Gastric Cancer‐Derived Exosomal MATN3 Favours Immunosuppressive Tumour Microenvironment by Activating Autophagy in an EGFR/ELK1/ATG12 Signalling Dependent Manner

ABSTRACT

Gastric cancer (GC)‐derived exosomes (Exos) have been identified to facilitate GC progression by inducing M2 macrophage polarization. This study investigated the biological function of exosomal matrilin‐3 (MATN3) in M2 macrophage polarisation during GC development and its underlying mechanism. Exos were isolated from GC cells and then co‐cultured with THP‐1‐derived macrophages. Macrophage polarisation was evaluated by measuring the levels of M1/M2 macrophage markers. Target molecule expression was evaluated by RT‐qPCR, Western blotting and immunohistochemical staining. LC3II expression and the co‐localisation of MATN3 and epidermal growth factor receptor (EGFR) were determined by immunofluorescent staining. In vivo growth of GC cells was assessed in a xenograft mouse model. Molecular mechanisms were analysed by Co‐IP, ChIP, dual‐luciferase reporter assay and ubiquitination assay. MATN3 was highly expressed in GC and its high expression was negatively associated with the overall survival and M1 macrophage marker expression of GC patients. The in vitro experiments validated that MATN3 was secreted by GC‐Exos, which promoted M2 macrophage polarisation via autophagy activation. In addition, exosomal MATN3 contributed to in vivo growth of GC cells via promoting M2 macrophage infiltration. Mechanistically, MATN3 interacted with EGFR to enhance its protein stability, which activated Ets‐like protein‐1 (ELK1) and consequently promoted ATG12‐mediated autophagy. Activation of the EGFR/ELK1 pathway abolished exosomal MATN3 silencing‐mediated inhibitory effect on autophagy and M2 macrophage polarisation. GC‐derived exosomal MATN3 exerted an oncogenic role by inducing M2 macrophage polarisation via activation of the EGFR/ELK1/ATG12 axis‐mediated autophagy, which provides potential therapeutic targets for GC.

GC‐delivered exosomal MATN3 interacted with EGFR to activate ELK1 that subsequently triggered ATG12 transcription to induce autophagy, thereby facilitating immunosuppressive M2 macrophage polarisation to drive GC tumourigenesis.

Abbreviations

ANOVA

analysis of variance

ChIP

chromatin immunoprecipitation

Co‐IP

co‐immunoprecipitation

EGFR

epidermal growth factor receptor

ELK1

Ets‐like protein‐1

EMT

epithelial‐mesenchymal transition

Exos

exosomes

GC

gastric cancer

MATN3

matrilin‐3

PMA

phorbol 12‐myristate 13‐acetate

RT‐qPCR

real‐time quantitative PCR

SD

standard deviation

sh‐MATN3

short hairpin RNA targeting MATN3

TEM

transmission electron microscopy

TME

tumour microenvironment

1. Introduction

Gastric cancer (GC) remains a frequent malignancy of the digestive system, with high morbidity and mortality [1]. GC ranks as the fourth main cause of cancer‐related mortality worldwide [2]. Currently, surgery remains one of the effective treatments for GC patients at early stage [3]. However, the mortality is still high for GC patients at advanced stage due to lack of available therapies [4]. Hence, the identification of early diagnostic biomarkers and precise therapeutic targets is urgently needed to improve the prognosis of GC patients.

Matrilin‐3 (MATN3) belongs to the MATN family, which participates in cartilage and bone development [5]. Recent studies have documented the abnormal expression of MATN3 in various cancers, including GC. For example, MATN3 was reported to be up‐regulated in GC, indicating a poor prognosis of GC patients [6]. Another study suggested that MATN3 might serve as a diagnostic and prognostic biomarker for GC patients [7]. Furthermore, previous studies indicated that the dysregulation of MATN3 was associated with immune escape, epithelial–mesenchymal transition (EMT) and macrophage infiltration during GC progression [8, 9, 10]. However, the biological function of MATN3 and its detailed mechanisms in GC remain unclear.

As reported, exosomes (Exos)‐derived MATN3 was correlated with stromal activation and immunosuppression, which could serve as a prognostic marker for GC [10], hinting at the potential role of MATN3 in tumour microenvironment (TME). Macrophages are a major component in TME, and play crucial roles in tumour development [11]. Macrophages can be polarised into the immunostimulatory M1 and immunosuppressive M2 phenotypes under specific conditions. M1 macrophages possess anti‐tumour effects, while M2 macrophages can facilitate cancer growth and immune evasion [12]. Autophagy is a biological process that is essential for the degradation of misfolded proteins and damaged organelles. Autophagy has been validated to modulate the polarisation of macrophages during tumour development. For example, autophagy activation contributes to M2 macrophage polarisation, thereby contributing to tumour growth and metastasis [13]. Zhang et al. reported that LL‐37 delayed pancreatic cancer progression via repressing autophagy to inhibit M2 macrophage polarisation [14]. Another study showed that gold nanoparticles served as an antitumour immunotherapy via reducing M2 polarisation of tumour‐associated macrophages by autophagic flux inhibition [15]. Therefore, modulation of macrophage polarisation via autophagy inhibition has been considered as a novel strategy for cancer immunotherapy. Notably, a previous study demonstrated that synovial mesenchymal stem cells‐derived exosomal MATN3 alleviated osteoarthritis via activating autophagy [16]. Whether GC‐derived exosomal MATN3 can affect autophagy and consequently regulate macrophage polarisation in TME of GC cells deserves to be investigated.

