Tumour‐Derived Extracellular Vesicles Reprogramme Tumour‐Associated Macrophages Into Immunosuppressive Phenotype via NOD1 Signalling in Clear Cell Renal Cell Carcinoma

ABSTRACT

Clear cell renal cell carcinoma (ccRCC)‐derived extracellular vesicles (EVs) play a pivotal role in myeloid cell differentiation toward immunosuppressive phenotypes, yet the underlying mechanisms remain poorly understood. Through integrated single‐nucleus RNA sequencing, EV proteomics, and phenotype validation, we identified NOD1, a pattern recognition receptor (PRR), as a critical mediator of EV‐driven macrophage differentiation. Clinically, NOD1 overexpression in ccRCC tissues correlated with increased immunosuppressive macrophage infiltration, immune checkpoint inhibitors (ICIs) therapy resistance, and poor prognosis. Mechanistically, ccRCC‐EVs enriched with SCRIB, a Rho GTPase activator, triggered sustained NOD1 activation in macrophages via Rho GTPase signalling. Crucially, NOD1 signalling converged on interferon regulatory factor 4 (IRF4), whose expression might be dynamically regulated through NF‐κB activation and MITF suppression. Furthermore, inhibition of NOD1 promotes the therapeutic effect of ICIs in vivo. Our findings unveil a novel EV‐mediated immune evasion mechanism in ccRCC and propose the SCRIB‐NOD1‐IRF4 axis as a therapeutic target to restore anti‐tumour immunity.

1. Introduction

Clear cell renal cell carcinoma (ccRCC) is the most prevalent histological subtype of renal malignancies (Young et al. 2024). Approximately 20% of ccRCC patients are diagnosed at advanced stages, and one‐third of those undergoing localized tumour resection eventually experience recurrence (Rose and Kim, 2024). Tumour‐infiltrating myeloid cells, including macrophages, have been identified as critical contributors to ccRCC progression, invasion, and resistance to immune checkpoint inhibitor (ICI) therapies (Mantovani et al. 2017; Yang et al. 2018). Consequently, targeting these immune cells to restore anti‐tumour immunity emerges as a promising therapeutic strategy for ccRCC.

As key components of the tumour microenvironment (TME), the functional states of myeloid cells are dynamically regulated by TME‐derived signals. Within ccRCC TME, macrophages often exhibit functional properties associated with immunosuppression, whereas populations with pro‐inflammatory characteristics are relatively diminished (Jin et al. 2023; Shen et al., 2021). This shift toward an immunosuppressive phenotype correlates with poor clinical outcomes in ccRCC patients. However, the mechanisms by which ccRCC TME facilitates this functional reprogramming remain to be fully elucidated. Specifically, the pathways by which tumour cells transmit immunosuppressive cues to macrophages to drive their phenotypic transition remain a subject of intense investigation.

Extracellular vesicles (EVs), including exosomes and microvesicles, serve as essential mediators of intercellular communication between tumour cells and other cellular components within TME (Cocozza et al., 2020). These vesicles carry diverse bioactive molecules (proteins, lipids, mRNAs and miRNAs) that can be transferred to neighbouring immune cells, thereby influencing tumour immune surveillance and immune evasion (Marar et al. 2021). Previous studies have demonstrated that ccRCC‐derived EVs can modulate the phenotype of macrophages and confer systemic treatment resistance (Qu et al. 2016; Lu et al. 2023). Nevertheless, the molecular mechanisms through which EVs mediate macrophage functional reprogramming in ccRCC require further investigation.

In this study, we identified NOD1, a host intracellular pattern‐recognition receptor (PRR), as a critical regulator of macrophage phenotypic programming. Mechanistically, we discovered that ccRCC‐derived EVs trigger the NOD1 signalling pathway via the protein SCRIB. Within macrophages, NOD1 activation subsequently drives a functional shift toward an immunosuppressive state by modulating IRF4 activity. These findings not only elucidate the process and mechanism underlying NOD1 activation in host immune cells by ccRCC‐derived EVs, but also reveal the critical role of the SCRIB‐NOD1‐IRF4 regulatory axis in macrophage reprogramming in ccRCC, offering a potential therapeutic target for ccRCC treatment.

2. Materials and Methods

2.1. Cell Lines and Cell Culture

The human clear cell renal cell carcinoma cell lines 786O and 769P, human normal renal tubular cell line HK2, and mouse ccRCC cell line Renca were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). Two primary human ccRCC cell lines (R0312 and R0221) were established from fresh tumour tissues obtained from donors. THP‐1 cells were obtained from Cell Bank of Typical Culture Preservation Committee of Chinese Academy of Sciences. 786O, 769P, HK2, R0312 and R0221 cells were cultured in RPMI 1640 (Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Thermo Fisher). THP‐1 cells were cultured in RPMI 1640 (Gibco) supplemented with 10% fetal bovine serum (Gibco), 1% penicillin/streptomycin (Thermo Fisher) and 0.05 mM 2‐mercaptoethanol (Thermo Fisher). To induce differentiation into resting macrophages, THP‐1 cells were treated with 50 ng/mL phorbol 12‐myristate 13‐acetate (PMA; Selleck) followed by a 48 h incubation period. Human peripheral blood monocytes (hPBM) were isolated and purified from PBMCs using Dynabeads FlowComp Human CD14 Kit (11367D, Thermo Fisher), and were cultured in RPMI 1640 (Gibco) containing 10% fetal bovine serum (Gibco) and 50 ng/mL recombinant human macrophage colony‐stimulating factor (M‐CSF; Peprotech) for 7 days at 37°C for differentiation into resting macrophages.

THP‐1 cells were transduced with lentiCRISPR lentivirus targeting NOD1 (NOD1‐KO), NOD1‐overexpressing lentiviral vector (NOD1‐OE) or control lentivirus. hPBM cells were transduced with shRNA lentivirus targeting NOD1 (NOD1‐KD) or control lentivirus. For rescue experiments, THP‐1 cells and hPBMs were transduced with a SCRIB‐overexpressing lentiviral vector or corresponding control vector. 786O, 769P, Renca cells were transduced with lentiCRISPR lentivirus targeting SCRIB (SCRIB‐KO), overexpressing lentiviral vector (SCRIB‐OE) or control lentivirus. Cells infected with virus were selected with 2 µg/mL puromycin (Selleck). C12‐iE‐DAP (HY‐P5524) and Nodinitib‐1 (ML130; HY‐18639) were purchased from MCE.

2.2. Patient Specimens

Written informed consent was obtained from all the patients. 196 patient specimens were obtained from patients who underwent radical or partial nephrectomy at Xinhua Hospital from 2016 to 2018. All tumour specimens were pathologically confirmed as ccRCC. After resection, the samples were immediately frozen in liquid nitrogen and stored at −80°C for further studies. Specimen collection for research purposes was approved by Clinical Research Ethics Board of Xinhua Hospital.

2.3. Processing of Single‐Nucleus RNA Sequencing (snRNA‐seq) Data

snRNA‐seq data were processed using R (v4.3.0) and Seurat (R package, v4.3.0.1). Low‐quality cells were filtered out by excluding those with fewer than 200 genes, over 10,000 genes, UMI counts exceeding 10,000, or mitochondrial read proportions above 10%. Additionally, only genes detected in at least three cells were retained. For each sample, DoubletFinder (R package, v2.0.4) was used to remove doublets. 2000 highly variable genes (HVGs) and top 30 principal components were used for clustering. Batch effect correction was performed using ‘Runharmony’ function and subsequently achieve UMAP plot for cell cluster identification. The same analytical pipeline—including variable gene identification, dimensionality reduction and clustering—was applied to define subclusters.

