon-γ response pathway was identified as the most significantly affected (Fig. 5A). The JAK1/STAT1 signaling pathway, a canonical downstream component of the interferon-γ response, is widely recognized for its involvement in macrophage polarization [25].

Fig. 5.

Exosomal RAB10 suppresses M1 macrophage polarization by inhibiting the JAK1/STAT1 signaling pathway. A Gene enrichment analysis of macrophage transcriptomes revealed significant involvement of the interferon-γ response pathway. B, C THP-1 and BMDMs cells were treated with exosomes for 48 h, followed by Western blotting to assess total and phosphorylated levels of JAK1 and STAT1. D The GRAMMresults database predicted an interaction between RAB10 and IFNAR1. E, F Immunoprecipitation (IP) of RAB10 or IFNAR1 from THP-1 and BMDMs lysates using specific antibodies; non-specific IgG was used as a negative control. Western blotting (IB) was performed to detect IFNAR1 and RAB10 in input and IP fractions. Molecular weights are indicated in kDa. Data represent at least three independent experiments per condition. G, H Confocal microscopy images showing co-localization of IFNAR1 and RAB10 in THP-1 and BMDMs cells. Statistical significance was analyzed by one-way ANOVA with Tukey’s test (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001

To explore the effect of exosomal RAB10 on this pathway, we performed Western blot analysis to examine the phosphorylation status of JAK1 and STAT1. As shown in Fig. 5B, exosomal RAB10 significantly reduced the levels of phosphorylated JAK1 and STAT1 (p-JAK1 and p-STAT1), without affecting the total expression levels of JAK1 and STAT1 in PMA-induced THP-1 macrophages (Fig. 5B). Similar results were obtained using macrophages isolated from human peripheral blood (Fig. 5C). We hypothesized that RAB10 modulates JAK1-STAT1 signaling through physical interaction with IFNAR1. Data from the GRAMMresults database indicated a potential interaction between RAB10 and IFNAR1 (Fig. 5D). To validate this interaction, we conducted Co-IP and fluorescence co-localization assays. Co-IP experiments were performed in bone marrow-derived macrophages (BMDMs) and THP-1 macrophages. Following specific immunoprecipitation of RAB10, IFNAR1 was detected, whereas it was absent in the negative control groups using non-specific antibodies (Fig. 5E–F). Confocal microscopy further confirmed co-localization of RAB10 and IFNAR1 in both THP-1 macrophages and human peripheral blood-derived macrophages (Fig. 5G–H). These findings demonstrate that exosomal RAB10 physically interacts with IFNAR1 and inhibits M1 macrophage polarization by downregulating IFNAR1, and suppressing activation of the JAK1/STAT1 signaling pathway.

RAB10 mediates the tumorigenic effects of macrophages

Having confirmed that exosomal RAB10 inhibits M1 macrophage polarization, we further investigated whether RAB10-enriched exosomes could promote breast cancer progression. It is well established that M1 macrophages exert tumor-suppressive effects, whereas M2 macrophages promote tumor progression. Single-cell suspensions from 4T-1 and 4T-1-shRAB10-KD mouse tumors were prepared and analyzed using Raman spectroscopy. Significant alterations were observed in several Raman spectral bands associated with key cellular components and metabolites between the two groups (Fig. 6A). The most pronounced differences were identified at 750 cm⁻¹ (nucleotides), 927 cm⁻¹ (proline and hydroxyproline), 1072 cm⁻¹ (DNA), and 1588 cm⁻¹ (lipids) (Fig. 6B). These results indicate that after silencing RAB10, the infiltration of M1 macrophages in the tissue increases. Exosomes were isolated via ultracentrifugation from equal volumes of serum-free, exosome-depleted medium collected from breast cancer cells with either RAB10 overexpression or knockdown, using equivalent initial cell densities. Macrophages treated with PMA and conditioned exosomes were assessed for cytokine expression using ELISA. Compared to controls, exosomes overexpressing RAB10 increased the expression of TGF-β, IL-10, and PD-L1, while suppressing TNF-α expression. In contrast, RAB10 knockdown reduced the expression of TGF-β, IL-10, and PD-L1 and enhanced TNF-α levels (Fig. 6C). To assess the influence of exosomal RAB10 on tumor cell invasion and migration, SKBR-3 and SKBR3/PR cells were pre-incubated with macrophage-conditioned supernatants. Transwell assays revealed that exosomes with RAB10 overexpression significantly enhanced, while those with RAB10 knockdown reduced, the migratory capacity of SKBR-3 and SKBR3/PR cells, respectively (Fig. 6D and E). These results indicate that RAB10-enriched exosomes derived from breast cancer cells promote M2 macrophage polarization and subsequently enhance the migration and invasion of breast cancer cells.