In addition, MATN3 contains four epidermal growth factor (EGF)‐like domains [17], suggesting the potential association between MATN3 and EGF/EGFR pathway. It has been reported that the EGF/EGFR pathway was an important oncogenic signalling pathway through activating autophagy [18, 19]. Based on the above background, we speculated that GC‐derived exosomal MATN3 promoted EGF/EGFR pathway‐mediated autophagy to trigger M2 macrophage polarisation, thus contributing to the progression of GC. This study verified this speculation using in vitro and in vivo experiments, which suggested that targeting MATN3 might be an effective therapy for GC patients.

2. Materials and Methods

2.1. Clinical Sample Collection

Forty‐six paired human GC and adjacent normal samples were collected from the First Affiliated Hospital of Guilin Medical University and stored in liquid nitrogen. The serum samples were obtained from 30 GC patients and 30 healthy volunteers. The experimental procedures were following the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of The First Affiliated Hospital of Guilin Medical University (No. 2022KJTLL‐10). Written informed consents were provided by all participants.

2.2. Cell Culture

Normal gastric epithelium cells (GES‐1), GC cells (AGS, HGC‐27, MKN‐45, MKN‐74) and THP‐1 cells were purchased from Procell (Wuhan, China). AGS cells were maintained in Ham's F‐12 (Thermo Fisher, Waltham, MA, USA), while HGC‐27, MKN‐74, GES‐1, MKN‐45 and THP‐1 cells were cultured in RPMI‐1640 medium (Thermo Fisher). All culture medium was supplemented with 10% fetal bovine serum (FBS, Thermo Fisher), 100 U/mL penicillin and 100 U/mL streptomycin. All cells were maintained at 37°C with a 5% CO₂ atmosphere.

2.3. Cell Transfection

Short hairpin RNA (shRNA) targeting MATN3 (sh‐MATN3#1, shMATN3#2, shMATN3#3), ELK1 (shELK1) and overexpression vectors for MATN3 (pcDNA3.1‐MATN3), ELK1 (pcDNA3.1‐ELK1) and ATG12 (pcDNA3.1‐ATG12) and negative controls (shNC, pcDNA3.1) were obtained from GenePharma (Shanghai, China). The above shRNAs or vectors were transfected into GC cells or THP‐1‐derived macrophages by using Lipofectamine 2000 (Thermo Fisher) according to the manufacturer's instructions.

2.4. Macrophage Induction

To induce M0 macrophages, THP‐1 cells were seeded into 6‐well plates (2 × 10⁶ cells per well) and treated with 50 nM phorbol 12‐myristate 13‐acetate (PMA, Sigma‐Aldrich, USA) for 48 h [20].

2.5. Real‐Time Quantitative PCR (RT‐qPCR)

Total RNA was isolated using the RNAsimple Total RNA Kit (TianGen, Beijing, China). The total RNA (1 μg) was subjected to reverse transcription using the All‐in‐One First‐Strand cDNA Synthesis Kit (GeneCopoeia, USA) and then amplified using the FastKing One Step RT‐qPCR Kit (SYBR) (TianGen). GAPDH was considered as the reference gene and the relative mRNA expression was calculated using the 2^–ΔΔCt method. The primer sequence was provided in Table 1.

TABLE 1.

Oligonucleotide primer sets for qRT‐PCR.

Name	Sequence (5′¬3′)	Length
MATN3 F	TCTCCCGGATAATCGACACTC	21
MATN3 R	CAAGGGTGTGATTCGACCCA	20
iNOS F	GGCCTGGAAACGCACAAGCTG	21
iNOS R	TTGGGGTTGAAGGCACAGCTG	21
CD86 F	ACAGATGTCCTACGGGAACG	20
CD86 R	ATCCCACCTTAGAGCCAGGT	22
Arg1 F	AAGATTCCCGATGTGCCAGG	20
Arg1 R	GTCCACGTCTCTCAAGCCAA	20
IL‐10 F	TGAGAACAGCTGCACCCACT	20
IL‐10 R	GGCAACCCAGGTAACCCTTA	20
ELK1 F	TCCCTGCTTCCTACGCATACA	21
ELK1 R	GCTGCCACTGGATGGAAACT	20
ATG12 F	TTTGCTAAAGGCTGTGGG	18
ATG12 R	AAGGAGCAAAGGACTGAT	18
GAPDH F	AAATCCCATCACCATCTTCC	20
GAPDH R	TCACACCCATGACGAACA	18

2.6. Western Blotting

Total protein was isolated using the RIPA lysis buffer (Beyotime, Haimen, China) and quantified using the BCA protein kit (Thermo Fisher). The equal amount of protein samples was separated using 10% SDS‐PAGE and then blotted onto the PVDF membranes. Subsequently, the membranes were blocked with 5% skim milk and probed with primary antibodies against MATN3 (PA5‐20727, 1:1000, Thermo Fisher), CD63 (bs‐43117R, 1:500, Bioss, Beijing, China), TSG101 (bs‐1365R, 1:500, Bioss), Calnexin (bsm‐52639R, 1:500, Bioss), LC3B (PA1‐46286, 1:1000, Thermo Fisher), p62 (bs‐55207R, 1:1000, Bioss), EGFR (MA5‐13269, 1:500, Thermo Fisher), p‐ELK1 (bs‐5335R, 1:500, Bioss), ELK1 (ab125085, 1:2000, Abcam, UK) and GAPDH (bs‐41373R, 1:5000, Bioss) at 4°C overnight. The membranes were then incubated with the secondary antibody (bs‐0295D, 1:10000, Bioss) for 1 h. The ECL Western Blotting Substrate (Thermo Fisher) was utilised for visualisation of protein blots, followed by quantification using the Image J software.

