NR4A3 potentials M1-like macrophage polarization to facilitate anti-tumor immune responses in breast cancer

Abstract

Breast cancer (BC) is commonly labeled a “cold tumor” due to its dense population of immunosuppressive cells, particularly M2-like macrophages, which contribute to its resistance to therapy. Thus, there is a pressing need to shift the macrophage polarization towards M1 and revitalize the tumor immune microenvironment (TIME) to improve BC prognosis. In this study, we leveraged published RNA-sequencing data and performed multiplex immunohistochemistry on clinical specimens to identify NR4A3 as a promising biomarker for favorable outcomes in BC. High NR4A3 expression correlates with an inflamed TIME, characterized by heightened T-cell infiltration and activation. NR4A3 was preferentially expressed in macrophages and fostered M1-like macrophage polarization through direct binding to p65, thereby enhancing NF-κB transcriptional activity. Overexpression of Nr4a3 in tumor-infiltrating macrophages significantly inhibited the growth of E0771 tumors in a syngeneic mouse model, accompanied by increased T-cell infiltration and elevated production of functional cytokines. Conversely, suppression of Nr4a3 in macrophages compromised T-cell recruitment and diminished their anti-tumor capabilities. Consistent with these findings, co-culture experiments involving human T cells and NR4A3-overexpressing THP1 cells further demonstrated enhanced T-cell functionality. Collectively, our findings uncover a novel role for NR4A3 in macrophage polarization and TIME remodeling, offering a potential biomarker for favorable BC prognosis and a therapeutic target to enhance immunotherapy efficacy.

Introduction

Breast cancer (BC) stands as the most commonly diagnosed malignancy and ranks as the third foremost contributor to cancer-associated mortality among women worldwide¹. Based on its molecular characteristics, BC is primarily classified into four molecular subtypes, including Luminal A, Luminal B, HER2-enriched, and triple-negative breast cancer (TNBC)². Currently, the efficacy of the therapeutic regimens for BC, including surgery, radiotherapy, chemotherapy, and hormone therapy, is limited by the emergence of drug resistance and pronounced side effects^3,4. Immunotherapy has revolutionized cancer treatment owing to its remarkable capacity to enhance the immune response, enabling the precise identification and eradication of malignant cells⁵. Nonetheless, its clinical application in BC treatment is still challenged by low response rates^6,7. The therapeutic effectiveness of immunotherapy is closely linked to the immune status in tumors, characterized as either ‘cold’ or ‘hot’^8,9. By inducing a pro-inflammatory tumor immune microenvironment (TIME), it becomes possible to enhance the responsiveness of immunotherapeutic interventions in BC patients.

Tumor-associated macrophages (TAMs), accounting for the largest fraction of TIME in BC, correlate with poor prognosis by inducing immune suppression¹⁰. By the expression of a series of inhibitory receptor ligands, including PD-L1, PD-L2, and HLA-E, TAMs have been suggested to negatively modulate T-cell and NK-cell activation^11,12. TAMs can also secrete several cytokines such as TGF-β and IL-10 to inhibit T-cell function or induce regulatory T-cell differentiation, displaying an immunosuppressive role¹². Nevertheless, a population of macrophages, imparting a pro-inflammatory M1-like phenotype, enhances the Th1 response and subsequently boosts the antitumor response¹³. In addition, as antigen-presenting cells, M1-like macrophages present tumor neoantigens to T cells, thereby activating T cells¹⁴. Therefore, the role of macrophages is highly dynamic and heterogeneous across tumors, malignant lesions, or patients, ranging from a pro-inflammatory M1-like phenotype to an anti-inflammatory (M2-like) phenotype¹⁰. The depletion of M2-like macrophages in the TIME and the reversion of M2-like macrophages to M1-like macrophages may be potential strategies for heating the TIME and improving the effectiveness of immunotherapy.

The nuclear receptor 4 A (NR4A) family, including NR4A1, NR4A2, and NR4A3, represents the orphan members of the steroid/thyroid/retinoid nuclear receptor superfamily. NR4As are widely recognized as molecular switches and act as transcription factors to regulate immune cell homeostasis, tumor cell proliferation, and cellular metabolism¹⁵. Recent studies have shown that the NR4As also contributed significantly to the immune response and tumor development¹⁶. In blood-derived tumors, NR4A3 was demonstrated to exert robust tumor suppressive function in aggressive lymphomas¹⁷. Similarly, NR4A3 could also attenuate the proliferation of lung cancer cells and upregulate pro-apoptotic genes to promote apoptosis via direct binding to p53¹⁸. In contrast, the absence of NR4As on chimeric antigen receptor T cells (CAR T cells) promoted tumor regression and prolonged patient survival in solid tumors¹⁹. Based on the controversial effects of NR4As in tumor development and treatment, the impact of the NR4A subfamily in BC remains to be further elucidated.

In this study, we established that NR4A3 expression was positively correlated with immunotherapy sensitivity and favorable clinical outcomes. Mechanistic exploration revealed that NR4A3 promoted M1-like macrophage polarization in an NF-κB-dependent manner and subsequently increased T-cell infiltration and cytotoxicity. Collectively, our findings have uncovered a novel property of NR4A3 in macrophages, suggesting a novel target for BC immunotherapy.

Results

The high expression of NR4A3 in BC is correlated with a satisfied immunotherapy response and a favorable prognosis

To assess the clinical relevance of the NR4A subfamily (NR4A1, NR4A2, and NR4A3) in BC, we analyzed overall survival (OS) and disease-free survival (DFS) using TCGA-BRCA datasets. Kaplan–Meier analysis revealed that neither NR4A1 nor NR4A2 expression levels significantly affected BC patient survival (Supplementary Fig. 1A). In contrast, patients with high NR4A3 expression exhibited significantly prolonged OS and DFS compared to those with low expression (Fig. 1A, B). This prognostic advantage was particularly evident in TNBC and Luminal B subtypes upon molecular subtype analysis (Supplementary Fig. 1B). To determine NR4A expression in tumor tissues, we examined 1118 BC tissues and 113 normal tissues from TCGA-BRCA cohorts. Tumors universally displayed reduced NR4A levels compared to normal tissues, with NR4A3 showing the most pronounced downregulation (Supplementary Fig. 1C). Additionally, among the four molecular subtypes, TNBC exhibited the highest NR4A3 expression (Supplementary Fig. 1D). It is well known that the status of the tumor immune microenvironment (TIME) is closely related to the prognosis of patients²⁰. To investigate the role of NR4A in TIME status, we analyzed the correlation between the expression of the NR4A receptors and immune profile using the “ESTIMATE” and “TIMER” databases. Interestingly, the level of NR4A3, other than NR4A1 or NR4A2, was positively correlated with ESTIMATE, immune, and stromal scores, and negatively correlated with tumor purity (Fig. 1C and Supplementary Fig. 1E). Furthermore, the analysis of the association between NR4A3 levels and immune subtypes of BC showed that C2 (IFN-gamma dominant) and C3 (inflammatory) subtypes had higher levels of NR4A3, whereas C4 (lymphocyte depleted) subtype had the lowest level of NR4A3 (Fig. 1D). To test whether NR4A3 can be a potential predictor for immunotherapy, we evaluated the correlation between NR4A3 levels and immunotherapy response using the immunophenoscore (IPS)²¹. As expected, patients with higher NR4A3 levels were significantly more sensitive to anti-PD-1 and/or anti-CTLA-4 treatment (Fig. 1E). Taken together, these findings suggest that NR4A3 is positively associated with inflamed tumor immune status, resulting in improvement of clinical outcomes.

[See PDF for image]

Fig. 1

The level of NR4A3 is positively related to the prognosis and responsiveness of immunotherapy in breast cancer.

A, B Kaplan–Meier curves showing the OS (A) and DFS (B) with high NR4A3 expression versus low NR4A3 expression in the TCGA-BRCA cohort. C Pearson correlation analysis of scores of ESTIMATE, immune cell, and stroma, and expression of NR4A subfamily (NR4A1, NR4A2, and NR4A3) in the TCGA-BRCA cohort. D Violin plot showing NR4A3 expression among different immune subtypes in the TCGA-BRCA cohort. The C2 subtype has the highest M1/M2-like macrophage polarization and the highest lymphocyte infiltration with the greatest TCR diversity. The C3 subtype is defined by elevated Th17- and Th1-related genes and low to moderate tumor cell proliferation. The C4 and C6 subtypes, characterized by Th1 suppression and high M2 response, confer the worst prognosis. E Differences in IPS between the high NR4A3 and low NR4A3 group for CTLA4_negative/PD-1_negative, CTLA4_negtive/PD-1_positive, CTLA4_positive /PD-1_negative, CTLA4_positive /PD-1_positive in the TCGA-BRCA cohort. pvalues were calculated by Pearson correlation analysis (C), Kruskal–Wallis (D), and Wilcoxon test (E).