Fig. 6.

RAB10 induces macrophages to enhance migration and invasion of breast cancer cells. A Averaged Raman spectra of 4T-1 and 4T-1-KD cells, highlighting key Raman peaks used for cellular identification. B Raman shifts corresponding to nucleotides (750 cm⁻¹), proline/hydroxyproline (927 cm⁻¹), DNA (1072 cm⁻¹), and lipids (1588 cm⁻¹). C ELISA analysis of secreted factors (e.g., TGF-β, IL-10, TNF-α, PD-L1) in the macrophage-conditioned media. D Transwell assays assessing migration and invasion of SKBR3 cells co-cultured with macrophage supernatants. E Transwell assays assessing migration and invasion of SKBR3/PR cells co-cultured with macrophage supernatants. One-way ANOVA was performed. Each experiment included three biological replicates and three technical replicates. Statistical significance was analyzed by one-way ANOVA with Tukey’s test (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001

Therapeutic potential of targeting RAB10 in combination with immune checkpoint blockade in vivo

In previous experiments, we observed that macrophages secreted a variety of anti-inflammatory cytokines, especially a significant increase in PD-L1. PD-L1 is a ligand of the PD-1 receptor on the surface of T cells and inhibits the activation and proliferation of T cells by binding to PD-1 [26]. Based on these findings, we hypothesize that exosomes derived from breast cancer cells, especially those containing RAB10, may induce tumor-associated macrophages to adopt a pro-tumor phenotype. The elevated expression of M2 macrophages and multiple anti-inflammatory cytokines such as PD-L1 further inhibit the response of CD8⁺ T cells and promote immune escape. To test this hypothesis, macrophages were seeded at a density of 1 × 10⁵ cells per well in six-well plates. Following polarization, 40 ng of exosomes were added, and macrophage supernatants were collected after 24 h. These supernatants were then co-cultured with human peripheral blood T cells. The changing trend of the proportion of CD8 + T cells in the co-culture system was consistent with the change of RAB10 content in exosomes (Fig. 6A). To evaluate the therapeutic potential of RAB10 targeting in vivo, 4T1 and shRAB10 tumor-bearing mice were treated with PBS or PD-L1 inhibitors. Visual inspection of mice and tumor tissues revealed that tumors in both the 4T1 and shRAB10 groups treated with PD-L1 inhibitors were significantly smaller than those in the PBS group. Notably, the mouse body weight curves demonstrated that the combination of RAB10 knockdown and PD-L1 inhibition produced the most significant tumor regression (Fig. 6B). Subsequently, tumor tissues were enzymatically digested into single-cell suspensions for flow cytometry analysis, where macrophages were characterized using CD11b, CD86, and CD206. In the PD-L1 inhibitor + shRAB10 group, CD86 (M1 marker) was highest, while CD206 (M2 marker) was lowest, indicating that immune suppression in the tumor microenvironment was alleviated (Fig. 6C–D). We further analyzed T cell populations by assessing CD3⁺ and CD8⁺ expression. Compared with the PBS-treated group, CD8⁺ T cell levels were consistently increased across all treatment groups, with the most pronounced increase observed in the PD-L1 inhibitor + shRAB10 combination group (Fig. 6E). Fluorescence immunohistochemistry was performed to validate these findings by assessing CD4 and CD8 expression in T cells, along with iNOS and ARG-1 in macrophages across randomly selected microscopic fields (Fig. 6F). The staining results were consistent with those obtained via flow cytometry, reinforcing the robustness of the data. In conclusion, RAB10 plays a pivotal role in shaping the tumor immune microenvironment and represents a promising therapeutic target for breast cancer treatment, especially when combined with immune checkpoint blockade.