2.4. Cell Type Annotation and Cluster Marker Identification

Cluster‐specific markers were identified using the ‘FindAllMarkers’ algorithm, with subsequent annotation performed according to established canonical markers from published single‐cell and single‐nucleus RNA sequencing datasets of renal tumours.

2.5. Gene set Enrichment Analysis

Pathway activity scores across cell clusters or samples were computed using GSVA (R package, version 1.48.2), with visualization via heatmaps incorporating gene sets from the Molecular Signatures Database (MSigDB). Differential pathway activities were subsequently analysed using the limma package in R.

2.6. Immunohistochemistry and Multiplex Immunofluorescence

Immunohistochemistry (IHC) and multiplex immunofluorescence (mIF) analyses were performed on paraffin‐embedded sections of ccRCC tissues. IHC analysis was performed using standard protocols. Tissue sections were initially deparaffinized in xylene and rehydrated through a descending ethanol gradient. Endogenous peroxidase was inactivated with 0.3% H₂O₂ treatment (20 min), followed by blocking with 10% normal goat serum. Primary antibodies were applied and incubated at 4°C overnight, followed by HRP‐conjugated secondary antibody incubation and DAB development. Nuclear counterstaining was completed using hematoxylin.

mIF staining was conducted using the PANO 7‐plex IHC kit (Panovue, 0004100100). The procedure included iterative rounds of primary antibody incubation, HRP‐conjugated secondary antibody application, and TSA‐based signal amplification, with microwave‐mediated antibody stripping between cycles. DAPI was used for final nuclear staining. Imaging was performed using an OLYMPUS FLUOVIEW VS200 system. Cellular segmentation and subsequent quantification were performed using PanoScore software (Panovue, Beijing, China). Nuclei were identified from DAPI signals, and cytoplasmic boundaries were delineated by applying a 5 µm radial expansion. To exclude artifacts and overlapping nuclei, cell areas were constrained to a range of 10–400 µm². Fluorescence positivity thresholds were established independently for each marker and tissue core. These thresholds were determined based on background signal intensity and staining quality. To ensure accurate detection across all tissue microarrays (TMAs), the automatically assigned thresholds were manually validated and adjusted as necessary. In mIF, CD86⁺ or CD80⁺ cells were used as markers for pro‐inflammatory macrophages and CD163⁺, CD206⁺, or IL‐10⁺ cells were used as markers for immunosuppressive macrophages. Tumours with a NOD1 positivity rate exceeding 50% in macrophages were classified into the NOD1‐high group, whereas those with a rate below 50% were assigned to the NOD1‐low group. The antibodies were provided in the Supplemental Materials.

2.7. Isolation of EVs

Cells were cultured in RPMI 1640 (Gibco) with EV‐depleted FBS (OriCell) for 72 h, reaching 80%–90% confluency. EVs were isolated from the conditioned medium through sequential centrifugation. Briefly, the collected supernatant was first spun at 2000 × g for 30 min and then at 10,000 × g for 45 min to eliminate larger vesicles. The clarified medium was passed through a 0.45‐µm filter and subjected to ultracentrifugation at 100,000 × g for 70 min (4°C). After discarding the supernatant, the pellet was washed with PBS and centrifuged again under the same conditions. The final EV pellet was resuspended in PBS.

EV protein content was quantified using a BCA assay (Pierce), and aliquots (20 µg per vial) were stored at −80°C for downstream applications. For TEM and NTA, EVs were diluted in PBS (1:100) prior to analysis.

2.8. Western Blotting Assay

Total protein was extracted from the cultured cells or specimens using RIPA buffer according to manufacturer's protocol, and resolved by SDS‐polyacrylamide gels, and then transferred to PVDF membranes. The antibodies were provided in the Supplemental Materials. The immunoreactive bands were analysed by chemiluminescence.

2.9. RNA Extraction and Real‐Time Quantitative PCR (RT‐qPCR)

Total RNA was isolated using TRIzol reagent (Invitrogen), and 2 µg of total RNA was subjected to reverse transcription using Superscript III transcriptase (Invitrogen). RT‐qPCR was conducted using a Bio‐Rad CFX96 system with SYBR green to determine the mRNA expression level of a gene of interest. Expression levels were normalized to the ACTB level using the 2−∆∆Ct methods. The sequences of the primers were provided in Supplemental Materials.

2.10. Data Independent Acquisition (DIA) Quantitative Proteomics Analysis of EVs

Proteins were extracted from EV samples using SDT lysis buffer (4% SDS, 100 mM DTT, 100 mM Tris‐HCl pH 8.0) with boiling and ultrasonication, followed by centrifugation and BCA quantification. Digestion used FASP method: proteins were reduced with DTT, alkylated with IAA in UA buffer, and digested overnight with trypsin (1:50 ratio, 37°C). LC‐MS/MS analysis was conducted on an Orbitrap Astral mass spectrometer coupled to a Vanquish Neo UHPLC system with a 50‐cm µPAC Neo column. Peptides were separated over an 8 min gradient (Buffer A: 0.1% FA in water; Buffer B: 0.1% FA in 80% ACN) at 1.25 µL/min. DIA parameters included: MS1 resolution 240,000 (380–980 m/z); DIA scans 150–2000 m/z with 2 m/z isolation windows. Data were processed in DIA‐NN against UniProtKB/Swiss‐Prot human database (trypsin digestion, 1 missed cleavage, carbamidomethylation fixed, oxidation/acetylation variable, 1% FDR).

2.11. Regulon Activity Analysis

PySCENIC (v0.12.1) was used to build the gene regulatory network (GRN) based on raw expression data combined with cis‐Target human motif database (V10). Enriched motifs were identified with default parameters of pySCENIC.

2.12. RNA‐seq Data Analysis

Total RNA was isolated from tissue samples using TRIzol reagent (Invitrogen) following the manufacturer's protocol. Subsequent steps, including RNA purification, reverse transcription, library preparation, and sequencing, were conducted at Hangzhou Cosmos Wisdom Biotechnology Co., Ltd. (Hangzhou, China) in accordance with Illumina's (San Diego, CA) guidelines. Paired‐end RNA sequencing libraries were processed on the NovaSeq 6000 platform. Raw paired‐end reads underwent trimming and quality assessment using fastp with default settings. Clean reads were then aligned to the reference genome in an orientation‐specific manner using the HISAT2 tool. Finally, the mapped reads for each sample were assembled using StringTie with a reference‐guided approach.

2.13. Animals Studies

All animal procedures were performed in compliance with the guidelines established by the Shanghai Medical Experimental Animal Care Commission and were ethically approved by the Institutional Animal Care and Use Committee of Xinhua Hospital, affiliated to Shanghai Jiao Tong University (XHEC‐F‐2025‐043). Male BALB/c mice (6–8 weeks old), purchased from SLAC Laboratory Animal Co., Ltd (Shanghai, China), were maintained under specific pathogen‐free (SPF) conditions in the university's animal facility throughout the experimental period.