NR4A3 augments CD4⁺ and CD8⁺ T-cell infiltration

To further characterize the alteration of BC TIME linked to NR4A3 levels, we performed the CIBERSORT algorithm to estimate the infiltration of immune cells in TCGA-BRCA datasets. The results showed that the abundance of tumor-infiltrating lymphocytes (TILs), including CD4⁺ memory T cells, CD8⁺ T cells, and follicular helper T cells, was significantly higher in the high NR4A3 expression group (Fig. 2A). Notably, NR4A3 expression correlated with an immune-activated microenvironment across all BC subtypes, particularly in TNBC and Luminal A/B (Supplementary Fig. 2A). We also employed the BC single-cell RNA-sequencing (scRNA-seq) data of 26 BC tissue samples from the GEO dataset (GSE176078). Cells were annotated by cell surface marker genes, and classified into seven major cell types, including T cells, B cells, fibroblasts, epithelial cells, endothelial cells, myeloid cells, and normal basal epithelial cells (Fig. 2B and Supplementary Fig. 3A). We divided these tissue samples into two groups by the level of NR4A3 mRNA expression (Supplementary Fig. 3B) and found that tumor tissues with high NR4A3 expression had a higher abundance of T cells (Fig. 2C). To further confirm this bioinformatic analysis, we performed immunohistochemistry staining (IHC) staining of NR4A3 and multiplex immunohistochemistry (mIHC) staining of CD4, CD8, and pan-cytokeratin (a marker for tumor cells) using clinical specimens from BC patients. The results revealed that more CD4⁺ and CD8⁺ T cells were infiltrated in tumors with high NR4A3 expression (Fig. 2D, E). Collectively, these data support that NR4A3 expressed in BC is positively associated with T-cell infiltration.

[See PDF for image]

Fig. 2

NR4A3 promotes the infiltration of CD4⁺ and CD8⁺ T cells.

A The CIBERSORT algorithm was used to assess the abundance of 22 immune cells between high NR4A3 and low NR4A3 groups in the TCGA-BRCA cohort. B UMAP plot showing the cell types in the microenvironment of the BC single-cell database (GSE176078), colored by cell types. C Boxplot showing the percentage of indicated cell types in the high NR4A3 and low NR4A3 samples (GSE176078). D Representative images of IHC staining of NR4A3 (left panel) and mIHC staining of CD4, CD8, pan-cytokeratin (PanCK), and DAPI (right panel) in high NR4A3 and low NR4A3 expression slides of BC (scale bar, 20 μm). The IHC staining was independently assessed by two pathologists without patient data, and the statistical analysis of NR4A3 was based on the IHC score (values, 0–12). Patients were divided into high- and low-expression groups according to IHC scores. E The number of CD4⁺ or CD8⁺ T cells in high NR4A3 and low NR4A3 expression slides of BC tissues within a 20 × field (n = 32). The mean number of CD4⁺ or CD8⁺ cells was quantified from four random fields of each section. p values were calculated by the Wilcoxon test (A, C) and Student’s t-test (E). ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

NR4A3 boosts infiltrating T-cell activation

Next, we sought to determine whether the level of NR4A3 was linked to the activation of TILs. We performed a correlation analysis using the TCGA-BRCA datasets and found that the NR4A3 expression was closely correlated with major histocompatibility complex (MHC) molecules and immunostimulatory molecules: the higher the level of NR4A3, the more MHC and immunostimulatory molecules were expressed (Fig. 3A, B). We further employed the data of T cells from the single-cell dataset to investigate the role of NR4A3 in the functions of TILs. A total of 595 differentially expressed genes (DEGs) (295 upregulated and 300 downregulated genes, abs logFC >1, p_adj value <0.05, Supplementary Table 1) were identified in the TILs of high NR4A3 tumors compared to the low NR4A3 tumors (Fig. 3C). Functional cytokines, including IFN-γ, IL-2, and TNF, and the immune-activated signaling molecules, including NFKBIZ, ICOS, IL21R, and NEK7, were significantly upregulated in the high NR4A3 group (Fig. 3C). Consistently, GSEA showed that the genes related to immune-activated pathways, including TNF-α signaling via NF-κB, interferon-gamma response, regulation of CD4 positive alpha/beta T-cell activation, and T-cell differentiation involved in immune response, were significantly enriched in the high NR4A3 group (Fig. 3D). Furthermore, we conducted mIHC staining in clinical specimens to verify these bioinformatic findings. The results showed that the frequencies of IFN-γ-producing CD4/8⁺ T cells and caspase 3-expressing tumor cells were increased in the high NR4A3 group (Fig. 3E, F), indicating an enhanced cytotoxic effect of T cells. Collectively, these data suggest that NR4A3 plays a critical role in the regulation of TIL activation and function.

[See PDF for image]

Fig. 3

NR4A3 enhances the activation of tumor-infiltrated T cells.

A, B The correlation between the level of NR4A3 and MHC molecules (A) and immune stimulatory molecules (B) in the TCGA-BRCA cohort. C Volcano plot of DEGs between high NR4A3 and low NR4A3 samples in a subpopulation of T cells from GSE176078. D The DEGs were related to TNF-α signaling via NF-κB, interferon-gamma response, regulation of CD4 positive alpha/beta T-cell activation, and T-cell differentiation involved in immune response between high NR4A3 and low NR4A3 samples by GSEA (GSE176078). E Representative images of IHC staining of NR4A3 (left panels) and mIHC staining of CD4, CD8, IFN-γ, pan-cytokeratin (PanCK), caspase 3 and DAPI (right panels) in high NR4A3 and low NR4A3 expression slides of BC (scale bar, 20 μm). The IHC staining was independently assessed by two pathologists without patient data, and the statistical analysis of NR4A3 levels was based on the IHC score (values, 0–12). Patients were divided into high- and low-expression groups according to IHC scores. F The numbers of IFN-γ-producing CD4⁺ or CD8⁺ T cells and the proportion of caspase 3⁺ tumor cells in high NR4A3 and low NR4A3 expression slides of BC tissues within a 20 × field (n = 32). The mean number of IFN-γ-producing CD4⁺ or CD8⁺ T cells and the proportion of caspase 3⁺ tumor cells were quantified from four random fields of each section. p values were calculated by Pearson correlation analysis (A, B) and Student’s t-test (F). *p < 0.05.

NR4A3 mainly distributes in M1-like macrophages

To clarify the distribution of NR4A3 in tumors, we examined the expression of NR4A3 in different cell types of tumors in the single-cell dataset. The results revealed that the NR4A3 was predominantly expressed in myeloid cells (Fig. 4A, B). Furthermore, the UMAP showed that CD68, a gene marker for macrophages, was closely co-expressed with NR4A3 according to the single-cell dataset (Fig. 4C). The analysis of other scRNA-seq datasets showed a similar phenomenon (Supplementary Fig. 4A). To verify that NR4A3 was preferentially expressed in macrophages, we detected the mRNA level of CD68 and NR4A3 in BC cell lines, including Luminal A cells (T74D and MCF-7), HER2-enriched cells (HCC1954) and TNBC cells (BT549 and MDA-MB-231), THP-1 monocytes, and THP-1-differentiated macrophages. We found that THP-1-differentiated macrophages expressed the highest mRNA level of NR4A3 compared to the other cells (Fig. 4D).

[See PDF for image]

Fig. 4

NR4A3 is located in macrophages, specifically M1-like macrophages.