Fig. 8.

RAB10 from tumor-derived exosomes interacts with IFNAR1 on macrophages, inhibiting the JAK1-STAT1 pathway and subsequent M1 polarization. This impairs CD8⁺ T cell activation, fostering an immunosuppressive microenvironment

Discussion

Cancer proliferation is closely associated with intercellular communication within the TME [27]. Exosomes, as a key medium for such communication, regulate tumor progression by transferring various biomolecules [28]. Tumor-derived exosomes deliver proteins, lipids, and miRNAs to macrophages, reprogramming their gene expression and metabolic pathways. This reprogramming promotes the shift of macrophages from an anti-tumor M1 phenotype to a pro-tumor M2 phenotype, thereby facilitating tumor growth and progression [29–31]. Experimental data from this study demonstrate that exosomal RAB10 suppresses JAK1-STAT1 pathway activation in macrophages through physical interaction with IFNAR1, thereby inhibiting M1 polarization. This impaired macrophage polarization further attenuates CD8⁺ T cell activation, reshaping the immunosuppressive tumor microenvironment and ultimately promoting tumor immune escape (Fig. 7).

Fig. 7.

Exosomal RAB10 enhances PD-L1 secretion from macrophages and suppresses CD8⁺ T cell activity. A Supernatants from macrophages treated with exosomes derived from SKBR3, SKBR3-OE, SKBR3/PR, and SKBR3/PR-KD cells were co-cultured with peripheral blood-derived human T cells, followed by flow cytometric analysis. B Tumorigenesis experiments were conducted in mice divided into four groups: 4T1, 4T1 + PD-L1 inhibitor, 4T1-KD, and 4T1-KD + PD-L1 inhibitor. Tumors were imaged, and tumor volume and weight were quantified. C Flow cytometry labeling of M1 macrophages using CD11b and CD86. D Flow cytometry labeling of M2 macrophages using CD11b and CD206. E Flow cytometry analysis of CD8⁺ T cells using CD3 and CD8 markers. F Multiplex fluorescence immunohistochemistry staining of tumor tissues for CD4, CD8, iNOS, and ARG1. Statistical significance was analyzed by one-way ANOVA with Tukey’s test (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001

The activation of oncogenes in tumor cells plays a critical role in driving the delivery of tumor-derived exosomes to macrophages, inducing M2 polarization and thereby contributing to tumor progression and therapeutic resistance [32]. In this study, exosomes derived from breast cancer (BRCA) cells were efficiently internalized by macrophages. Subsequent experiments showed that both PMA-induced THP-1 cells co-cultured with BRCA exosomes and peripheral blood-derived macrophages treated with these exosomes exhibited phenotypic characteristics that promote tumor growth and metastasis. Notably, although the exosomes derived from SKBR3/PR (drug-resistant) cells did not significantly differ in size or quantity from those of the parental SKBR3 cells, the SKBR3/PR-derived exosomes demonstrated stronger tumor-promoting effects.