2.7. Immunohistochemical Staining

The tumour tissues were fixed with 4% paraformaldehyde, followed by hydration with ethanol solution, embedding in paraffin and cutting into 5 μm sections. After treatment with 3% hydrogen peroxide, the sections were incubated with the primary antibodies against MATN3 (PA5‐112604, 1:50, Thermo Fisher), CD163 (bsm‐54015R, 1:100, Bioss) or CD86 (bsm‐52375R, 1:100, Bioss) overnight at 4°C. After incubation with the secondary antibody (bs‐0295D, 1:200, Bioss) at room temperature for 1 h, the sections were developed by DAB and counterstained with haematoxylin. The images were examined under a light microscope (Olympus, Japan).

2.8. Extraction and Identification of Exos

Exos were isolated from HGC‐27 and MKN‐45 cells according to a previous study [20]. In short, the cell supernatant was centrifuged at 3000 g to remove the cell debris. Subsequently, the cell supernatant was centrifuged at 10 000 g for 30 min to remove large‐sized shedding vesicles. Finally, the Exos were isolated by centrifugation at 110 000 g for 70 min, dissolved in phosphate‐buffered saline and stored at −80°C. The morphology of Exos was observed by the transmission electron microscopy (TEM, Thermo Fisher, USA). The size of Exos was evaluated by nanoparticle tracking analysis (NTA) using the NanoSight NS 300 system (Malven, UK). In addition, the protein levels of Exo markers CD63, TSG101 and endoplasmic reticulum membrane marker Calnexin were assessed by Western blotting.

2.9. Uptake of Exos

GC cells‐derived Exos (50 μL) were labelled with the PKH26 solution (100 μL, Sigma‐Aldrich) for 5 min. After stopping the staining by 250 μL of sterile FBS, the labelled Exos were added to THP‐1 cells and incubated for 24 h. Then, THP‐1 cells were fixed with 4% paraformaldehyde and stained with DAPI. The uptake of Exos by THP‐1 cells was observed under a fluorescence microscope (Olympus).

2.10. Immunofluorescent Staining

THP‐1‐derived macrophages were fixed with 4% paraformaldehyde, permeabilized with 0.3% Triton X‐100 and then blocked with 5% BSA (Thermo Fisher). For autophagy evaluation, the macrophages were incubated with LC3B antibody (PA1‐46286, 1:100, Thermo Fisher) overnight at 4°C. To observe the colocalization of EGFR and MATN3, the macrophages were incubated with primary antibodies against EGFR (MA5‐13269, 1:100, Thermo Fisher) and MATN3 (PA5‐20727, 1:400, Thermo Fisher) overnight at 4°C. Subsequently, the cells were washed and incubated with horseradish peroxidase‐conjugated secondary antibodies for 1 h, followed by nuclear staining with DAPI. Finally, the images were taken under a fluorescence microscope (Olympus).

2.11. Flow Cytometry

After various treatments, THP‐1‐derived macrophages were collected and suspended in PBS, followed by incubation with FITC Mouse Anti‐Human CD163 (BD Biosciences, USA) for 30 min. After washing with PBS, the percentage of CD163 in macrophages was analysed on the Flow Cytometer System (BD Biosciences).

2.12. Co‐Immunoprecipitation (Co‐IP)

The interplay between MATN3 and EGFR proteins was validated by Co‐IP assay using the Pierce Classic Magnetic IP/Co‐IP Kit (Thermo Fisher). Briefly, THP‐1‐derived macrophages were lysed with IP lysis buffer containing protease inhibitor cocktail (MCE, USA). The cell lysates were incubated with anti‐EGFR (MA5‐13269, 1:100, Thermo Fisher) or anti‐IgG at 4°C overnight. Then, the protein A/G PLUS‐Agarose (SantaCruz, USA) was added to the cell lysates and incubated for another 3 h. After washing, the precipitated proteins were subjected to Western blotting.

2.13. In Vitro Ubiquitination Assay

THP‐1‐derived macrophages were co‐transfected with HA‐Ub plasmid and pcDNA3.1‐MATN3 or pcDNA3.1 vector using Lipofectamine 2000. Subsequently, the cells were collected and treated with MG132 for 6 h. The lysates of THP‐1‐derived macrophages were incubated with anti‐EGFR (MA5‐13269, 1:100, Thermo Fisher) at 4°C overnight and further incubated with the protein A/G PLUS‐Agarose (SantaCruz) for 3 h. Next, the precipitated protein complexes were collected and subjected to Western blotting using anti‐HA (#32‐6700, 1:500, Thermo Fisher) and anti‐EGFR (MA5‐13269, 1:1000, Thermo Fisher).