A Violin plot showing the NR4A3 expression levels in different cell types (GSE176078). B Heatmap showing the NR4A3 expression levels in different cell types (GSE176078). C UMAP plot showing the CD68 and NR4A3 double-positive cells in the BC single-cell database (GSE176078). D The mRNA expression of CD68 and NR4A3 in MCF-7, T74D, HCC1954, BT549, MDA-MB-231, THP-1-monocytes, and THP-1-macrophages was determined by RT-qPCR (normalized with GAPDH mRNA). E The correlation between NR4A3 expression level and M1-like scores in a subpopulation of myeloid cells from GSE176078. F The correlation between NR4A3 expression level and M2-like scores in a subpopulation of myeloid cells from GSE176078. G CIBERSORT algorithm to assess the percentages of M0-like macrophages, M1-like macrophages, and M2-like macrophages in high NR4A3 and low NR4A3 groups in the TCGA-BRCA cohort. H Heatmap showing the expression of markers of M1-like macrophages in THP-1-derived macrophages after treatment with IL-4 or IFN-γ/LPS for 24 h (GSE140235). I The mRNA expression of iNOS, TNFA, CD206, and NR4A3 in THP-1-derived macrophages after treatment with IL-4 or IFN-γ/LPS for 24 h was determined by RT-qPCR (normalized with GAPDH mRNA). J Representative images of mIHC staining of NR4A3, CD68, and iNOS within BC tissue slides (scale bar, 20 μm). pvalues were calculated by Student’s t-test (I), Pearson correlation analysis (E, F), and Wilcoxon test (G). ns p > 0.05, ***p < 0.001.

Given that the macrophages harbor two distinct polarized states, namely M1-like (pro-inflammatory) and M2-like (anti-inflammatory), playing different roles in the TIME reshaping and tumor progression^10,12,13, we wondered whether NR4A3 participated in the macrophage polarization. To test this hypothesis, we leveraged the M1- and M2-like signature lists to assess the association of NR4A3 levels with the polarization of macrophages according to correlation analysis. The results found that the expression of NR4A3 was significantly and positively correlated with the M1-like scores, but not with the M2-like scores (Fig. 4E, F). A similar result was also observed in TCGA-BRCA via CIBERSORT (Fig. 4G). Meanwhile, we analyzed a GEO dataset containing IFN-γ/ lipopolysaccharide (LPS)-polarized M1-like macrophages and IL-4-induced M2-like macrophages. The expression of NR4A3 was much higher in M1-like macrophages compared to M0-like and M2-like macrophages (Fig. 4H). More importantly, THP-1 cells were differentiated into M1-like and M2-like macrophages with IFN-γ/LPS and IL-4, respectively, and their iNOS (a marker for M1), TNFA (a marker for M1), CD206 (a marker for M2), and NR4A3 mRNA levels were determined with RT-qPCR. In line with bioinformatic analysis, higher NR4A3 was expressed in M1-like THP-1 cells compared to M0-like and M2-like THP-1 cells (Fig. 4I). The results from mIHC staining verified that in clinical samples, M1-like macrophages exhibited coexpression of NR4A3 (Fig. 4J). Collectively, these results indicate that the NR4A3 is predominantly expressed in macrophages, specifically M1-like macrophages.

NR4A3 promotes M1-like macrophage polarization via the NF-κB signaling pathway

Our previous findings revealed a significantly elevated expression of NR4A3 in M1 macrophages, leading us to hypothesize that NR4A3 might promote M1 polarization. To test this, we silenced the expression of NR4A3 in THP-1 cells and found that knocking down NR4A3 substantially inhibited the levels of CD86 and CD80 following treatment with IFN-γ and LPS (Fig. 5A). In contrast, overexpression of NR4A3 substantially promoted the expression of these markers (Fig. 5B). In addition to the alteration of the surface markers, the gene expression markers for M1-like macrophages (TNFA, CD86, iNOS, IL1B, and IL6) were also decreased in NR4A3-silenced THP-1 cells and increased in NR4A3-overexpressed THP-1 cells (Fig. 5C, D). Therefore, these results suggest that NRA43 enhances M1-like macrophage differentiation. To decode the mechanism underlying the role of NR4A3 in M1-like macrophage polarization, GSEA was conducted using the myeloid cell data derived from the scRNA-seq dataset by comparing the differentially enriched signaling pathways between high NR4A3 myeloid cells and low NR4A3 myeloid cells. The results revealed that genes related to TNF-α signaling via the NF-κB signaling pathway were dramatically enriched in the high NR4A3 myeloid cells (Fig. 5E). To verify this observation, we assessed the NF-κB activation of THP-1 cells with and without NR4A3-overexpression. As expected, an increased level of phosphorylated p65 (p-p65), a marker for NF-κB activation, was observed in THP-1 cells with NR4A3 overexpression (Fig. 5F). Importantly, inhibiting NF-κB activation abrogated the enhancement of M1 differentiation induced by NR4A3 overexpression (Fig. 5G). The maintenance of p-p65 stability is critical for the enhancement of NF-κB activity^22,23. Therefore, we sought to determine whether NR4A3 could stabilize p-p65 by direct interaction. The AlphaFold3 algorithm was employed and revealed that NR4A3 could directly bind to p65 with multiple hydrogen bonding sites (Fig. 5H). Excitingly, endogenous immunoprecipitation (IP) and Immunofluorescence (IF) staining verified this observation (Fig. 5I–K). Together, these findings suggest that NR4A3 binds to p65 and subsequently strengthens NF-κB activation to facilitate M1-like macrophage polarization.

[See PDF for image]

Fig. 5

NR4A3 contributes to M1-like macrophage polarization via the NF-κB signaling pathway.

A Knocking down NR4A3 expression in THP-1 cells with siRNA, the production of CD86⁺ and CD80⁺ in CD68⁺ cells in THP-1-derived macrophages after treatment with IFN-γ/LPS for 24 h was determined by flow cytometry. B The production of CD86⁺ and CD80⁺ in CD68⁺ cells in THP-1-derived macrophages with or without NR4A3-overexpressing (NR4A3-OE) after treatment with IFN-γ/LPS for 24 h was determined by flow cytometry. C Knocking down NR4A3 expression in THP-1 cells with siRNA, the mRNA expression of NR4A3, TNFA, CD86, iNOS, IL1B, and IL6 in THP-1-derived macrophages after treatment with IFN-γ/LPS for 24 h was determined by RT-qPCR (normalized with GAPDH mRNA). D The mRNA expression of NR4A3, TNFA, CD86, iNOS, IL1B, and IL6 in THP-1-derived macrophages with or without NR4A3-OE after treatment with IFN-γ/LPS for 24 h was determined by RT-qPCR (normalized with GAPDH mRNA). E Dot plot of the HALLMARK enrichment analysis. The size of the dot represents the gene count, and the color of the dot represents the p-value (GSE176078). F The protein levels of NR4A3, p65, p-p65, and β-Tubulin in THP-1-derived macrophages with or without NR4A3-OE were measured after treatment with IFN-γ/LPS for 24 h. G The mRNA expression of iNOS, TNFA, CD86, IL1B, and IL6 in NR4A3-OE M1-like macrophage with NF-κB inhibitor BAY 11-7082 (5 μM, 24 h) treatment was determined by RT-qPCR (normalized with GAPDH mRNA). H Predicted protein–protein docking model between NR4A3 (slate cartoon) and NF-κB p65 (cyan cartoon). I, J Immunoprecipitation (IP) and western blot analysis of the NR4A3/p65 interaction in THP-1-derived macrophages after treatment with IFN-γ/LPS for 24 h. K The NR4A3/p65 interaction was analyzed using confocal microscopy in THP-1-derived macrophages after treatment with IFN-γ/LPS for 24 h (scale bar, 10 μm). p values were calculated by one-way ANOVA (A, C) and Student’s t-test (B, D, G). ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

NR4A3 expression in macrophages suppresses tumor growth by activating antitumor immunity

To investigate the critical role of NR4A3 in the antitumor capacity of macrophages in vivo, we established a syngeneic mouse tumor model by injecting E0771 cells into the fat pads of female C57BL/6 mice, either alone or co-injected with wild-type bone marrow-derived macrophage cells (BMDMs) or Nr4a3-overexpressing BMDMs (Fig. 6A, B). Notably, Nr4a3-overexpressing BMDMs significantly delayed E0771 tumor progression, while the control BMDMs could not (Fig. 6C–E). The proportion of M1 macrophages (iNOS⁺CD68⁺ cells or CD86⁺CD68⁺ cells) was significantly increased in the tumors with Nr4a3-overexpressing BMDMs, and the mean fluorescence intensity (MFI) of p-p65 of macrophages was also elevated (Fig. 6F–I). Furthermore, genetic knockdown of Nr4a3 in BMDMs significantly reduced M1-like polarization and enhanced M2-like polarization in vivo, as analyzed by flow cytometry (Fig. 6J–N).