RAS-related binding protein 10 (RAB10), a member of the RAS-related GTPase family within the RAS superfamily [33], primarily regulates intracellular vesicle trafficking and signal transduction and plays a critical role in tumor development [34]. The expression level of RAB10 is strongly correlated with tumor cell proliferation, invasiveness, and metastasis, and its overexpression often portends a poor prognosis [17]. Our findings revealed that RAB10 was significantly upregulated in drug-resistant breast cancer cells and their exosomes. Single-cell sequencing analysis identified macrophages as the immune cell population most actively interacting with tumor cells, with the most significant phenotypic changes observed following RAB10 knockdown. Based on these results, we hypothesized that RAB10 inhibits M1 macrophage polarization in the breast cancer microenvironment through paracrine signaling. Furthermore, our previous clinical validation using a tissue microarray of 74 breast cancer patients confirmed that RAB10 is significantly upregulated in tumor tissues compared to adjacent normal tissues, and its elevated expression is associated with advanced disease stage [35].

To validate this hypothesis, we overexpressed RAB10 in BRCA cells and observed a significant reduction in M1 polarization in co-cultured macrophages. We isolated exosomes from SKBR3 and SKBR3/PR cells with either silenced or overexpressed RAB10 and used them to treat human monocytes. qRT-PCR, Western blotting, flow cytometry, and immunofluorescence analyses consistently indicated that exosomes from RAB10-overexpressing BRCA cells significantly suppressed M1 polarization, whereas exosomes from RAB10-silenced cells promoted it. This finding not only confirms the inhibitory effect of tumor cell-derived RAB10 on M1 macrophage polarization but also reveals a novel mechanism through which exosomal RAB10 mediates this regulation. Consistent with these results, our data showed that exosomal RAB10 influenced macrophage metabolite expression. M2-polarized macrophages secreted high levels of immunosuppressive factors, including TGF-β, IL-10, and PD-L1, whereas M1 macrophages expressed elevated levels of TNF-α. Moreover, the conditioned medium from M2 macrophages significantly enhanced the migration and proliferation of breast cancer cells.

The JAK1/STAT1 signaling pathway is central to cellular responses to interferons (e.g., IFN-γ) and various interleukins [36]. IFN-γ is a potent inducer of M1 macrophage polarization [37]. Upon binding to its receptors, IFNAR1 and IFNAR2, IFN-γ activates JAK1 and JAK2, which in turn phosphorylate STAT1 [38]. Phosphorylated STAT1 (Tyr701) forms homodimers that translocate to the nucleus, bind the γ-activated sequence (GAS) in promoter regions, and initiate the transcription of genes associated with M1 polarization [39–41]. The present study demonstrated that exosomal RAB10 inhibits the activation of the JAK1/STAT1 pathway by downregulating IFNAR1, thereby preventing M1 polarization. Additionally, we evaluated the therapeutic potential of targeting RAB10 in combination with PD-L1 blockade using a mouse tumor model. Our results confirmed that M2 macrophages, promoted by RAB10, suppressed CD8⁺ T cell activation and remodeled the tumor immune microenvironment to facilitate immune evasion.

Although this study revealed a new mechanism by which exosome RAB10 promotes immune escape by inhibiting M1 polarization in macrophages and verified the potential therapeutic value of targeting RAB10 in mouse models, the number and types of in vivo models used in the study were relatively limited. More importantly, this study has not directly analyzed the correlation between the RAB10 level of tumor tissue or blood-derived exosomes and the polarization status of macrophages (M1/M2 ratio) and the infiltration/activation degree of CD8⁺ T cells in the tumor microenvironment in a breast cancer patient cohort. Collectively, these findings suggest that RAB10 plays multiple roles in tumor cell signaling and behavior, representing a promising target for novel cancer therapies.

Conclusions

This study identifies RAB10, an exosomal protein derived from drug-resistant breast cancer cells, as a key effector molecule driving the formation of an immunosuppressive TME. We demonstrated that tumor-derived exosomal RAB10 is transferred to macrophages and targets IFNAR1. This effectively inhibits the activation of the JAK1/STAT1 signaling pathway, ultimately suppressing IFN-γ-induced M1 macrophage polarization. Consequently, macrophages are skewed towards a pro-tumor M2 phenotype that secretes immunosuppressive factors such as TGF-β, IL-10, and PD-L1, resulting in impaired CD8⁺ T cell activation. This reprogramming of the TME facilitates tumor immune evasion, growth, metastasis, and resistance to therapy. The findings elucidate a novel mechanism of macrophage polarization regulation mediated by tumor-derived exosomal RAB10 and provide a theoretical basis for the development of macrophage-targeted therapeutic strategies.