For subcutaneous murine tumour models, the SCRIB‐KO and wild‐type Renca tumour cells (1 × 10⁶ cells/mouse) were injected into the right flank of each mouse. Tumour volumes were measured every 3 days using the formula length × width² × 0.5 (mm³). Tumours were harvested and then used for subsequent assay. For anti‐PD‐1 treatment groups, anti‐mouse PD‐1 (Bioxcell) 100ug was injected on day 7 for every 3 days, and Nodinitib‐1 (MCE) 20ug for every 6 days, until day 34. Tumour sizes were measured every 5 days. Tumours were harvested on day 35 and then used for subsequent assay.

2.14. Statistical Analysis

All data were analysed by GraphPad Prism V.9.0 and presented as the mean ± SD. Statistical tests for data analysis included Student's t test, Spearman correlation analysis. Unpaired two‐tailed Student's t‐tests were used to compare data of two sets. Spearman correlation analysis was employed to assess the strength and direction of the monotonic relationship between two variables. P values are indicated with *p < 0.05, **p < 0.01, and ***p < 0.001 on graphs. All in vitro experiments were repeated with at least three replicates.

3. Results

3.1. Integrated Transcriptomic Analysis Identifies NOD1 as a Potential Mediator of Macrophage Functional States in ccRCC

To identify potential therapeutic targets that modulate macrophage functional states, we conducted a comprehensive correlation analysis using the TCGA‐KIRC dataset. We specifically examined genes significantly associated with either pro‐inflammatory or immunosuppressive macrophage signatures. Notably, NOD1 expression exhibited a marked positive correlation with immunosuppressive macrophage markers (Figure 1A, 1B). Furthermore, high NOD1 expression in tumour tissue was associated with poor prognosis in ccRCC patients (Figure 1C).

FIGURE 1.

Integrated transcriptomic analysis identifies NOD1 as a potential mediator of macrophage functional states in ccRCC. (A and B) Correlation analysis between NOD1 expression and signatures of pro‐inflammatory macrophages (A) and immunosuppressive macrophages signatures (B) in TCGA‐KIRC database. (C) Kaplan–Meier survival curves for NOD1 in the TCGA‐KIRC cohort, grouped by optimal cutoff value. (D) The cell type clusters of myeloid cells in snRNA‐seq. (E) Proportion of 7 groups of myeloid cells in NOD1‐high and NOD1‐low ccRCC samples. (F) Ratio of Macro_PI to Macro_II clusters in NOD1‐high and NOD1‐low ccRCC samples. (G) Differentially expressed genes of macrophages between 2 groups, colored by macrophage clusters. (H) GSVA analysis of myeloid clusters between 2 groups. (I) GSVA analysis of RCC cells between 2 groups. Data were presented as mean ± SD.

NOD1 has been reported to exhibit dual pro‐inflammatory and anti‐inflammatory potential (Maisonneuve et al. 2021; Wei et al., 2022). Maisonneuve et al. demonstrated that in peritonitis models, prolonged NOD1 pathway activation shifts macrophage function from pro‐inflammatory to anti‐inflammatory to reduce excessive inflammation (Maisonneuve et al. 2021). To validate the association between NOD1 and immunosuppressive macrophages infiltration in ccRCC, we analysed published single‐nucleus RNA sequencing (snRNA‐seq) data from 40 ccRCC and 14 normal kidney samples (Terekhanova et al. 2023; Wu et al. 2023; Muto et al. 2021; Lake et al. 2023). Among myeloid cells (CD14, CD68), we identified and annotated seven distinct clusters: classical monocytes (CD14_Mono; CD14, LYZ), nonclassical monocytes (CD16_Mono; FCGR3A, FCGR3B), macrophages expressing pro‐inflammatory markers (Macro_PI; CD86, IL1B, TNF), macrophages with inhibitory‐inflammatory features (Macro_IIa and Macro_IIb; CD163, MRC1, ARG1), GPNMB+ macrophages (Macro_GPNMB; GPNMB, APOC1, CTSD), and dendritic cells (DC; CD1C, FCER1A) (Figure 1D; Figure S2).

Based on NOD1 expression levels across all myeloid clusters, ccRCC samples were classified into NOD1‐high and NOD1‐low groups. Cell proportion analysis revealed significantly reduced ratio of PI/II macrophages in the NOD1‐high group (Figure 1E and 1F), suggesting an enhanced shift toward an immunosuppressive functional state. We further analysed the functional differences between the two groups. Differentially expressed gene analysis demonstrated marked alterations in immune regulatory genes and NF‐κB pathway‐related genes across all macrophage clusters between two groups (Figure 1G). Gene Set Variation Analysis (GSVA) further revealed marked change in NF‐κB pathway, which was critical for macrophage functional regulation, in myeloid subsets between NOD1‐high and NOD1‐low groups. Moreover, in Macro_PI cluster, pro‐inflammatory pathways (e.g. IL‐1/IL‐6 cytokine production) were downregulated in parallel with NOD1 pathway activation, whereas anti‐inflammatory pathways (e.g., IL‐10 production) were upregulated in Macro_IIa/IIb clusters (Figure 1H). These findings collectively indicate that NOD1 pathway activation correlates with an immunosuppressive shift in the tumour macrophage compartment.

To explore potential mechanisms by which TME mediates the activation of NOD1 in macrophages, we performed GSVA on RCC cell cluster. Intriguingly, Rho GTPase activity pathway was remarkably enriched in RCC cells from NOD1‐high group samples (Figure 1I). Given that NOD1 is known to sense Rho GTPase activation states (Keestra et al. 2013), with hyperactivated Rho GTPases serving as critical NOD1 stimuli, this suggests that NOD1 activation in macrophages may be regulated by RCC cells via Rho GTPase signalling. However, whether the Rho GTPase pathway directly regulates macrophage phenotype and how its signals are transduced across cells remain to be further investigated. Additionally, RCC cells from NOD1‐high group samples exhibited enhanced blood vessel remodelling and glycolytic fermentation signatures, implying these tumours may possess greater malignant potential and TKI resistance.

3.2. NOD1 Upregulation in ccRCC Tissues Correlates With Immunosuppressive Macrophage Infiltration and Poor Clinical Prognosis

To validate our findings in clinical cohorts, we performed immunohistochemical (IHC) staining on 196 ccRCC specimens from the Xinhua Hospital cohort. NOD1 expression was significantly elevated in tumour tissues compared to adjacent normal tissues and was predominantly localized in immune cells rather than tumour or vascular endothelial cells (Figure 2A). Kaplan–Meier analysis demonstrated that high NOD1 expression correlated with reduced progression‐free survival (PFS), consistent with TCGA RNA‐seq data and further confirming its association with poor prognosis (Figure 2B).

FIGURE 2.

Elevated NOD1 in ccRCC correlates with immunosuppressive macrophages infiltration and poor prognosis. (A) IHC staining of NOD1 in ccRCC adjacent para‐tumour tissues (left), NOD1‐low tumour (middle) and NOD1‐high tumour (right) (scale bar = 50 µm). (B) Kaplan–Meier curves of survival probability in Xinhua Hospital cohort. Patients were divided by IHC score. (C) Quantitative analysis of NOD1 score in patients responding to ICI therapy (left) and patients non‐responding to ICI therapy (right). (D) H&E staining of patient tissue sections confirms pathological type as ccRCC. (E) Representative mIF images of adjacent kidney tissue (left), NOD1‐low tumour (middle) and NOD1‐high tumour (right) from patient tissue sections. (row1: NOD1 (purple), row2: CD86 (red), row3 CD163 (green), row4: merged image of NOD1, PANCK (light blue), CD86, CD163 and DAPI (blue)) (scale bar = 50 µm), and (F) quantitative analysis of CD86⁺ and CD80⁺ macrophages (indicative of a pro‑inflammatory phenotype) and CD163⁺, CD206⁺ and IL10⁺ macrophages (indicative of an immunosuppressive phenotype). Data were presented as mean ± SD. A significant difference between the groups, *p < 0.05.