[See PDF for image]

Fig. 6

Overexpressing NR4A3 in macrophages limited tumor growth.

A The schematic illustration of E0771 alone or co-injected with wild-type BMDMs or Nr4a3-overexpressing (Nr4a3-OE)/Nr4a3-knockdown BMDMs model with C57BL/6 mice. B The efficacy of Nr4a3 overexpression in BMDMs was determined by RT-qPCR (normalized with Gapdh mRNA). C–E Immunocompetent C57BL/6 mice were transplanted with E0771 alone or co-injected with wild-type BMDMs or Nr4a3-OE BMDMs for 4 weeks. The volumes of tumors. Mean values ± SEM (C), the images of tumors (D), and the weight of tumors (E) were quantified (n = 5). F Representative images of mIHC staining of CD68, iNOS, CD86, p-p65, and DAPI in tumor tissues derived from the syngeneic mouse tumor model (scale bar, 20 μm). G–I Quantification of iNOS⁺ CD68⁺ cells (G), CD86⁺ CD68⁺ cells (H), and the mean fluorescence intensity (MFI) of p-p65 of CD68⁺ cells (I), about two random fields (1 mm² per field) were selected under a microscope in each section. J The efficacy of Nr4a3 knockdown in BMDMs was determined by RT-qPCR (normalized with Gapdh mRNA). K–N The production of CD86⁺ cells (K), CD80⁺ cells (L), CD206⁺ cells (M), and CD163⁺ cells (N) among F4/80⁺ and CD11b⁺ cells in E0771 tumor-bearing mice co-injected with Nr4a3-knockdown BMDMs was examined by flow cytometry. p values were assessed by the Student’s t-test (B), one-way ANOVA (E, G–I, K–N), and two-way analysis of variance in C. ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

Next, we evaluated the status of TIME using IHC staining. The results showed a significant increase in the proportion of CD4⁺, CD8⁺, granzyme B⁺, and IFN-γ⁺ cells within tumors containing Nr4a3-overexpressing BMDMs, compared to tumors with control BMDMs (Fig. 7A, B). In parallel, knockdown of Nr4a3 in TAMs significantly impaired infiltration and cytotoxicity of CD4⁺ and CD8⁺ T cells (Fig. 7C, D). To further validate these findings, we cocultured primary human CD3⁺ T cells with THP-1 macrophages with or without overexpressing NR4A3, respectively (Fig. 7E). We found that the frequencies of IFN-γ-producing CD4⁺ and CD8⁺ T cells were significantly increased after cocultured with NR4A3-overexpressing macrophages, as well as TNF-α-producing CD4⁺ T cells (Fig. 7F, G). Overall, our results further demonstrate that expression of NR4A3 in macrophages effectively suppresses BC by boosting T-cell infiltration and functions.

[See PDF for image]

Fig. 7

Expression of NR4A3 in macrophages boosts immune response in vivo.

A, B Immunocompetent C57BL/6 mice were transplanted with E0771 alone or co-injected with wild-type BMDMs or Nr4a3-overexpressing (Nr4a3-OE) BMDMs for 4 weeks. Representative images of CD4, CD8, Granzyme B, and IFN-γ IHC staining in tumor sections (scale bar, 20 μm) (A) and numbers of indicated immune cells per mm² (B), about two random fields (1 mm² per field) were selected under a microscope in each section. C, D Immunocompetent C57BL/6 mice were transplanted with E0771 alone or co-injected with wild-type BMDMs or Nr4a3-knockdown BMDMs for 3 weeks. Representative images of CD4, CD8, Granzyme B, and IFN-γ B IHC staining in tumor sections (scale bar, 20 μm) (C) and numbers of indicated immune cells per mm² (D), about two random fields (1 mm² per field) were selected under a microscope in each section. E The schematic illustration of T cells co-culturing with THP-1 macrophages with or without NR4A3-overexpressing (NR4A3-OE). F The production of IFN-γ and TNF-α in CD4⁺ T cells was examined by flow cytometry. G The production of IFN-γ and TNF-α in CD8⁺ T cells was examined by flow cytometry. p values were assessed by the one-way ANOVA (B, D), and Student’s t-test (F, G). ns p > 0.05, *p < 0.05, ***p < 0.001.

Discussion

In this study, we substantiated that NR4A3 was associated with favorable tumor prognosis and high levels of activated CD4⁺ and CD8⁺ T-cell infiltration in BC. NR4A3 was predominantly expressed in macrophages and was able to promote M1-like macrophage polarization via the NF-κB signaling pathway activation. Specifically, overexpressing Nr4a3 in macrophages dramatically suppressed mouse tumor growth and inflamed the TIME. Our findings provide a novel understanding of the mapping of NR4A3 to the TIME and offer a potential therapeutic target for BC.

NR4As belong to the superfamily of nuclear receptors due to the unidentified physiological and endogenous ligand, the activation of which is generally transient, and the cellular outcome is stimulus- and cell-context dependent¹⁸. The target genes of NR4As, which regulate the cell cycle, apoptosis, inflammation, atherogenesis, metabolism, or DNA repair, have recently been implicated in tumorigenesis^24,25. Compared to the non-transformed parental cells, NR4As are upregulated in the transformed cervical cancer cell line^26,27. NR4A1 has also been identified as an important determinant of the hyperactivation of the pro-oncogenic TGF-β signaling pathway in BC²⁸. In contrast, overexpression of NR4A3 attenuates cancer cell proliferation and promotes apoptosis by increasing the expression of pro-apoptotic genes in both a p53-dependent and p53-independent manner¹⁸. Bioinformatics analyses of the GENT2 and the Kaplan–Meier plotter databases imply that a higher expression level of NR4As contributed to a better prognosis of BC²⁹. Therefore, the controversial views on the role of NR4As in tumors highlight an urgent need to precise the role and function of NR4As in BC remains unclear. Here, we revealed that the expression of NR4A3 but not NR4A1/2 was closely associated with better survival of BC by analyzing the TCGA database. Furthermore, tumors with higher NR4A3 expression exhibited greater sensitivity to ICB therapy. These observations substantially advance the knowledge of NR4As in BC.

NR4A3 displays distinct functions in a variety of cells and tissues, including neurons³⁰, vascular smooth muscle cells³¹, T lymphocytes³², dendritic cells³³, and tumor cells³⁴. In neuronal and vascular smooth muscle cells, NR4A3 plays a role in survival and proliferation^30,31. NR4A3 also contributes to TCR-mediated cell death and thymocyte-negative selection, suggesting a link between NR4A3 and T-cell development³². Toll-like receptor (TLR) mediated stimulation markedly upregulates NR4A3 in dendritic cells, and the knockdown of NR4A3 in dendritic cells impairs T-cell proliferation and IL-2 production³³. In addition, NR4A3 hinders epithelial-mesenchymal transition, migration, and invasion of tumor cells by inhibiting the MAPK/ERK signaling pathway in malignant breast cells³⁴. Therefore, mapping NR4A3 expression on specific cell types may provide a better understanding of its role in BC TIME. Here, we proposed that NR4A3 was predominantly expressed in macrophages of BC. A previous study has reported that the presence of NR4A3 increased the generation of anti-inflammatory genes in macrophages³⁵. In this study, we found that NR4A3 was associated with M1-like macrophage polarization in BC. The cytotoxic function of T cells cocultured with macrophages overexpressing NR4A3 was significantly enhanced. Also, tumors with Nr4a3-overexpressing macrophages showed reduced growth rates compared to those with WT macrophages in a BC mouse model. Taken together, our results significantly expand our understanding of the immunoediting effects of NR4A3 in BC macrophages.

Research has established that NR4A3 is crucial for toll-like receptor (TLR)-mediated NF-κB activation of dendritic cells³³. Its elevated expression in cartilage is implicated in osteoarthritis development by amplifying IL-1β-induced NF-κB activation³⁶. Thus, these studies indicate a propensity for NR4A3 to promote inflammatory responses via the NF-κB signaling pathway. Here, leveraging GSEA of TCGA-BRCA data and scRNA-seq data, we revealed that the majority of upregulated genes were enriched in the activation of the NF-κB signaling pathway within NR4A3 highly expressed macrophages. NR4A3 directly bound to p65 and enhanced NF-κB signaling activity in macrophages, which promoted M1-like macrophage polarization both in vitro and in vivo. Inhibition of its activity significantly reversed the polarization of M1-like macrophages after overexpression of NR4A3. Thus, we identified the underlying mechanism governing macrophage polarization induced by NR4A3.