Abbreviations

TME

tumor microenvironment

BC

Breast cancer

ER

estrogen receptor

PR

progesterone receptor

HER2

human epidermal growth factor receptor

TAMs

tumor-associated macrophages

PD-L1

programmed death-ligand 1

SKBR3/PR

paclitaxel-resistant SKBR3 variant

PMA

phorbol 12-myristate 13-acetate

NanoFCM

nano-flow cytometry

IF

Immunofluorescence

Co-IP

Co-immunoprecipitation

IHC

Immunohistochemistry

ELISA

enzyme-linked immunosorbent assay

BMDMs

bone marrow-derived macrophages

UMAP

uniform manifold approximation and projection

kNN

K-nearest neighbor

PBS

phosphate buffer saline

FBS

fetal bovine serum

PCA

principal component analysis

PCs

principal components

HVGs

highly variable genes

t-SNE

t-distributed stochastic neighbor embedding

Authors’ contributions

Tang Guohui: Conceptualization, methodology, writing—original draft. Pang Bo : Formal analysis, Investigation. Yuting Liu: Formal analysis, software. Shaopeng Xu: Resources. Ruonan Li: Resources, Data curation . Zhu Chengle: Resources. Wu Qiong: Data curation . Ran Ruorong: Validation. Haotian Cai: Resources. Wenrui Wang: Supervision, Visualization. Changjie Chen: Supervision, Visualization. Qingling Yang: Writing – review & editing, Funding acquisition.T.G. conceived the overall framework, designed specific experimental methods, participated in most of the experiments, and was mainly responsible for writing the first draft of the paper, including the introduction, methods, results, and the core part of the discussion.P.B. and T.G.: Participated in the result analysis of multiple experiments, transformed them into readable data, and conducted statistical analysis on them, specifically including Figures 1 A, 1B, 1 C, 1D, 1E, 1 F, 1G, 6 C, 6D, and 6E. Specifically carried out key experiments, data collection or case investigation work, and participated in the preliminary interpretation of the data.Y.L. And T.G. were responsible for the analysis of single-cell data, including Figures 2 A, 2B, 2 C, 2D, 2E, 2 F, 2G, 2 H, 2I, 2 J, and 2 K.R.L.: and Z.C.: mainly involved in animal experiments and the use of Raman spectroscopy.S.X. Provided the clinical samples used in this study.W.Q. participated in the data processing of qPCR and ELISA experiments, mainly in Figures 4 A, 4B, 4 C, and 4D.R.R.: The key research results were reviewed.H.C.: The materials mainly involved in the preparation of the immunofluorescence experiment are shown in Figures 3 H and 3I.W.W. Provided decision-making and academic support at key nodes and assisted in optimizing the data charts presenting research results.C.C. provided crucial supervision and guidance during the research process and directed the production of charts.Q.Y.: Conducted multiple rounds of in-depth reviews of the paper manuscript, put forward key revision suggestions, enhanced the scientific nature, logic and language expression of the content, and was responsible for providing research funding support for this study. As the corresponding author, one usually also assumes the responsibility of reviewing the final manuscript, submitting it and communicating with the journal.All authors: Have read the final version of the paper and agree to submit it for publication.

Funding

The authors declare that they have received financial support for the research, writing, and/or publication of this manuscript. Anhui Province Postgraduate Academic Innovation Project (2024xscx136), the Anhui Higher Education Collaborative Innovation Program (GXXT-2022-064), and the Anhui Higher Education Excellent Research Innovation Team (2024AH010021).

Data availability

The data is provided in the manuscript and supplementary information files.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Bengbu Medical College.

Competing interests

The authors declare no competing interests.