To investigate the relationship between NOD1 expression and ICI resistance, we scored NOD1 IHC intensity (0–5) in samples from patients who received postoperative ICI therapy. Significantly higher NOD1 expression was observed in non‐response group compared to response group (Figure 2C), suggesting that NOD1‐high ccRCC tumours are more prone to ICI therapy resistance. These results highlight NOD1 as a potential biomarker for predicting prognosis and therapeutic efficacy in ccRCC.

To further validate the association between NOD1 and macrophage functional states, we conducted multiplex immunofluorescence (mIF) staining for NOD1 and common macrophage markers on patient tissue sections. All samples were pathologically confirmed as ccRCC (Figure 2D). CD86⁺ cells were used to indicate a pro‐inflammatory phenotype, while CD163⁺ cells were used as a marker for an immunosuppressive phenotype. NOD1 expression was predominantly observed in macrophages and showed strong co‐localization with CD163⁺ macrophages (Figure 2E; Figure. S5). Stratifying samples into NOD1‐high and NOD1‐low subgroups based on fluorescence intensity revealed increased CD163⁺ macrophages infiltration and decreased CD86⁺ macrophages infiltration in the NOD1‐high group (Figure 2F), indicating a shift toward an immunosuppressive phenotype. We further incorporated established immunophenotypic markers to validate these findings: CD80 expression (associated with a pro‐inflammatory phenotype), and CD206 and IL‐10 (both indicative of an immunosuppressive phenotype) (Figure 2F, S6). Collectively, these results demonstrate that elevated NOD1 expression in tumour tissues is associated with an immunosuppressive shift in the tumour immune landscape, ICI resistance and adverse clinical outcomes in ccRCC patients.

3.3. CcRCC‐EVs Activate NOD1 Signalling to Promote an Immunosuppressive Phenotype in Macrophages

Our prior analyses suggested that NOD1 activation in macrophages might originate from tumour cell‐derived molecules. Given the pivotal role of EVs in intercellular communication, we hypothesized that ccRCC‐derived EVs (ccRCC‐EVs) mediate NOD1 activation in macrophages. To test the hypothesis, EVs were isolated via ultracentrifugation from two ccRCC cell lines (786O and 769P), two primary ccRCC cell lines (R0312 and R0221), and normal renal tubular cells (HK2) (Figure S4A). Transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) confirmed the presence of EVs with characteristic lipid bilayers and diameters averaging 70 nm (Figure S4B, S4C). Western blotting validated EV identity through detection of positive markers (CD9, Flotillin‐1, CD63, ALIX, TSG101) and absence of the negative marker calnexin (Figure S4D). After labelling the EVs with DID, they were added to two distinct macrophage model systems: PMA‐differentiated THP‐1 cells and M‐CSF‐differentiated hPBM. Confocal microscopy showed that macrophages could uptake EVs derived from ccRCC cells (Figure 3A).

FIGURE 3.

ccRCC‐EVs drive sustained activation of NOD1 signalling and promote an immunosuppressive phenotype in macrophages. (A) Confocal microscope shows that macrophages could uptake EV (DID‐labelled) derived from ccRCC cells (scale bar  =  10 µm). (B) Western blot detecting the protein expression levels of p‐RIP2, RIP2, p‐P65 and P65 in THP‐1 cell treated with EVs derived from ccRCC cell line (EV‐786O, EV‐769P, EV‐R0312, EV‐R0221) or normal renal tubular cell line (EV‐HK2). C12‐iE‐DAP represents the positive control. NC represents negative control obtained from RPMI 1640 with 5% “EV‐depleted FBS” with differential ultracentrifugation and 0.45‐µm filtration. (C) Western blot detecting levels of pro‐IL1β, IL6, iNOS IL10, ARG1 and CD163. (D) RT‐qPCR detecting relative levels of pro‐IL1β, IL6, iNOS, IL10, ARG1 and CD163 after EV treatment for 2 h and 24 h. (E) Western blot detecting levels of p‐RIP2, RIP2, p‐P65 and P65 in ML130‐treated and untreated groups. ML130 represents NOD1 inhibitor. (F) Western blot detecting levels of p‐RIP2, RIP2, p‐P65, P65, pro‐IL1β, IL6, iNOS, IL10, ARG1 and CD163 in NOD1‐KO and WT THP‐1 macrophage model, as well as NOD1‐KD and Ctrl hPBM macrophage model. Data were presented as mean ± SD. A significant difference between the groups, **p < 0.01 and ***p < 0.001.

To assess NOD1 pathway activation in macrophages, PMA‐differentiated THP‐1 cells were treated with EVs from ccRCC sources (EV‐786O, EV‐769P, EV‐R0312, EV‐R0221) or normal renal tubular cells (EV‐HK2) for 24 h. C12‐iE‐DAP, a NOD1 agonist, served as the positive control, while the supernatant from RPMI 1640 medium containing 5% EV‐depleted FBS after ultracentrifugation was used as the negative control (NC). Compared to EV‐HK2 or NC, ccRCC‐EVs significantly increased the levels of phosphorylated RIP2 (p‐RIP2) and P65 (p‐P65), mirroring the effect of NOD1 agonist treatment, indicating enhanced NOD1 signalling (Figure 3B). Functional profiling revealed significant upregulation of IL‐10, ARG1, and CD163—markers commonly associated with an immunosuppressive phenotype—without altering the expression of pro‐IL‐1β, IL‐6, and iNOS, molecules typically linked to pro‐inflammatory responses. This suggests ccRCC‐EVs preferentially drive a shift toward an immunosuppressive function (Figure 3C).

Notably, RT‐qPCR analysis uncovered biphasic regulation over time: transient pro‐IL‐1β, IL‐6, and iNOS elevation at 2h post‐EV stimulation shifted to sustained IL‐10, ARG1, and CD163 dominance at 24h (Figure 3C), consistent with the report by Maisonneuve et al. in peritonitis models (Maisonneuve et al. 2021). This dynamic response may explain how chronic EV exposure in ccRCC TME favours immunosuppressive reprogramming.

To confirm that RIP2 phosphorylation and downstream NF‐κB activation depend on NOD1, THP‐1 cells were pretreated with the selective NOD1 antagonist ML130 prior to ccRCC‐EV stimulation. NOD1 inhibition effectively blocked the ccRCC‐EV‐induced increases in p‐RIP2 and p‐P65 levels (Figure 3E). For further validation, we generated NOD1‐knockout (NOD1‐KO) THP‐1 cells. Compared to wild‐type (WT) controls, NOD1‐KO THP‐1 cells exhibited significantly reduced levels of NOD1 pathway activation markers (p‐RIP2, p‐P65), IL‐10, ARG1 and CD163 after ccRCC‐EV treatment (Figure 3F). hPBM with NOD1 knockdown were further used to verify these findings (Figure 3F). Together, these data demonstrate that ccRCC‐EVs specifically activate the NOD1 signalling pathway, and sustained NOD1 pathway activation drives macrophages toward an immunosuppressive phenotype.