Conclusively, we uncovered that NR4A3 promotes M1-like macrophage polarization via NF-κB signaling, establishing the inflammatory TIME and contributing to a satisfactory clinical outcome in BC. These findings not only reveal a potential biomarker for predicting the prognosis of BC but also offer a deeper understanding of the role of NR4A3 in BC, suggesting that NR4A3 may be a promising therapeutic target in BC.

Methods

Tissue collection

Tumor tissues were collected from 32 patients with BC who underwent tumorectomy (Supplementary Table 2). The use of all the patient material was approved by the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (approval number: TJ-IRB202502105), and an informed consent document was signed by all the participants in the study. This study was performed in accordance with the Declaration of Helsinki.

Cell culture

Human BC cell lines (BT549, MCF-7, T-47D, HCC1954, and MDA-MB-231), HEK293T cells, RAW264.7 cells, E0771 cells, and THP-1 monocytes were purchased from the China Center for Type Culture Collection (CCTCC). MDA-MB-231, HEK293T, E0771, and RAW264.7 cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM; Gibco, 11965092) containing 10% fetal bovine serum (FBS) (Corning, 35-076), 1% penicillin-streptomycin solution (PS) (Servicebio, G4003). BT549, T-47D, HCC1954, and THP-1 cells were cultured in RPMI 1640 medium (Gibco, 11875093) supplemented with 10% FBS and 1% PS. MCF-7 cells were cultured in MEM medium (Gibco, 11095080) supplemented with insulin (10 μg/mL), 10% FBS, and 1% PS. Cells were cultured in a thermostatic incubator at 37 °C and 5% CO₂. After resuscitation, cells were used within 10–15 passages or 3 months. When cells reached 70%–80% confluence, we harvested these cells for further experiments.

RNA interference

Cells were cultured to approximately 60% confluence before transfection. Small interfering RNA (siRNA) were transfected into the cells using Lipofectamine 3000 (Invitrogen, L3000075), following the manufacturer’s protocol. The transfected cells were grown in an antibiotic-free medium for 24 h and then subjected to specific treatments. The knockdown efficiency was examined by qRT-PCR analysis. The siRNA sequences were obtained from Wuhan Qinda Biological Technology (Wuhan, China).

The siRNA sequences used were as follows (5′-3′):

si-h-NR4A3_001 CGAGCAACUACGAACUCAATT

si-h-NR4A3_002 CAAGAGAACAGUGCAGAAATT

si-control TTCTCCGAACGTGTCACGTDTDT

Lentiviral transduction

Full-length cDNA of NR4A3 and Nr4a3 was cloned into the pLV3-CMV-N-CopGFP-puro plasmid. The shRNA against Nr4a3 was cloned into the pLKO.1-puro plasmid. For the lentivirus-mediated systems, the lentiviral transfer plasmid was transfected into HEK293T cells with psPAX2 (Addgene, 12260) and pMD2G (Addgene, 12259) at a ratio of 2:1:1. Supernatants of HEK293T cells containing lentivirus were collected 2 days later and used to infect THP-1 or primary BMDMs. The overexpression or silencing efficacy was confirmed by RT-qPCR and western blot. The plasmids used above were purchased from Wuhan Qinda Biological Technology (Wuhan, China).

The sequences targeting Nr4a3 were as follows (5′-3′):

sh-Nr4a3_001 CGGCCTTTGATCAAGATGGAA

sh-Nr4a3_002 GCAGACTTATGGCTCGGAATA

sh-Nr4a3_003 CCTCCGATCTGTATGATGAAT

THP-1 differentiation, polarization, and treatment with NF-κB inhibitor

THP-1 monocytes were differentiated into M0-like macrophages in the presence of 100 ng/mL phorbol 12-myristate 13-acetate (PMA, MedChemExpress, HY-18739) for 24 h. To obtain M1-like macrophages, M0-like macrophages were incubated with 100 ng/mL LPS (Sigma-Aldrich, L2630) and 10 ng/mL IFN-γ (PeproTech, 300-02) for 48 h. For M2-like macrophages, M0-like macrophages were polarized by 20 ng/mL IL-4 (PeproTech, 200-04) for 48 h. For NF-κB inhibition, cells were treated with NF-κB inhibitor BAY 11-7082 (Selleck, S2913) at 5 μM for 24 h during M1 polarization.

Isolation of primary T cells and cocultured models

Peripheral blood was obtained from healthy donors provided with informed consent. Human peripheral blood mononuclear cells were isolated using Ficoll gradient purification (TBD sciences, LTS1077) and sorted with human CD3 positive selection kit II (STEMCELL Technologies, 17851). Purified T cells were activated in complete RPMI-1640 medium supplemented with human CD3/CD28 T-cell activator (STEMCELL Technologies, 10971) and 10% FBS for 5 days, and expanded in complete RPMI-1640 medium supplemented with 20 ng/mL recombinant human IL-2 (Peprotech, AF-200-02) and 10% FBS. T cells were cocultured with macrophages and treated with Cell Activation Cocktail (with Brefeldin A) (Biolegend, 423303) for 4 h to detect intracellular cytokines.

RNA isolation and RT-qPCR

Total RNA was extracted using the FastPure RNA Isolation Kit V2 (Vazyme, RC112-01) according to the manufacturer’s instructions and then reverse-transcribed into cDNA with HiScript SuperMix (Vazyme, R222-01). The RT-qPCR analysis was performed using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711-02) on a LightCycler 480 (Roche, Basel). The mRNA levels of gene expression were analyzed by the 2^−ΔΔCt method. The primers used in the study were summarized in Supplementary Table 3.

Mouse tumor models

Bone marrow cells were harvested from donor C57BL/6 WT mice and cultured in the medium containing 50 ng/mL M-CSF (Peprotech, 315-02) for 7 days to generate BMDMs. For macrophage adoptive transfer experiments, female C57BL/6 mice were subcutaneously (s.c.) injected with 5 × 10⁵ E0771 cells either alone or co-injected with 2 × 10⁵ wild-type BMDMs or Nr4a3-overexpressing/Nr4a3-knockdown BMDMs. The BMDMs and E0771 cells were physically mixed immediately just before in vivo administration. Tumor growth was measured every other day after tumor inoculation using digital calipers.

Immunohistochemistry staining

After deparaffinization in xylene and rehydration through graded alcohol solutions, the tissue slides underwent antigen retrieval for 15 min. To block nonspecific binding, 5% BSA was applied for 30 min at 37 °C. The slides were then incubated with anti-NR4A3 (Atlas Antibodies, HPA043360), anti-CD8 alpha (Abcam, ab217344), anti-CD4 (Abcam, ab183685), anti-granzyme B (Abcam, ab255598), and anti-IFN-gamma (R&D Systems, MAB485) primary antibody overnight at 4 °C. After three washes with TBST, the slides were incubated with HRP-conjugated secondary antibodies for 1 h. The HRP activity of the secondary antibodies was determined using the DAB kit (ZLI-9017, ZSGB-BIO).

Multiplex immunohistochemistry staining

The mIHC was performed on tissue slides according to the manufacturer’s instructions (Treble-Fluorescence immunohistochemical mouse/rabbit kit, Immunoway, RS0037). Slides were stained with CD4 (Abcam, ab133616), CD8 (Abcam, ab101500), IFN-γ (Abcam, ab231036), CD68 (Abcam, ab233172), NR4A3 (Atlas Antibodies, HPA043360), iNOS (Abcam, ab283655), CD68 (Abcam, ab283654), CD86 (Invitrogen, 14-0862-85), phospho-NF-κB p65 (Cell Signaling Technology, 3033), pan-Cytokeratin (Abcam, ab7753), caspase 3 (Abcam, ab4051), and DAPI in proper order. All the primary antibodies were incubated overnight in a cryogenic laboratory for more specific antigen-antibody binding. The slides were then blocked with an antifade mounting medium and scanned with PHENO IMAGER HT (Akoya Biosciences, America).