2.14. Chromatin Immunoprecipitation (ChIP)

The molecular interplay between ELK1 and ATG12 was validated using ChIP assay. Briefly, THP‐1‐derived macrophages were fixed with formaldehyde and then treated with glycine. Next, the chromatins in cell lysates were truncated ultrasonically. Thereafter, the lysates were immunoprecipitated with anti‐ELK1 antibody (ab125085, Abcam) conjugated with magnetic bead with rotation. Next, the immunoprecipitated chromatins were eluted and the targeted fragments at the ATG12 promoter region were measured by RT‐qPCR method. The primer sequences are as follows: F1, 5′‐ACACCTGAAAGTCACACTGGG‐3′ (forward) and 5′‐CTGCCTTCCTGCCGCC‐3′ (reverse); F2, 5′‐GAGCCTCGCCTCGTAATCTC‐3′ (forward) and 5′‐TGTCTGTGCTCGGTCGAATG‐3′ (reverse).

2.15. Dual Luciferase Reporter Assay

The fragments at the ATG12 promoter region containing the binding sites to ELK1 were cloned and inserted into the pGL3 vector (Promega, USA) to establish the pGL3‐ATG12 luciferase reporter vector. Then, the recombinant pGL3‐ATG12 luciferase reporter vector was co‐transfected with pcDNA3.1‐ELK, sh‐ELK1 or their negative controls (pcDNA3.1, shNC) into macrophages using Lipofectamine 2000. Forty‐eight hours later, the luciferase activity was assessed using the dual‐luciferase reporter gene assay kit (Promega).

2.16. Xenograft Experiment

Female BALB/C nude mice (6 weeks old) were obtained from Slac Jingda Laboratory Animal Co. Ltd. (Hunan, China). After adaptive feeding for 1 week, 5 × 10⁶ HGC‐27 cells were mixed with 2 × 10⁶ THP‐1‐derived macrophages that were pretreated with Exo^‐control, Exo^‐shMATN3, Exo^‐shMATN3 + Rapamycin and were subcutaneously inoculated into the left armpit of mice (n = 5 per group). Subcutaneous tumour volume was determined via the formula (V = 1/2 × length × width [2]). Four weeks after the inoculation, all mice were euthanatized by cervical dislocation and tumours were collected and weighed. The animal experiments were performed according to the Guide for the Guilin Medical University Institutional Ethical Guidelines for animal experiments (GLMU‐IACUC‐202610036).

2.17. Statistical Analysis

Data are presented as the mean ± standard deviation (SD) from at least three independent biological replicates. Each biological replicate included three technical replicates. Normality of data distribution was assessed using the Shapiro–Wilk test and homogeneity of variances was verified using Levene's test. All analysed datasets were confirmed to meet the assumptions of normality and equal variance. All statistical analyses were performed with SPSS 19.0 software. Comparisons between two groups were performed using the unpaired Student's t‐test. Comparisons among multiple groups were analysed by one‐way analysis of variance (ANOVA) followed by Tukey's post hoc test. Survival curves were generated using the Kaplan–Meier method. p < 0.05 was indicated as statistical significance.

3. Results

3.1. High Expression of MATN3 Was Positively Correlated With Poor Prognosis and M2 Macrophage Infiltration

First, the differential expression of MATN3 was detected in 46 paired GC tumour and adjacent normal samples. It was shown that MATN3 mRNA and protein levels were higher in GC tissues, as compared with para‐cancerous tissues (Figure 1A,B). Kaplan–Meier analysis suggested that the overall survival was shorter in high MATN3 expression group than that in low MATN3 expression group (Figure 1C). In addition, we collected 30 paired serum samples from GC patients and normal volunteers. Whereas there was no significant difference in MATN3 mRNA level between GC and healthy groups (Figure 1D). Notably, we observed that MATN3 protein level was strikingly elevated in the serum of GC patients (Figure 1E). To further evaluate the association between MATN3 and macrophage infiltration, GC tissues were divided into MATN3‐low and ‐high groups according to the median expression of MATN3. Immunohistochemical staining showed that the expression of M2 macrophage marker CD163 was higher, while M1 macrophage marker CD86 expression was lower in MATN3‐high group as compared with MATN3‐low group (Figure 1F). Similarly, in comparison with MATN3‐low group, the levels of M2 macrophage markers (IL‐10, Arg1) were elevated, whereas the levels of M1 macrophage markers (iNOS, CD86) were down‐regulated in MATN3‐high group (Figure 1G). Collectively, these results indicated that high expression of MATN3 was correlated with a poor prognosis and M2 macrophage infiltration in GC.

FIGURE 1.

High expression of MATN3 indicated a poor prognosis and M2 macrophage infiltration of GC patients. (A, B) MATN3 expression in GC and paracancerous tissues was detected by RT‐qPCR and immunohistochemical staining. (C) Kaplan–Meier analysis evaluated the correlation between MATN3 expression and overall survival of GC patients. (D, E) RT‐qPCR and Western blotting analysis of MATN3 level in the serum samples of GC patients and healthy volunteers. GC patients were divided into MATN3‐high and MATN3‐low groups according to the median expression level of MATN3. (F) CD163 and CD86 expressions were determined by immunohistochemical staining. (G) RT‐qPCR analysis of iNOS, CD86, IL‐10 and Arg1 mRNA levels in GC samples. Student's t‐test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.2. GC Cells‐Derived Exosomal MATN3 Was Transferred to Macrophages