3.4. CcRCC‐EVs Activate NOD1 via SCRIB‐Mediated Rho GTPase Signalling

Next, we tried to decipher how ccRCC‐EVs activate NOD1 in macrophages. Given the established role of NOD1 in recognizing pathogen proteins (Caruso et al. 2014; Trindade and Chen, 2020), we performed liquid chromatography‐mass spectrometry (LC‐MS)‐based proteomic profiling of EVs isolated from human ccRCC cells (EV‐786O) and normal renal tubular cells (EV‐HK2). Significant differences were observed in protein composition patterns between EV‐786O and EV‐HK2 (Figure 4A). Through GSVA pathway analysis of proteins contained in EVs, we systematically screened known mechanisms capable of activating NOD1, including Rho GTPase, calcium ion import and endoplasmic reticulum (ER) stress pathway. Our analysis revealed that the Rho GTPase cycle and Rho protein signal transduction pathways exhibited the most significant enrichment (Figure 4B). Recent studies have demonstrated that NOD1 not only recognizes pathogen‐derived signals but can also be activated by small Rho GTPases (Keestra‐Gounder and Tsolis, 2017; Chen et al. 2023). Therefore, we hypothesized that EV‐transmitted Rho GTPase signalling activates NOD1 in macrophages.

FIGURE 4.

SCRIB‐mediated Rho GTPase signalling as a potential route for ccRCC‐EV‐induced NOD1 activation. (A) PCA plot of proteomic samples of EVs derived from 786O and HK2 cells. (B) GSVA analysis of NOD1 upstream signals. (C) Differentially expressed Rho signal proteins in EVs. (D) Dotplot of small GTPase superfamily protein in EVs. (E) IHC staining of SCRIB in ccRCC adjacent para‐tumor tissues (left), SCRIB‐low tumour (middle) and SCRIB‐high tumour (right) (scale bar = 50 µm). (F) Kaplan–Meier curves of SCRIB in Xinhua Hospital cohort and TCGA‐KIRC cohort (grouped by optimal cutoff value). (G and H) Correlation analysis between SCRIB expression in RCC cell and Macro_PI proportion (G) or Macro_II proportion (H) in snRNA‐seq. (I) Western blot detecting expression levels of SCRIB and CDC42 in EVs.

To identify the key proteins that cause NOD1 activation in EVs, we systematically analysed differentially expressed Rho signal proteins in EVs based on two criteria: (1) significant upregulation in EV‐786O (fold change > 1.5, p < 0.05); (2) clinical correlation with poor prognosis. SCRIB, a Rho GTPase activator, showed the highest differential expression (Figure 4C). We also profiled all detected small GTPase superfamily proteins in EVs and observed no remarkable changes of expression levels between two groups (Figure 4D). These results suggest that NOD1 may be activated by SCRIB‐mediated Rho signalling.

Clinical validation in the Xinhua cohort demonstrated elevated SCRIB expression in tumour tissues versus adjacent normal tissues (Figure 4E), with high SCRIB levels correlating with poor survival, consistent with TCGA transcriptomic data (Figure 4F). We further investigated the association between SCRIB expression in tumour cells and macrophage composition in the TME. snRNA‐seq analysis revealed a positive correlation between SCRIB expression in RCC cells and the abundance of macrophages with immunosuppressive features (Macro_II subset) (Figure 4G and 4H), aligning with the observed NOD1‐mediated functional shift in macrophages.

To validate the proteomics findings, we performed western blot to confirm SCRIB enrichment in all four ccRCC‐EVs compared to EV‐HK2 (Figure 4I). Notably, we found that CDC42—a Rho GTPase previously reported to activate NOD1 in colorectal cancer (Wei et al., 2022)—did not show significant expression differences between ccRCC‐EVs and normal renal tubular cell EVs, suggesting a ccRCC‐specific mechanism. These findings nominate SCRIB as a key mediator of NOD1 pathway activation via ccRCC‐EVs.

3.5. SCRIB is Essential for ccRCC‐EV‐Mediated NOD1 Activation in Macrophages

To validate the role of SCRIB in NOD1 activation, we generated SCRIB‐knockout (SCRIB‐KO) 786O and 769P cell lines. EVs isolated from these cells were confirmed to lack SCRIB protein (Figure 5A). Treatment of THP‐1 and hPBM macrophages models with SCRIB‐KO EVs markedly reduced p‐RIP2 and p‐P65 levels compared to WT EVs (Figure 5B), indicating impaired NOD1 signalling.

FIGURE 5.

SCRIB is essential for NOD1 activation by ccRCC‐EVs in macrophages. (A) Western blot verifying the knockdown efficiency of SCRIB in cell lysate and EVs derived from SCRIB‐KO 786O and 769P cells. (B) Western blot detecting levels of p‐RIP2, RIP2, p‐P65 and P65 in THP‐1 and hPBM macrophage models treated with SCRIB‐KO EVs and WT EVs. (C) Western blot detecting levels of pro‐IL1β, IL6, iNOS, IL10, ARG1 and CD163. (D) RT‐qPCR detecting relative levels of pro‐IL1β, IL6, iNOS, IL10, ARG1 and CD163. (E) Western blot detecting levels of SCRIB, p‐RIP2, RIP2, p‐P65, P65, pro‐IL1β, IL6, iNOS, IL10, ARG1 and CD163 in SCRIB‐OE and Ctrl THP‐1 and hPBM macrophage model treated with SCRIB‐KO EVs. Data were presented as mean ± SD. A significant difference between the groups, *p < 0.05, **p < 0.01 and ***p < 0.001.

To further investigate the regulatory effect of EV‐derived SCRIB on macrophage functional phenotype, we assessed alterations in macrophage markers mediated by NOD1 pathway activation. Western blot analysis showed that, compared to WT controls, macrophages treated with SCRIB‐KO EVs exhibited significantly lower expression levels of IL‑10, ARG1, and CD163 in both THP‐1 and hPBM models, whereas no significant changes were observed in pro‑IL‑1β, IL‑6, and iNOS (Figure 5C). These findings were further corroborated by RT‐qPCR (Figure 5D). To determine whether SCRIB is sufficient to regulate NOD1 signalling, rescue experiments were performed by treating SCRIB‐overexpressing (SCRIB‐OE) and control macrophages (THP‐1 and hPBMs models) with SCRIB‐KO EVs. Western blot analysis confirmed increased SCRIB levels in macrophage from SCRIB‐OE group. Consequently, this group exhibited enhanced activation of the NOD1 pathway and significantly elevated expression of IL‑10, ARG1, and CD163 compared to the Ctrl group (Figure 5E). Together, these results demonstrate that SCRIB is indispensable for ccRCC‐EV‐induced NOD1 activation and subsequent acquisition of an immunosuppressive phenotype by macrophages.