Immunofluorescence staining

Cells were fixed in 4% paraformaldehyde or absolute ethanol and permeabilized with 0.2% Triton-X in PBS. The anti-NR4A3 (Origen, TA804893) and anti-phospho-NF-κB p65 (Cell Signaling Technology, 3033) primary antibodies were incubated with the cells overnight at 4 °C. After three washes with PBS, cells were incubated with anti-rabbit IgG (Alexa Fluor 488 Conjugate) (Cell Signaling Technology, 4412) or goat anti-mouse IgG H&L (Alexa Fluor 647) (Abcam, ab150115) for 1 h at 37 °C in the dark. Nuclei were stained with DAPI, and images were captured using a ZEISS LSM 900 confocal laser scanning microscope.

Flow cytometry

Fresh mouse tumors were collected, cut into small pieces, and dissociated into single-cell suspensions using the tumor dissociation kit (Miltenyi Biotec, Germany). After staining the single-cell suspensions with Human TruStain FcX (Biolegend, 422301/101320), Fixable Viability Dye (Biolegend, 423105), and surface markers, cells were fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences, 554715). After two washes, cells were incubated with the indicated antibodies for 1 h at 4 °C. Following three washes with ice-cold FACS buffer, cells were examined on a Beckman Coulter Cytoflex. Flow Jo was used for data analysis. The following antibodies were used: BV421 anti-human CD4 antibody (BD, 562424), PE anti-human IFN-γ antibody (BD, 554701), PerCP-Cy5.5 anti-human CD3 antibody (BD, 560835), PE-Cy7 anti-human CD8 antibody (BD, 557750), BV786 anti-human TNF antibody (BD, 571508), APC anti-human CD86 (BD, 560956), BV605 anti-human CD80 (BD, 563315), FITC anti-human CD68 (Biolegend, 333806), BV510 anti-mouse F4/80 (Biolegend, 123135), PE-Cy7 anti-mouse CD86 (Biolegend, 105014), PerCP anti-mouse CD11b (Biolegend, 101230), AF647 anti-mouse CD206 (BD, 568808), FITC anti-mouse CD45 (Biolegend, 103108), PE anti-mouse CD163 (BD, 557495), and BV421 anti-mouse CD80 (BD, 562611) diluted in FACS buffer (PBS containing 0.5% BSA) for 30 min on ice.

Western blot and immunoprecipitation (IP) assays

Cells were harvested and lysed using RIPA lysis buffer (Sigma, V900854) supplemented with protease inhibitor (Sigma, P8340) and phosphatase inhibitor (Sigma, P0044). The prepared samples were then loaded into SDS-PAGE with constant voltage and transferred to polyvinylidene fluoride (PVDF) membranes with a constant current. After blocking with 5% BSA for 1 h at room temperature, the membranes were incubated with anti-NR4A3 (Origen, TA804893), anti-phospho-p65 (Cell Signaling Technology, 3033), anti-p65 (Cell Signaling Technology, 8242), and anti-beta Tubulin (ABclonal, AC015) antibodies overnight at 4 °C. The membranes were incubated with HRP-conjugated secondary antibodies (1:2000) for 1 h at room temperature. The blots were visualized using ECL chemiluminescence technology (Bio-Rad).

For IP, cells were lysed in IP Lysis/Wash Buffer (Beyotime, P0013J). Dynabeads™ Protein G beads (Invitrogen, 10004D) were incubated with affinity-purified antibodies to allow antibody binding. The antibody-coated beads were then added to the protein lysate and incubated at 4 °C overnight with end-over-end rotation. The immune complexes were washed three times with IP Lysis/Wash Buffer and eluted with 2 × SDS-PAGE Buffer at 100 °C for 10 min. The immunoprecipitated samples were analyzed by immunoblot.

Data collection

RNA-sequencing (RNA-seq) data and clinical information for 1118 BC tissues and 113 normal tissues were downloaded from the Genomic Data Commons data portal (https://portal.gdc.cancer.gov/) and processed via the TCGAbiolinks package in R using TCGAWorkflow guided practices^37,38. The tumor molecular subtypes were classified using PAM50 based on previous publications³⁹. A BC single-cell RNA-sequencing (scRNA-seq) dataset of GSE176078 was retrieved from the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/). The RNA-seq data of polarized macrophage from THP-1 cells was obtained from the GEO database (GSE140235).

Single-cell RNA-sequencing data filtering and the standard process

ScRNA-seq data were acquired and transformed into a count matrix. Subsequently, scRNA-seq data analysis was conducted using the Seurat package (version 5.0.0) implemented in the R programming language (version 4.2.0). Quality control criteria were applied to retain cells for further analysis. Only cells that met the following criteria were included: 200 < nFeature < 5000, 500 < nCount < 100,000, mitochondrial gene expression accounting for less than 25% of the total expressed genes per cell, and erythroid gene expression accounting for less than 5% of the total expressed genes per cell. To mitigate variations in library size and cell-specific biases, the count matrix was normalized. Highly variable genes were identified based on their expression variance across cells. Normalization was performed after identifying the top 2000 highly variable genes. We employed the Harmony R package (version 0.1.1) to alleviate batch effects observed within the samples, as previously reported in the relevant literature⁴⁰. Further downscaling and clustering analysis were conducted using the top 30 principal components (PCs). To annotate the cell clusters, we referred to known cell markers from previous literature and the CellMarker 2.0 database (http://bio-bigdata.hrbmu.edu.cn/CellMarker/)⁴¹. The UMAP method was employed to visualize the cell clusters in a two-dimensional map. Subsequently, the proportion of different cell types in the dataset was evaluated by assessing the distribution of cell types across clusters. To visualize and explore the coexpression of NR4A3 and CD68, we employed the scCustomize R package (version 2.0.1).

Differentiation and enrichment analysis

To identify DEGs within the group of NR4A3, we employed the “pseudobulk” function in the “Libra” package (version 1.0.0). This analysis was based on the edgeR-LRT with default parameters. Reference genomes (Hallmark, c5go) were obtained from the Molecular Signatures Database (MSigDB). Gene set enrichment analysis (GSEA) was performed using the R package “clusterProfiler” R package (version 0.1.1).

Subpopulation analysis of myeloid cells

To investigate the characteristics of myeloid cells, gene signature scores were used to evaluate the strength of various cellular phenotypes or biological processes based on gene sets corresponding to each signature and gene expression data. The signature lists used to define the macrophage phenotypes were obtained from previously published studies, including classically activated M1-like, and alternatively activated M2-like macrophages (Supplementary Table 4)⁴². The R package AUCell (version 1.16.0) was used to analyze the activity of gene sets (Hallmark).

The expression analysis between gene and cell type

The GSE161529, GSE150660, GSE148673, GSE138536, GSE110686, and EMTAB8107 datasets from the Tumor Immune Single Cell Hub (TISCH) (http://tisch.comp-genomics.org/) database were used to analyze the expression of NR4A3 among TME-associated cells⁴³.

Survival and immunotherapy response analysis

Kaplan–Meier analysis of OS and DFS was obtained from the GEPIA database (http://gepia.cancer-pku.cn/)⁴⁴. The anti-PD-1 and anti-CTLA4 immunophenoscore (IPS) data were downloaded from the TCIA database (https://tcia.at/home/) to assess the responsiveness to immune checkpoint inhibitors^45,46.

Tumor microenvironment analysis

We used the R package “ESTIMATE” to perform tumor microenvironmental analysis and calculate the ESTIMATE score, immune cell score, and stroma score for each sample. The TIMER2.0 database (http://timer.cistrome.org/) was used for tumor purity analysis. The proportion of 22 types of immune infiltrating cells was calculated by the CIBERSORT algorithm (https://cibersort.stanford.edu/).

Immunological subtypes of cancer analysis

An online integrated repository portal called TISIDB (http://cis.hku.hk/TISIDB/) gathers data from TCGA, which contains extensive datasets related to human cancer⁴⁷. In the TISIDB database, correlations were examined between NR4A3 expression and immunological subtypes of cancers (C1 (wound healing), C2 (IFN-gamma dominant), C3 (inflammatory), C4 (lymphocyte depleted), C5 (immunologically quiet), C6 (TGF-b dominant))⁴⁸.

Protein structure prediction

The protein structures were initially generated using AlphaFold3. Structural optimization was subsequently performed through explicit solvent removal and hydrogen atom addition implemented in AutoDockTools-1.5.7 to prepare for molecular docking. Protein–protein docking was subsequently performed on the AlphaFold3 docking server. The resultant complex structure underwent additional refinement involving manual solvent removal and hydrogen atom optimization using AutoDockTools-1.5.7. Molecular interaction analysis was ultimately performed through PyMOL visualization to generate the protein–protein interaction network diagram.