Considering that MATN3 high expression was correlated with M2 macrophage infiltration in GC tissues, we further validate this finding in vitro. Consistent with the clinical findings, we further confirmed that MATN3 mRNA expression was higher in multiple GC cell lines (AGS, HGC‐27, MKN‐45 and MKN‐74) than that in normal GES‐1 cells (Figure 2A). Given that the MATN3 protein level, but not mRNA level, was elevated in the serum samples of GC patients, we speculated that MATN3 protein might be released by GC cells via Exos. To verify our hypothesis, Exos were isolated from GC cells (HGC‐27, MKN‐45) and normal GES‐1 cells. TEM showed that Exos exhibited a cup‐shaped double‐membrane structure (Figure 2B) and NTA analysis further revealed that the diameter of Exos was 30–160 nm (Figure 2C). Additionally, western blotting confirmed the positive expression of CD63 and TSG101 proteins, but negative expression of Calnexin in Exos (Figure 2D). Importantly, the MATN3 protein level in GC cells‐derived Exos was significantly higher than that in GES‐1 cells‐derived Exos (Figure 2E). Next, GC cells‐derived Exos were co‐cultured with THP‐1‐differentiated macrophages. PKH26 staining demonstrated that GC cells‐derived Exos were successfully internalised by THP‐1 macrophages (Figure 2F). Moreover, co‐culture with GC cells remarkably enhanced the MATN3 protein level in macrophages, which was counteracted by co‐treatment with GW4869 (Figure 2G). All together, these findings proved that GC cells‐derived Exos could be transferred to macrophages and subsequently enhance MATN3 protein expression.

FIGURE 2.

GC cells transferred MATN3 to macrophages via Exos. (A) RT‐qPCR analysis of MATN3 level in GC cells and normal GES‐1 cells. (B) Exos isolated from GES‐1, HGC‐27 and MKN‐45 cells were observed by TEM. (C) Nanoparticle tracking analysis evaluated the diameter of Exos from GES‐1, HGC‐27 and MKN‐45 cells. (D) Western blotting analysis of the protein levels of CD63, TSG101 and Calnexin. (E) MATN3 expression in GES‐1, HGC‐27 and MKN‐45 cells‐derived Exos was detected by Western blotting. (F) The internalisation of GC cells‐derived Exos by macrophages was measured by PKH26 staining. (G) THP‐1‐derived macrophages were co‐cultured with HGC‐27 and MKN‐45 cells with or without GW4869 treatment. MATN3 level in macrophages was determined by Western blotting. One‐way ANOVA with Tukey's post hoc test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.3. Exosomal MATN3 Delivered From GC Cells Promoted M2 Macrophage Polarisation

To determine whether GC cells‐derived exosomal MATN3 could modulate macrophage polarisation, MATN3 was silenced by transfection with shMATN3#1,2,3 in HGC‐27 and MKN‐45 cells, respectively. RT‐qPCR and western blotting revealed that shMATN3#1/2/3 transfection effectively reduced MATN3 expression, among which shMATN3#2 possessed the highest knockdown efficiency in GC cells (Figure 3A,B). Consistently, exosomal MATN3 protein level was reduced after shMATN3#2 transfection (Figure 3C). Subsequently, GC cells‐derived Exos remarkably raised the percentage of CD163 positive macrophages; however, this effect was weakened by exosomal MATN3 depletion (Figure 3D). Accordingly, the mRNA levels of M2 macrophage markers (IL‐10, Arg1) were increased, whereas the mRNA levels of M1 macrophage markers (iNOS, CD86) declined after co‐culture with GC cells‐derived Exos and these results were abolished after MATN3 was depleted in GC cells‐derived Exos (Figure 3E). Thus, these results demonstrated that GC cells‐derived exosomal MATN3 contributed to M2 macrophage polarisation in vitro.

FIGURE 3.

GC cells induced M2 macrophage polarisation via delivering exosomal MATN3. (A, B) RT‐qPCR and western blotting detected MATN3 expression in GC cells transfected with shMATN3#1, 2, 3 or shNC. (C) MATN3 level in Exos isolated from GC cells from different groups was assessed by RT‐qPCR. THP‐1‐derived macrophages were co‐cultured with Exos from different groups for 48 h. (D) CD163 percentage in macrophages was determined by flow cytometry. (E) The mRNA levels of iNOS, CD86, Arg1 and IL‐10 in macrophages were measured by RT‐qPCR. One‐way ANOVA with Tukey's post hoc test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.4. Autophagy Played an Important Role in M2 Macrophage Polarisation Mediated by Exosomal MATN3

It has been recognised that autophagy is an essential biological process that modulates macrophage polarisation [21]. Thus, we further investigated the influence of GC cells‐derived exosomal MATN3 on autophagy of macrophages. Here, immunofluorescent staining showed that co‐culture with Exos^‐control could enhance LC3II expression in macrophages, which was abated by MATN3 knockdown in GC cells derived Exos (Exo^‐shMATN3 ) (Figure 4A). Accordingly, GC cells derived Exos‐mediated up‐regulation of LC3II/I ratio and down‐regulation of p62 in macrophages were partly counteracted by Exo^‐shMATN3 (Figure 4B). Notably, Exo^‐shMATN3‐induded decrease in percentage of CD163 positive macrophages was restored by treatment with autophagy activator rapamycin (Figure 4C). Furthermore, the down‐regulation of IL‐10, Arg1 and up‐regulation of iNOS, CD86 in Exo^‐shMATN3 group were also reversed by rapamycin treatment (Figure 4D). Collectively, the above observations indicated that exosomal MATN3 induced M2 macrophage polarisation via autophagy activation.