3.6. NOD1 Activation by ccRCC‐EVs Induces an Immunosuppressive Phenotype in Macrophages via IRF4

The molecular mechanisms through which sustained NOD1 pathway activation drives macrophages toward an immunosuppressive functional state remain to be fully elucidated. To investigate NOD1‐mediated molecular alterations in macrophages, bulk RNA‐sequencing was performed on NOD1‐knockout (KO) and WT THP‐1 macrophage model. First, we examined alterations in upstream pathways regulating NOD1 activation. GSVA analysis revealed remarkable enhancement of Rho GTPase pathway activity but no changes in calcium ion response pathways (Figure 6A), aligning with our prior findings (Figure 4B). Next, we systematically analysed the signal changes downstream of NOD1. NOD1 signalling converges on three well‐characterized downstream mechanisms: 1) activation of the NF‐κB pathway via polyubiquitinated RIP2‐mediated TAK1 recruitment; 2) suppression of inflammatory cytokine expression through ATG16L1‐dependent autophagy; and 3) modulation of inflammatory responses and apoptosis by caspase‐induced mitochondrial damage (Omaru et al. 2022; Watanabe et al. 2010). Our data demonstrate that compared to NOD1‐KO cells, WT macrophages exhibited markedly enhanced TAK1 complex activity and NF‐κB pathway activation. In contrast, autophagy signalling showed only marginal enhancement (non‐significant p‐value), and mitochondrial signalling remained unaltered (Figure 6A). Protein‐level validation confirmed these findings: NOD1 knockout reduced levels of phosphorylated RIP2 (p‐RIP2) and the NF‐κB subunit P65 (p‐P65), while autophagy‐related proteins (ATG16L1, p62) showed no significant changes (Figure 6B).

FIGURE 6.

NOD1 activation by ccRCC‐EVs induces immunosuppressive macrophages phenotypes through IRF4. (A) GSVA analysis of NOD1‐KO and WT THP‐1 macrophage model using RNA‐seq data. (B) Western blot detecting levels of p‐RIP2, RIP2, p‐P65, P65, ATG16L1 and P62 in NOD1‐KO and WT THP‐1 macrophage model treated with EVs. (C) Differentially expressed genes between NOD1‐KO (right) and WT group (left). (D) Heatmap of TF activity. (E) Differentially expressed gene of IRF family between NOD1‐KO (right) and WT group (left). (F) Western blot detecting levels of IRF4, IRF5 and IRF7. (G) Western blot detecting levels of p‐RIP2, RIP2 and IRF4 in OE‐Ctrl, OE‐NOD1, WT and KO‐NOD1 THP‐1 macrophage model. (H) Western blot detecting levels of IRF4, IL10, ARG1 and CD163 in NOD1‐KO and WT THP‐1 macrophage model treated with EVs. (I) GSEA analysis of JAK‐STAT pathway. (J) Western blot detecting levels of pSTAT6 and MITF in OE‐Ctrl, OE‐NOD1, WT and KO_NOD1 THP‐1 macrophage model.

Next, we further investigated the impact of NOD1 knockout on macrophage gene expression profiles. Through differential gene expression analysis, we identified that NOD1 KO significantly downregulated the expression of ZNF family genes, indicating substantial transcriptional alterations induced by NOD1 knockout. Additionally, we observed upregulated expression of inflammatory genes CXCL1/8 and IL1B following KO. The expression of S100A9 and IL4L1—factors promoting immunosuppressive macrophage programming and suppressing inflammatory activation—was markedly reduced, along with decreased expression of TNFAIP3, a regulator of transient inflammatory responses (Figure 6C). Collectively, these findings demonstrate that NOD1 KO shifts macrophages toward a less immunosuppressive, more inflammation‑permissive state.

To further explore the mechanism by which NOD1 signalling mediates the acquisition of an immunosuppressive phenotype, we analysed the differences in transcription factor (TF) activities between the two groups. Compared to WT cells, NOD1‐KO macrophages exhibited increased activity of the pro‐inflammatory transcription factor pSTAT1 and decreased activity of the anti‐inflammatory transcription factor pSTAT6, displaying a pattern opposite to that of macrophages with immunosuppressive features (Figure 6D). Interestingly, IRF4 and its key downstream TF PRDM1 (also known as BLIMP‐1) showed significantly lower activity in NOD1‐KO cells.

Previous studies have indicated that NOD1 exerts regulatory effects on the interferon regulatory factor (IRF) family. Watanabe et al. demonstrated that the NOD1 pathway regulates IRF7 to induce IFN‐β expression (Watanabe et al. 2010), but its regulatory relationships with other IRF family members remain unclear. To investigate whether IRF family members are involved in NOD1‐mediated immunosuppressive programming, we analysed the differential expression of IRF family genes between NOD1‐KO and WT groups and identified IRF4 as showing the most significant fold‐change difference between the two groups (Figure 6E). IRF4 is a key determinant of differentiation, activation and effector function for various immune cells, including CD8⁺ cytotoxic T cells, Treg cells and macrophages. IRF4 has been extensively reported to mediate macrophage immunosuppressive programming and exhibits an antagonistic relationship with IRF5, which drives pro‐inflammatory responses (Lu et al. 2023; Satoh et al. 2010; Negishi et al. 2005). However, whether IRF4 is regulated by NOD1 has not been previously reported. We performed Western blot analysis to validate the expression patterns of IRF4, IRF5, and IRF7. The results demonstrated pronounced downregulation of IRF4 levels upon NOD1 knockout (KO), whereas IRF5 and IRF7 expression showed no significant alterations (Figure 6F), suggesting a strong association between IRF4 expression and NOD1 activation.

Next, we validated the regulatory role of NOD1 on IRF4. Western blotting confirmed NOD1‐dependent IRF4 regulation: overexpression or pharmacological activation of NOD1 increased IRF4 levels, while NOD1 KO abolished this effect (Figure 6G). To confirm that EV‐induced NOD1 activation regulates macrophage function through IRF4, we assessed the expression of ARG1, IL‐10 and CD163. Correspondingly, these markers decreased proportionally with IRF4 reduction in KO cells (Figure 6H), suggesting that NOD1 depends on IRF4 to exert its regulatory effect on macrophage phenotype.

We next sought to investigate the mechanisms underlying NOD1‐mediated regulation of IRF4. It has been reported that the immunomodulatory function of IRF4 is activated by NF‐κB and pSTAT6, and is inhibited by MITF (Shaffer et al. 2009; Lee et al. 2024). Our prior analyses confirmed NF‐κB pathway activation (Figure 6A, 6B) but revealed no significant alterations in the JAK‐STAT pathway via GSEA (Figure 6I). Western blot analysis further demonstrated modest alterations in pSTAT6 levels following NOD1 overexpression (OE) or knockout (KO) (Figure 6J), suggesting that pSTAT6 may not be the dominant mediator of NOD1's regulatory effects on IRF4. Notably, we observed that MITF—a transcriptional repressor of IRF4—exhibited decreased protein levels in NOD1‐OE cells and increased levels in NOD1‐KO cells (Figure 6J). These findings suggest that NOD1 may regulate IRF4 through NF‐κB signalling and MITF suppression. However, the detailed mechanistic interactions between these factors require further investigation.

3.7. Targeting the SCRIB‐NOD1 Axis in Vivo Suppresses Tumour Growth and Enhances Anti‐PD‐1 Efficacy

Next, we investigated the in vivo functional role of the SCRIB‐NOD1 axis and its influence on ICI efficacy. To validate the role of tumour cell‐derived SCRIB in activating NOD1 in macrophages in vivo, we established subcutaneous ccRCC models in mice using SCRIB‐knockout (SCRIB‐KO) or WT Renca cells. Compared to WT group, mice in the SCRIB‐KO group exhibited significantly slower tumour growth (Figure 7A, 7B). H&E staining of harvested mouse tumours are shown in Figure S7. mIF analysis of harvested tumours revealed increased infiltration of CD86⁺ macrophages, reduced accumulation of CD163⁺ macrophages, and diminished exhausted CD8⁺ T cells (CD8E) in tumours from the SCRIB‐KO group (Figure 7C, 7D). Additionally, tumours from the SCRIB‐KO group exhibited elevated IL‐1β and IL‐6 levels, but reduced expression of phosphorylated RIP2 (p‐RIP2) and IRF4 (Figure 7E). These findings indicate that SCRIB knockout in tumour cells suppresses NOD1 activation in macrophages, drives a phenotypic shift within the macrophage compartment toward a pro‐inflammatory profile, and thereby enhances anti‐tumour immunity.