Statistical analysis

GraphPad Prism 5 was employed in our statistical analyses and for the graph of our data. Unpaired Student’s t-test and Wilcoxon test were used to compare differences between the two groups. Pearson tests were used to analyze the correlation between two paired factors. The Kruskal–Wallis test was applied to the comparison of more than two groups. Data obtained from the Kaplan–Meier plotter presented as hazard ratio (HR) and pvalues upon the log-rank test. GSEA analysis was presented as normalized enrichment score (NES), and padjust. Except for the tumor volume in mice, which is presented as the values ± standard error (SEM), all results are represented as the values ± standard deviation (SD). The significant difference was considered when pvalues were less than 0.05.

Acknowledgements

This work was supported by grants from the National Key Technology Research and Development Program of China (2022YFC2704200 and 2022YFC2704205), the National Science Foundation of China (81902933, 82403749, and 82403616), and the Natural Science Foundation of Hubei Province (2023AFB760, 2024AFB029). We would like to thank the online database (TCGA-BRCA, GSE176078, GSE140235, GSE161529, GSE150660, GSE148673, GSE138536, GSE110686, and EMTAB8107) for the availability of the data.

Author contributions

Y.-y.Q.: Planning, performing, analyzing in vitro experiments, and writing the original draft. N.J. and W.P.: Analyzing TCGA and scRNA-seq data and correcting the article. S.-s.R. and Y.-k.L: Contributed to the analysis of IHC staining. Y.W., X.L., and P.H.: Contributed to the flow cytometry experiments. S.-y.W. and P.-f.L.: Contributed to the animal experiments. Q.-l.G.: Responsible for hypothesis formulation and supervision. Y.X.: Responsible for hypothesis formulation, experimental design, writing, and supervision.

Data availability

Publicly available datasets can be accessed by the websites provided in the methods and the data supporting this study are available from the corresponding author upon reasonable request.

Code availability

The underlying code for this study is not publicly available, but it may be made available to researchers upon reasonable request to the corresponding author.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41523-025-00785-0.

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

References

1. Siegel, RL; Miller, KD; Wagle, NS; Jemal, A. Cancer statistics, 2023. CA Cancer J. Clin.; 2023; 73, pp. 17-48. [DOI: https://dx.doi.org/10.3322/caac.21763]

2. Xiong, X et al. Breast cancer: pathogenesis and treatments. Signal Transduct. Target. Ther.; 2025; 10, 49. [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11836418][DOI: https://dx.doi.org/10.1038/s41392-024-02108-4]

3. Nitheesh, Y et al. Surface engineered nanocarriers for the management of breast cancer. Mater. Sci. Eng. C Mater. Biol. Appl.; 2021; 130, 112441.1:CAS:528:DC%2BB3MXitFanu77N [DOI: https://dx.doi.org/10.1016/j.msec.2021.112441]

4. Loibl, S; Poortmans, P; Morrow, M; Denkert, C; Curigliano, G. Breast cancer. Lancet; 2021; 397, pp. 1750-1769.1:CAS:528:DC%2BB3MXhsVegsLnK [DOI: https://dx.doi.org/10.1016/S0140-6736(20)32381-3]

5. Zhang, Y; Zhang, Z. The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications. Cell. Mol. Immunol.; 2020; 17, pp. 807-821.1:CAS:528:DC%2BB3cXht12itbbK [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7395159][DOI: https://dx.doi.org/10.1038/s41423-020-0488-6]

6. Zou, Y et al. The single-cell landscape of intratumoral heterogeneity and the immunosuppressive microenvironment in liver and brain metastases of breast cancer. Adv. Sci.; 2023; 10, [DOI: https://dx.doi.org/10.1002/advs.202203699] e2203699.

7. Keenan, TE; Tolaney, SM. Role of immunotherapy in triple-negative breast cancer. J. Natl Compr. Cancer Netw.; 2020; 18, pp. 479-489.1:CAS:528:DC%2BB3cXitV2js7fK [DOI: https://dx.doi.org/10.6004/jnccn.2020.7554]

8. Gao, W; Wang, X; Zhou, Y; Wang, X; Yu, Y. Autophagy, ferroptosis, pyroptosis, and necroptosis in tumor immunotherapy. Signal Transduct. Target. Ther.; 2022; 7, 196. [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9208265][DOI: https://dx.doi.org/10.1038/s41392-022-01046-3]

9. Niu, X; Chen, L; Li, Y; Hu, Z; He, F. Ferroptosis, necroptosis, and pyroptosis in the tumor microenvironment: perspectives for immunotherapy of SCLC. Semin. Cancer Biol.; 2022; 86, pp. 273-285.1:CAS:528:DC%2BB38XhsVGhs7zJ [DOI: https://dx.doi.org/10.1016/j.semcancer.2022.03.009]

10. Cassetta, L; Pollard, JW. Targeting macrophages: therapeutic approaches in cancer. Nat. Rev. Drug Discov.; 2018; 17, pp. 887-904.1:CAS:528:DC%2BC1cXitVSisrnO [DOI: https://dx.doi.org/10.1038/nrd.2018.169]

11. Morandi, F; Pistoia, V. Interactions between HLA-G and HLA-E in physiological and pathological conditions. Front. Immunol.; 2014; 5, 394. [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4141331][DOI: https://dx.doi.org/10.3389/fimmu.2014.00394]

12. Noy, R; Pollard, JW. Tumor-associated macrophages: from mechanisms to therapy. Immunity; 2014; 41, pp. 49-61.1:CAS:528:DC%2BC2cXhtFOrsLvL [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4137410][DOI: https://dx.doi.org/10.1016/j.immuni.2014.06.010]

13. Biswas, SK; Mantovani, A. Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nat. Immunol.; 2010; 11, pp. 889-896.1:CAS:528:DC%2BC3cXhtFyhtrrL [DOI: https://dx.doi.org/10.1038/ni.1937]

14. Salmon, H et al. Expansion and activation of CD103(+) dendritic cell progenitors at the tumor site enhances tumor responses to therapeutic PD-L1 and BRAF inhibition. Immunity; 2016; 44, pp. 924-938.1:CAS:528:DC%2BC28XmsVehsLw%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4980762][DOI: https://dx.doi.org/10.1016/j.immuni.2016.03.012]

15. Crean, D; Murphy, EP. Targeting NR4A nuclear receptors to control stromal cell inflammation, metabolism, angiogenesis, and tumorigenesis. Front. Cell Dev. Biol.; 2021; 9, [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7901948][DOI: https://dx.doi.org/10.3389/fcell.2021.589770] 589770.

16. Hibino, S et al. Inhibition of Nr4a receptors enhances antitumor immunity by breaking treg-mediated immune tolerance. Cancer Res.; 2018; 78, pp. 3027-3040.1:CAS:528:DC%2BC1cXhtVKjsb%2FK [DOI: https://dx.doi.org/10.1158/0008-5472.CAN-17-3102]

17. Deutsch, AJA et al. NR4A3 suppresses lymphomagenesis through induction of proapoptotic genes. Cancer Res.; 2017; 77, pp. 2375-2386.1:CAS:528:DC%2BC2sXmvFarsL4%3D [DOI: https://dx.doi.org/10.1158/0008-5472.CAN-16-2320]

18. Fedorova, O et al. Orphan receptor NR4A3 is a novel target of p53 that contributes to apoptosis. Oncogene; 2019; 38, pp. 2108-2122.1:CAS:528:DC%2BC1cXit1Cgtr7I [DOI: https://dx.doi.org/10.1038/s41388-018-0566-8]

19. Chen, J et al. NR4A transcription factors limit CAR T cell function in solid tumours. Nature; 2019; 567, pp. 530-534.1:CAS:528:DC%2BC1MXmsVyqur4%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6546093][DOI: https://dx.doi.org/10.1038/s41586-019-0985-x]

20. Iglesias-Escudero, M; Arias-González, N; Martínez-Cáceres, E. Regulatory cells and the effect of cancer immunotherapy. Mol. Cancer; 2023; 22, 1:CAS:528:DC%2BB3sXis1ejs7s%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9898962][DOI: https://dx.doi.org/10.1186/s12943-023-01714-0] 26.