FIGURE 4.

Exosomal MATN3 contributed to M2 macrophage polarisation by autophagy activation. THP‐1‐derived macrophages were co‐cultured with Exos from various groups for 48 h. (A) LC3II expression in macrophages was observed by immunofluorescent staining. (B) The protein abundance of LC3 I/II, p62 in macrophages was analysed by Western blotting. THP‐1‐derived macrophages were treated with Exos^‐control, Exos^‐shMATN3 or Exos^‐shMATN3 combined with rapamycin for 48 h. (C) Flow cytometry determined CD163 percentage in macrophages. (D) The mRNA levels of iNOS, CD86, Arg1 and IL‐10 in macrophages were detected by RT‐qPCR. One‐way ANOVA with Tukey's post hoc test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.5. GC Cells‐Derived Exosomal MATN3 Accelerated In Vivo Tumour Growth by Promoting M2 Macrophage Infiltration

To further validate the above findings, HGC27 cells were mixed with macrophages with different treatments (Exo^‐control, Exo^‐shMATN3, Exo^‐shMATN3 + Rapamycin) and then intravenously injected into the nude mice to construct a xenograft tumour model. As shown in Figure 5A–C, the tumour volume and weight in Exo^‐control group were significantly elevated as compared with sham group; however, these changes were diminished in Exo^‐shMATN3 group. Meanwhile, the inhibitory effect of exosomal MATN3 knockdown on tumour growth was reversed by rapamycin administration (Figure 5A–C). Immunohistochemical staining showed that CD163 expression was higher in the Exo^‐control group, as compared with the sham group. Whereas, the elevated CD163 mediated by Exo^‐control was restored after MATN3 was depleted in Exos, and the inhibitory role of exosomal MATN3 knockdown in CD163 expression was abolished by rapamycin (Figure 5D). Furthermore, the RT‐qPCR assay indicated that the Exos^‐control‐mediated reduction in CD86 and iNOS levels and elevation in Arg1 and IL‐10 levels were abrogated by exosomal MATN3 deficiency; however, these changes mediated by exosomal MATN3 knockdown were abolished by co‐treatment with rapamycin (Figure 5E). The above results revealed that MATN3 secreted by GC cells‐derived Exos accelerated tumour growth by inducing macrophage M2 polarisation via autophagy activation.

FIGURE 5.

GC cells‐derived exosomal MATN3 promoted GC growth in vivo by inducing M2 macrophage infiltration. (A) Photographs of tumours collected from nude mice that were injected with HGC27 cells combined with Exos from different groups. (B) The tumour volume was detected. (C) The tumour weight was measured. (D) CD163 expression in tumours was evaluated by immunohistochemical staining. (E) The mRNA levels of iNOS, CD86, Arg1 and IL‐10 in tumour tissues were assessed by RT‐qPCR. One‐way ANOVA with Tukey's post hoc test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.6. MATN3 Directly Interacted With EGFR to Enhance EGFR Protein Stability in Macrophages

Next, we aimed to uncover the underlying molecular mechanism through which exosomal MATN3 induced M2 macrophage polarisation. Given that MATN3 contains four EGF‐like domains [17], we further explored the potential association between MATN3 and EGF. Immunofluorescence staining showed that MATN3 and EGFR were co‐localised in THP‐1‐derived macrophages (Figure 6A). Moreover, MATN3 protein could be immunoprecipitated by EGFR antibody (Figure 6B), indicating MATN3 could directly interact with EGFR protein. Furthermore, we observed that EGFR protein level in Exos^‐control‐co‐cultured macrophages was increased, which was weakened after down‐regulation of MATN3 in Exos (Figure 6C). To probe how MATN3 promoted EGFR protein expression, pcDNA3.1‐MATN3 was transfected into macrophages to overexpress MATN3 expression (Figure 6D). In vitro ubiquitination assay showed that MATN3 overexpression distinctly reduced the ubiquitin level of EGFR protein, thus enhancing EGFR protein level in macrophages (Figure 6E). To sum up, exosomal MATN3 enhanced EGFR protein stability in macrophages via inhibiting its ubiquitination.

FIGURE 6.

MATN3 was co‐localised with EGFR to enhance its stability in macrophages. (A) Co‐localisation of MATN3 and EGFR in macrophages was confirmed by immunofluorescent staining. (B) The interaction between MATN3 and EGFR proteins was validated by Co‐IP. (C) EGFR protein level in macrophages after co‐culture with Exos from each group was measured by Western blotting. (D) RT‐qPCR analysis evaluated MATN3 mRNA in macrophages transfected with pcDNA3.1‐MATN3. (E) The ubiquitin level of EGFR was evaluated by in vitro ubiquitination assay. Student's t test (for D) or one‐way ANOVA with Tukey's post hoc test (for C) was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.7. EGFR‐Dependent ELK1 Activation Enhanced ATG12 Transcription in Macrophages

EGFR has been documented to activate ELK1 in human brain endothelial cells [22]. Thus, we sought to examine the role of the EGFR/ELK1 axis in macrophages. As shown in Figure 7A, the protein levels of EGFR and p‐ELK1 in MATN3‐overexpressed macrophages were enhanced, whereas these alterations were neutralised by treatment with the EGFR inhibitor EGFR‐IN‐70. ELK1 plays a transcriptional regulator and its binding motif was obtained from the JASPAR database (Figure 7B). Notably, there were several putative binding sites of ELK1 to the ATG12 promoter (Figure 7C). Next, ChIP assay demonstrated that both the F1 and F2 fragments of the ATG12 promoter could be immunoprecipitated by ELK1 antibody in macrophages (Figure 7D). The luciferase activity of the ATG12 promoter was strikingly enhanced by ELK1 overexpression, but reduced by ELK1 silencing (Figure 7E). Besides, overexpression of ELK1 strikingly enhanced ATG12 expression, whereas ELK1 depletion resulted in the opposite result (Figure 7F,G). Taken together, EGFR facilitated ATG12 transcription and expression in macrophages by up‐regulation of ELK1.