FIGURE 7.

Targeting the SCRIB‐NOD1 axis in vivo suppresses tumour growth and boosts anti‐PD‐1 efficacy. (A) Harvested tumour volume from mice implanted with SCRIB‐KO and WT Renca cells. (B) Tumour growth curves of mice in SCRIB‐KO and WT group. (C) Representative mIF images of tumours from mice in SCRIB‐KO and WT group (markers: Cd86 (purple), Cd163 (light blue), Cd8a (red), Pdcd1 (green). Cd86⁺ cells were used as an indicator of a pro‑inflammatory macrophage phenotype. Cd163⁺ cells were used as an indicator of an immunosuppressive macrophage phenotype. Cd8a⁺ Pdcd1⁻ representing non‐exhausted CD8⁺T cells, Cd8a⁺ Pdcd1⁺ representing exhausted CD8⁺T cells) (scale bar = 50 µm), and (D) quantitative analysis of pro‑inflammatory macrophage, immunosuppressive macrophage, non‐exhausted CD8⁺T cells and exhausted CD8⁺T cells. (E) Relative levels of SCRIB in Renca cell and IL‐1β, IL6, CD163, IL10, p‐RIP2, RIP2, IRF4 in MM. (F) Design of in vivo therapeutic experiment of the NOD1 inhibitor Nodinitib‐1 combined with ICI (Nodinitib‐1 group) versus ICI monotherapy (Ctrl group). (G) Harvested tumour volume from mice in Nodinitib‐1 group and Ctrl group. (H) Tumour growth curves of mice in Nodinitib‐1 group and Ctrl group. (I) Representative mIF images of tumours from mice in Nodinitib‐1 group and Ctrl group (markers: Cd86 (purple), Cd163 (light blue), Cd8a (red), Pdcd1 (green). Cd86⁺ cells were used as an indicator of a pro‑inflammatory macrophage phenotype. Cd163⁺ cells were used as an indicator of an immunosuppressive macrophage phenotype. Cd8a⁺ Pdcd1⁻ representing non‐exhausted CD8⁺T cells, Cd8a⁺ Pdcd1⁺ representing exhausted CD8⁺T cells) (scale bar = 50 µm) and (J) quantitative analysis of pro‑inflammatory macrophage, immunosuppressive macrophage, non‐exhausted CD8⁺T cells and exhausted CD8⁺T cells. Data were presented as mean ± SD. A significant difference between the groups, *p < 0.05.

Given the clinical application of ICIs in advanced and metastatic ccRCC and our prior findings linking NOD1 to ICI resistance, we further assessed whether NOD1 inhibition improves ICI efficacy. In mice implanted with Renca cells, we compared the therapeutic efficacy of the NOD1 inhibitor Nodinitib‐1 combined with anti‐PD‐1 (Nodinitib‐1 group) versus anti‐PD‐1 monotherapy (Ctrl group). Compared to the Ctrl group, tumour growth was remarkably suppressed in the Nodinitib‐1 group (Figure 7G, 7H). Immunofluorescence analysis demonstrated reduced CD163⁺ macrophages recruitment and decreased CD8⁺ T cell exhaustion in the Nodinitib‐1 group compared to the Ctrl group (Figure 7I, 7J).

To further delineate the functional status of macrophages, we incorporated the myeloid marker Cd11b, the pan‐macrophage marker F4/80, and the immunosuppressive marker Cd206 into the immunofluorescence panel (Figure S8–S11). The results revealed that the majority of F4/80⁺ cells co‐expressed Cd11b, thereby enabling precise identification of macrophages as F4/80⁺Cd11b⁺ cells. Notably, most Cd86 signals were detected within this macrophage population, and similarly, Cd206⁺ or Cd163⁺ signals were predominantly localized to macrophages, thereby confirming the robustness of the data. We subsequently defined F4/80⁺Cd11b⁺Cd86⁺ cells as pro‐inflammatory macrophages and F4/80⁺Cd11b⁺Cd163⁺ or F4/80⁺Cd11b⁺Cd206⁺ cells as immunosuppressive macrophages. Consistent with our previous findings, SCRIB‐KO resulted in a reduction of immunosuppressive macrophages (Figure S12A). The combination therapy using Nodinitib‐1 and anti‐PD‐1 also showed a reduction in immunosuppressive macrophages, and an increase in pro‐inflammatory macrophages (Figure S12B). These results indicate that blocking the SCRIB‐NOD1 regulatory axis represents a potential therapeutic approach for ccRCC. A schematic summary of the key findings of this study is presented in Figure 8.

FIGURE 8.

Schematic depiction of tumour‐derived EVs reprogramming macrophages into immunosuppressive phenotype via SCRIB‐NOD1‐IRF4 regulatory axis in ccRCC. In normal tissues, transient NOD1 activation by pathogens promotes inflammatory gene expression. However, in ccRCC, tumour‐derived EVs deliver SCRIB to macrophages, leading to sustained NOD1 activation. Prolonged exposure to ccRCC‐EVs results in persistent NOD1 pathway activation, which reprograms macrophages toward an immunosuppressive phenotype via IRF4 regulation. Mechanistically, NOD1 may regulate IRF4 through both NF‐κB signalling activation and MITF suppression. Abbreviations: ccRCC: clear cell renal cell carcinoma.

4. Discussion

EVs act as intercellular “messengers” by delivering their cargo (proteins, lipids or nucleic acids) to neighbouring or distant cells. Previous studies have demonstrated that tumour‐derived EVs can modulate diverse immune cells—including macrophages and T cells—to facilitate immune evasion and metastasis (Maacha et al. 2019; Maybruck et al. 2017), underscoring the broad research potential of EV‐mediated regulatory networks. However, the complexity of EV composition and intercellular communication poses challenges for mechanistic investigations. In this study, we reveal that a pattern recognition receptor, NOD1, is activated by ccRCC‐derived EVs, leading to functional reprogramming of macrophages toward an immunosuppressive phenotype.

NOD1 signalling has been implicated in tumour progression across multiple malignancies, particularly in colorectal and gastric cancers where microbial involvement is prominent (Maisonneuve et al. 2021; Zhang et al. 2024; Wang et al. 2012; Jiang, 2020; Li et al. 2024; Ma et al. 2024). To our knowledge, no prior studies have investigated the role of NOD1 in renal cell carcinoma. Through integrated snRNA‐seq, IHC, and multiplex immunofluorescence analyses, we demonstrate that NOD1 in ccRCC is predominantly expressed in myeloid cells and is associated with an immunosuppressive shift in the tumour immune microenvironment. In contrast to existing reports, Wei et al. observed that NOD1 activation in colorectal cancer macrophages exacerbates inflammation and metastasis (Wei, 2022), while Zhang et al. found that NOD1 activation in macrophages enhances CD8+ T cell‐mediated antitumor immunity in hepatocellular carcinoma (Zhang et al. 2024). Our findings reveal a time‐dependent functional shift in ccRCC: transient EV exposure initially elevates inflammatory gene expression, whereas sustained EV stimulation induces immunosuppressive reprogramming. This temporal shift reflects evolutionarily conserved features of NOD1 signalling observed in infection models (Maisonneuve et al. 2021), yet appears to be co‐opted by tumours to establish a chronic immunosuppressive niche. The functional discrepancies observed across different tumour types may be closely associated with their specific TME contexts. In microbially rich tumours (e.g., colorectal/hepatic cancers), pathogen‐derived signals likely dominate NOD1 activation. Conversely, in ccRCC—a tumour type with limited microbial influence—NOD1 activation is primarily mediated by tumour‐derived EVs, potentially explaining its divergent downstream regulatory consequences.