21. Lv, W et al. A novel immune score model predicting the prognosis and immunotherapy response of breast cancer. Sci. Rep.; 2023; 13, 1:CAS:528:DC%2BB3sXosVCqurc%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10115816][DOI: https://dx.doi.org/10.1038/s41598-023-31153-2] 6403.

22. Xu, X et al. Phosphorylation of NF-κBp65 drives inflammation-mediated hepatocellular carcinogenesis and is a novel therapeutic target. J. Exp. Clin. Cancer Res.; 2021; 40, 253.1:CAS:528:DC%2BB3MXislGnsb3J [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8359590][DOI: https://dx.doi.org/10.1186/s13046-021-02062-x]

23. Wang, Y; Hu, L; Tong, X; Ye, X. Casein kinase 1γ1 inhibits the RIG-I/TLR signaling pathway through phosphorylating p65 and promoting its degradation. J. Immunol.; 2014; 192, pp. 1855-1861.1:CAS:528:DC%2BC2cXitFWms78%3D [DOI: https://dx.doi.org/10.4049/jimmunol.1302552]

24. Smith, AG et al. Melanocortin-1 receptor signaling markedly induces the expression of the NR4A nuclear receptor subgroup in melanocytic cells. J. Biol. Chem.; 2008; 283, pp. 12564-12570.1:CAS:528:DC%2BD1cXltVyqtro%3D [DOI: https://dx.doi.org/10.1074/jbc.M800480200]

25. Malewicz, M et al. Essential role for DNA-PK-mediated phosphorylation of NR4A nuclear orphan receptors in DNA double-strand break repair. Genes Dev.; 2011; 25, pp. 2031-2040.1:CAS:528:DC%2BC3MXhtlyhtr%2FF [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3197202][DOI: https://dx.doi.org/10.1101/gad.16872411]

26. Boylan, MO; Athanassiou, M; Houle, B; Wang, Y; Zarbl, H. Activation of tumor suppressor genes in nontumorigenic revertants of the HeLa cervical carcinoma cell line. Cell Growth Differ.; 1996; 7, pp. 725-735.1:CAS:528:DyaK28XjsFaqtLk%3D

27. Ke, N et al. Nuclear hormone receptor NR4A2 is involved in cell transformation and apoptosis. Cancer Res.; 2004; 64, pp. 8208-8212.1:CAS:528:DC%2BD2cXpvVGlur0%3D [DOI: https://dx.doi.org/10.1158/0008-5472.CAN-04-2134]

28. Zhou, F et al. Nuclear receptor NR4A1 promotes breast cancer invasion and metastasis by activating TGF-β signalling. Nat. Commun.; 2014; 5, [DOI: https://dx.doi.org/10.1038/ncomms4388] 3388.

29. Yousefi, H; Fong, J; Alahari, SK. NR4A family genes: a review of comprehensive prognostic and gene expression profile analysis in breast cancer. Front. Oncol.; 2022; 12, 1:CAS:528:DC%2BB3sXhtlGqsL3J [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9082356][DOI: https://dx.doi.org/10.3389/fonc.2022.777824] 777824.

30. Chio, CC et al. Neuron-derived orphan receptor 1 transduces survival signals in neuronal cells in response to hypoxia-induced apoptotic insults. J. Neurosurg.; 2016; 124, pp. 1654-1664.1:CAS:528:DC%2BC1cXht1Chsb3L [DOI: https://dx.doi.org/10.3171/2015.6.JNS1535]

31. Takahashi, H et al. GLP-1 receptor agonist exendin-4 attenuates NR4A orphan nuclear receptor NOR1 expression in vascular smooth muscle cells. J. Atheroscler. Thromb.; 2019; 26, pp. 183-197.1:CAS:528:DC%2BC1MXhtFyntrzN [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6365156][DOI: https://dx.doi.org/10.5551/jat.43414]

32. He, YW. Orphan nuclear receptors in T lymphocyte development. J. Leukoc. Biol.; 2002; 72, pp. 440-446.1:CAS:528:DC%2BD38XntlOnurY%3D [DOI: https://dx.doi.org/10.1189/jlb.72.3.440]

33. Nagaoka, M et al. The orphan nuclear receptor NR4A3 is involved in the function of dendritic cells. J. Immunol.; 2017; 199, pp. 2958-2967.1:CAS:528:DC%2BC2sXhs1amtbbM [DOI: https://dx.doi.org/10.4049/jimmunol.1601911]

34. Ma, Y; Chen, SS; Jiang, F; Ma, RY; Wang, HL. Bioinformatic analysis and validation of microRNA-508-3p as a protective predictor by targeting NR4A3/MEK axis in pulmonary arterial hypertension. J. Cell. Mol. Med.; 2021; 25, pp. 5202-5219.1:CAS:528:DC%2BB3MXhsFaqu7jI [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8178270][DOI: https://dx.doi.org/10.1111/jcmm.16523]

35. De Paoli, F et al. The neuron-derived orphan receptor 1 (NOR1) is induced upon human alternative macrophage polarization and stimulates the expression of markers of the M2 phenotype. Atherosclerosis; 2015; 241, pp. 18-26. [DOI: https://dx.doi.org/10.1016/j.atherosclerosis.2015.04.798]

36. Ma, C et al. The pro-inflammatory effect of NR4A3 in osteoarthritis. J. Cell. Mol. Med.; 2020; 24, pp. 930-940.1:CAS:528:DC%2BB3cXpsFeqtQ%3D%3D [DOI: https://dx.doi.org/10.1111/jcmm.14804]

37. Comprehensive molecular portraits of human breast tumours. Nature490, 61–70. (2012).

38. Mounir, M et al. New functionalities in the TCGAbiolinks package for the study and integration of cancer data from GDC and GTEx. PLoS Comput. Biol.; 2019; 15, e1006701. [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6420023][DOI: https://dx.doi.org/10.1371/journal.pcbi.1006701]

39. Thennavan, A. et al. Molecular analysis of TCGA breast cancer histologic types. Cell Genom. 1, 100067 (2021).

40. Korsunsky, I et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods; 2019; 16, pp. 1289-1296.1:CAS:528:DC%2BC1MXitFOqsr7N [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6884693][DOI: https://dx.doi.org/10.1038/s41592-019-0619-0]

41. Hu, C et al. CellMarker 2.0: an updated database of manually curated cell markers in human/mouse and web tools based on scRNA-seq data. Nucleic Acids Res.; 2023; 51, pp. D870-d6.1:CAS:528:DC%2BB3sXhtlWrsL7J [DOI: https://dx.doi.org/10.1093/nar/gkac947]

42. Yang, Q et al. Single-cell RNA sequencing reveals the heterogeneity of tumor-associated macrophage in non-small cell lung cancer and differences between sexes. Front. Immunol.; 2021; 12, 1:CAS:528:DC%2BB38XhtlGnur3F [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8602907][DOI: https://dx.doi.org/10.3389/fimmu.2021.756722] 756722.

43. Sun, D et al. TISCH: a comprehensive web resource enabling interactive single-cell transcriptome visualization of tumor microenvironment. Nucleic Acids Res.; 2021; 49, pp. D1420-d30.1:CAS:528:DC%2BB3MXisFOnsLjF [DOI: https://dx.doi.org/10.1093/nar/gkaa1020]

44. Tang, Z et al. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res.; 2017; 45, pp. W98-w102.1:CAS:528:DC%2BC1cXhs1aitr8%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5570223][DOI: https://dx.doi.org/10.1093/nar/gkx247]

45. Van Allen, EM et al. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science; 2015; 350, pp. 207-211. [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5054517][DOI: https://dx.doi.org/10.1126/science.aad0095]

46. Hugo, W et al. Genomic and transcriptomic features of response to anti-PD-1 therapy in metastatic melanoma. Cell; 2016; 165, pp. 35-44.1:CAS:528:DC%2BC28XksVKrsbw%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4808437][DOI: https://dx.doi.org/10.1016/j.cell.2016.02.065]

47. Ru, B et al. TISIDB: an integrated repository portal for tumor-immune system interactions. Bioinformatics; 2019; 35, pp. 4200-4202.1:CAS:528:DC%2BB3cXhtVektL7M [DOI: https://dx.doi.org/10.1093/bioinformatics/btz210]

48. Thorsson, V et al. The immune landscape of cancer. Immunity; 2018; 48, pp. 812-30.e14.1:CAS:528:DC%2BC1cXntFSjt7o%3D [PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5982584][DOI: https://dx.doi.org/10.1016/j.immuni.2018.03.023]