FIGURE 7.

EGFR‐mediated ELK1 activation facilitated ATG12 transcription and expression in macrophages. (A) Macrophages were divided into control, pcDNA3.1‐NC, pcDNA3.1‐MATN3, pcDNA3.1‐MATN3 + EGFR‐IN‐70 groups. The protein levels of EGFR, p‐ELK1 and ELK1 in macrophages were determined by Western blotting. (B) The schematic diagram of ELK1 binding motif. (C) The schematic diagram of the potential binding sites of ELK1 to ATG12 promoter. (D) The enrichment of ATG12 promoter (F1 and F2 fragments) after immunoprecipitation by ELK1 antibody was measured by ChIP. (E) Dual luciferase assay validated the binding relationship between ELK1 and ATG12 promoter. (F, G) The mRNA levels of ELK1 and ATG12 in macrophages transfected with pcDNA3.1‐ELK1 or sh‐ELK1 were detected by RT‐qPCR. Student's t test (for D) or one‐way ANOVA with Tukey's post hoc test (for A, E–G) was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

3.8. Activation of EGFR/ELK1 Pathway Reversed Exosomal MATN3‐Induced Autophagy and M2 Macrophage Polarisation

To validate the involvement of the EGFR/ELK1 pathway in the biological functions of GC cells‐derived exosomal MATN3, the macrophages were co‐cultured with Exo^‐shMATN3 in combination with pcDNA3.1‐ATG12 transfection or treatment with NSC228155 (EGFR activator). Firstly, the overexpression efficiency of ATG12 in macrophages was confirmed, suggesting that ATG12 was successfully overexpressed after pcDNA3.1‐ATG12 transfection (Figure 8A). Next, immunofluorescent staining revealed that NSC228155 treatment or ATG12 overexpression evidently augmented LC3II expression in Exo^‐shMATN3 ‐co‐cultured macrophages (Figure 8B). Consistently, western blotting demonstrated that the protein level of p62 was reduced, but the LC3II/I ratio was raised by NSC228155 treatment or ATG12 overexpression in Exo^‐shMATN3 ‐treated macrophages (Figure 8C). Moreover, M2 macrophage polarisation was further promoted by NSC228155 co‐treatment or ATG12 overexpression as evidenced by up‐regulating CD163, Arg1, IL‐10 and down‐regulating iNOS and CD86 in the presence of Exo^‐shMATN3 co‐culture (Figure 8D,E). These data proved that exosomal MATN3 triggered autophagy and M2 macrophage polarisation via activation of the EGFR/ELK1 pathway.

FIGURE 8.

EGFR/ELK1 pathway was involved in exosomal MATN3‐induced autophagy and M2 macrophage polarisation. (A) The overexpression efficiency of pcDNA3.1‐ATG12 in macrophages was determined by RT‐qPCR. Exo^‐shMATN3‐treated macrophages were transfected with pcDNA3.1‐ATG12 or co‐treated with EGFR activator NSC 228155. (B) LC3 expression in macrophages was analysed by immunofluorescent staining. (C) The protein levels of LC3 I/II, p62 in macrophages were assessed by Western blotting. (D) The percentage of CD163‐positive macrophages was determined by flow cytometry. (E) The mRNA levels of iNOS, CD86, Arg1 and IL‐10 in macrophages were evaluated by RT‐qPCR. One‐way ANOVA with Tukey's post hoc test was performed for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

4. Discussion

As a main component of TME, tumour‐associated macrophages have been considered a crucial driver of cancer progression. Promotion of the transformation of TAMs from M2 to M1 phenotype has been recognised as an effective strategy for treating cancer [15]. Exos are small vesicles containing complex RNAs and proteins, which have been proved to mediate the communication between tumour cells and TME. For example, Liu et al. reported that HMGB1 released from GC cells‐derived Exos accelerated GC progression via promoting M2 macrophage polarisation [23]. Another study documented that GC‐derived Exos contributed to lung metastasis of GC cells through inducing immunosuppressive M2 macrophage polarisation [24]. In this work, we discovered that GC cells‐derived Exos transferred MATN3 to stabilise EGFR that subsequently activated ELK1/ATG12 axis‐mediated autophagy, thus triggering M2 macrophage polarisation to accelerate GC development. Our findings suggested that MATN3 might be a potential intervention target for GC patients.