Deciphering the complex composition of EVs remains a major challenge in EV research. Through EV proteomic profiling, we identified SCRIB—an activator of small Rho GTPases—as the key mediator of NOD1 activation in ccRCC‐derived EVs. SCRIB, a non‐secretory protein, relies on EVs for intercellular transfer. Our results identified SCRIB as the critical EV cargo driving macrophages toward an immunosuppressive phenotype. SCRIB primarily acts as a scaffold protein within the cell (Santoni et al. 2020). SCRIB can recruit or coordinate the spatial distribution and activity of Rho family guanine nucleotide exchange factors (GEFs) and GTPase‐activating proteins (GAPs), thereby regulating the activity of Rho GTPases (such as RhoA and Rac1) (Matsuda et al. 2020; Matsuda et al. 2022; Alexander et al., 2021). This Rho GTPase signalling serves as a critical initiating signal for NOD1 activation (Keestra et al. 2013). However, the intermediary mechanisms underlying the SCRIB‐NOD1 axis regulation remain to be explored. Notably, residual NOD1 pathway activation observed in THP‐1 macrophage model treated with SCRIB‐KO EVs suggests compensatory mechanisms from other EV components. Recent studies have demonstrated that EV‐delivered CDC42 (a small Rho GTPase) activates NOD1 in colorectal cancer macrophages (Wei et al., 2022). While we observed no significant difference in CDC42 expression between ccRCC‐EVs and normal renal tubular cell EVs via Western blot, its presence may contribute to partial signalling. Similarly, ccRCC‐EVs contain RNA cargoes (e.g., circSAFB2, lncARSR) reported to independently induce an immunosuppressive phenotype in macrophages (Zhang et al. 2022; Huang, 2022), potentially explaining the modest immunosuppression‐related markers elevation in NOD1‐KO macrophages following EV treatment. Notably, our functional assays demonstrate SCRIB is the dominant NOD1 activator in ccRCC‐EVs. Therapeutically, blocking the SCRIB‐NOD1 axis has the potential to alleviate macrophage‐mediated immunosuppression, particularly when combined with ICIs. Future studies may focus on the precise molecular mechanisms by which SCRIB‐NOD1 axis influences anti‐PD‐1 efficacy.

IRF4 is a well‐established mediator of immunoregulatory functions in macrophages—known to promote IL‐4 and IL‐10 secretion through the IRF4/PRDM1 axis to drive an immunosuppressive phenotype and to competitively bind MyD88, thereby inhibiting IRF5/MyD88‐mediated pro‐inflammatory responses (Satoh et al. 2010; Negishi et al. 2005; Jefferies et al. 2019; Wang et al. 2024; Lin et al. 2004). Our study reveals that IRF4 is regulated by tumour cell‐derived SCRIB/NOD1 signalling in ccRCC. This finding suggests that the regulation of IRF4 may contribute to the NOD1‐driven immunosuppressive programming observed in our models. While our data point to the potential involvement of NF‐κB and MITF in this process, the precise upstream mechanism by which NOD1 regulates IRF4 falls beyond the scope of this study and remains an important area for future investigation.

It is noteworthy that the conceptual framework of macrophage biology has evolved considerably. With the advent of single‐cell technologies, the heterogeneity and plasticity of macrophages within tumours are increasingly appreciated, moving toward a spectrum of context‐dependent states defined by precise gene expression profiles (Ma et al. 2022; Chu et al. 2024; Mulder et al. 2021; Jiang et al. 2025). In this study, while we acknowledge the complexity revealed by high‐resolution approaches, we employed a set of widely reported marker genes associated with pro‐inflammatory (e.g., pro‐IL1B, IL6, CD86) or immunosuppressive (e.g., IL10, ARG1, CD163) functions as practical and interpretable surrogates to capture dominant functional shifts in macrophages upon EV‐mediated NOD1 activation. This approach allows us to align our mechanistic findings with a substantial body of existing literature and to communicate the functional consequences of the SCRIB–NOD1–IRF4 axis with clarity, while remaining consistent with the evolving understanding that macrophage phenotypes exist along a functional continuum shaped by the TME.

In summary, our findings demonstrate that SCRIB packaged within ccRCC‐derived EVs is transported to macrophages, where it activates NOD1 via Rho GTPase signalling. Sustained NOD1 activation enhances IRF4 transcriptional activity, thereby driving macrophage reprogramming toward an immunosuppressive phenotype that is associated with adverse clinical outcomes in ccRCC. Furthermore, our results suggest that targeting NOD1 could serve as a complementary strategy to ICI therapy. However, whether NOD1 regulates IRF4 through MITF‐dependent mechanisms requires further investigation. Additionally, detailed studies are needed to delineate the precise molecular interplay between SCRIB and NOD1. Collectively, this work provides novel insights into EV‐mediated NOD1 activation in ccRCC and identifies the SCRIB‐NOD1‐IRF4 regulatory axis as a promising therapeutic target for restoring antitumor immunity.

Author Contributions

Xin‐gang Cui: Supervision, Funding acquisition, Resources, Project administration. Ke‐qin Dong: Writing – review & editing, Resources, Formal analysis. Bin Fu: Methodology, Resources, Project administration. Yi‐fan Tang, Hong‐feng Zheng: Data curation, Investigation. Yi‐fan Liu, Jia‐lin Zhou: Investigation, Validation. Mu‐chen Li: Writing – original draft, Formal analysis, Visualization, Software. Wen‐jie Ma: Validation. Jian‐gui Liu, Zi‐chang Liu, Wen‐yan Li, Zi‐xuan Gong, Shun Zhang, Hang‐biao Zhang, Bing‐nan Lu, Yun‐tao Yao, Dong‐hao Lyu, Guan‐yu Zhu, Yuan‐bo Zong, Han‐lin Yao, Ai‐ping Zhang: Data curation. Xiu‐wu Pan: Writing – review & editing, Formal analysis, Visualization. Tian‐yue Yang: Investigation, Validation, Software. Jin Zeng: Methodology. Si‐tu Xiong: Methodology, Resources. Wang Zhou: Writing – review & editing, Supervision, Methodology, Conceptualization, Project administration.

Ethic Statements

All study protocols adhered to the REMARK guidelines and were approved by the Research Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (No: XHEC‐C‐2021‐145‐1). Written informed consent was obtained from all participating patients before initiation of the study.

Conflicts of Interest

The authors report no conflict of interest.

Supporting information

Supporting Information: jev270317‐sup‐0001‐SuppMat.docx

Data Availability Statement

The TCGA‐KIRC sample data were obtained from https://xena.ucsc.edu/ of UCSC Xena. snRNA‐seq data were obtained from the National Center for Biotechnology Information Gene Expression Omnibus (accession no. GSE240822, GSE227898, GSE151302, GSE183277). All primary data presented in this study are available from the corresponding author upon reasonable request.