Previous study have reported that high expression of MATN3, a key member of the ECM protein family, correlated with the poor prognosis in GC patients [7, 25]. Similar data were also confirmed in our investigation. Besides, our study revealed a remarkable increase in MATN3 protein level in the serum of GC patients, whereas there was no significant alteration in MATN3 mRNA level. This discrepancy led us to speculate that MATN3 might be encapsulated within Exos secreted by GC cells in a form encapsulated within exosomes, entering the bloodstream. Moreover, our data confirmed that MATN3 expression was substantially elevated in both GC cells and their secreted Exos. A prior study suggested the implication of the MATN3 gene in a cancer‐associated fibroblast score within the TME of GC patients [26], hinting at the potential role of MATN3 in tumour‐associated macrophages. Xu et al. reported a significant association between high expression level of MATN3 and macrophage infiltration in high‐risk GC patients [8]. In line with this study, we found that high expression of MATN3 was correlated with high expression of CD163 and low expression of CD86 in clinical tissues. On this basis, our findings uncovered a mechanism through which GC cells transferred MATN3 to macrophages via Exos, thereby orchestrating their polarisation towards an M2 phenotype. This provided the first evidence for the association of MATN3 with the immunomodulatory dynamics in the GC TME, which enriched our understanding of the complex interplay between cancer cells and TME.

As a pivotal biological process, autophagy is responsible for the delivery of cytoplasmic materials to lysosomes for degradation. It has been recognised that abnormal autophagy leads to immune dysfunction via regulation of TME [27]. Moreover, growing evidence has revealed the modulation of autophagy in tumour‐associated macrophage polarisation in the TME [13]. Autophagy activation could promote M2 macrophage polarisation by activating the STAT3 pathway [28]. Xu et al. documented that glioma‐derived Exos triggered autophagy to facilitate M2 macrophage polarisation [29]. In line with these findings, our data proved that exosomal MATN3 secreted by GC cells triggered M2 macrophage polarisation via autophagy induction. Our results explained the mechanism through which exosomal MATN3 affected macrophage polarisation.

MATN3 possesses four EGF domains [30], indicating its potential downstream mechanisms through the EGF pathway. A previous study found that MATN3 led to articular cartilage degeneration via transcriptional inhibition of SOX9 in an EGF‐dependent manner [31]. Another work indicated that EGF domains were responsible for MATN3‐induced SMAD1 inactivation via interaction with BMP2, which suppressed chondrocyte hypertrophy [17]. EGFR, the receptor of EGF, is recognised as an indicator of cancer aggressiveness and its overexpression has been verified in the pathogenesis of various cancers [32]. Blockage of the EGF/EGFR pathway could profoundly delay GC progression, which was considered an effective therapeutic approach [33]. So far, the influence of MATN3 on the EGF/EGFR pathway has not been explored. In this work, we found for the first time that MATN3 could directly interact with EGFR in macrophages and consequently stabilise EGFR via inhibiting its ubiquitination. More importantly, EGFR overexpression counteracted exosomal MATN3 deficiency‐induced M1 macrophage polarisation. Therefore, EGFR was involved in the downstream mechanism of exosomal MATN3 in macrophage polarisation.

ELK1, one of the key transcription factors, has been demonstrated as a downstream target of EGFR. For example, EGFR facilitated the phosphorylation of ELK1 to trigger GDH1 transcription, thereby promoting glutamine metabolism in glioblastoma [34]. Another study documented that EGFR could enhance the phosphorylation of ELK1 in human brain endothelial cells [22]. As reported by previous studies, ELK1 played key roles in M2 macrophage polarisation during the malignant progression of colorectal cancer [35], breast cancer [36] and prostate cancer [37]. Our data showed that EGFR overexpression led to the activation of ELK1 in macrophages. As an autophagy‐related protein, ATG12 is a key regulator of autophagosome formation [38]. ATG12 has been found to drive GC progression via promoting oncogenic autophagy [39]. In addition, high expression of ATG12 was closely associated with an advanced TNM stage, which represented as a new prognostic biomarker for GC [40]. In this work, we observed that ELK1 bound to the ATG12 promoter to trigger its transcription and expression in macrophages. Furthermore, the inhibitory effects of exosomal MATN3 depletion on autophagy and M2 macrophage polarizsation were partly abolished by ATG12 overexpression.

5. Conclusion

In summary, our study for the first time revealed that GC cells‐derived exosomal MATN3 accelerated GC progression. Mechanistically, exosomal MATN3 promoted M2 macrophage polarisation via stabilisation of EGFR protein and subsequent ELK1 activation, which in turn facilitated ATG12‐dependent autophagy. This work not only elucidates the oncogenic function of exosomal MATN3 in GC, but also identifies a promising therapeutic target for GC.

Author Contributions

Qianwen Zhao: conceptualization, methodology, investigation, data curation, software, writing – original draft, writing – review and editing. Shanshan Liu: investigation, data curation, resources, writing – original draft. Xin She: investigation, data curation, writing – original draft; Shiyue Ma: formal analysis, writing – original draft. Hui Tang: formal analysis, writing – original draft. Danli Peng: formal analysis, writing – original draft. Haonan Guo: formal analysis, validation, visualization, supervision, writing – original draft, writing – review and editing.

Funding

This work was supported by the Guangxi Natural Science Foundation (2025GXNSFHA069059), the Guangxi Science and Technology Base and Talent Special Project (Guike AD23026091), the Research Capability Improvement Project for Young and Middle‐aged Teachers in Guangxi Universities (2025KY0523), the Project of Guilin Scientific Research and Technology Development Plan (20220138z) and the Scientific Research Project of Guangxi Health Committee (Z‐C20220801).

Disclosure

Permission to Reproduce: Not applicable.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

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