MiR-668-3p in M2 macrophage-derived exosomes activates autophagy through the ETS1/EGFR axis and promotes cisplatin resistance in gastric cancer Authors: Zunqi Hu, Dapeng Xu, Zhi Zheng, Xin Huang, Zhenxin Zhu, Xin Zhang, Ziran Wei, Hongbing Fu, Weijun Wang, Dejun Yang, Ronglin Yan Published: 2025 ABSTRACT BackgroundPrevious research suggests that tumor-associated macrophages (TAMs) influence the cisplatin (DDP) tolerance of gastric cancer (GC) cells via the secretion of microRNA-containing exosomes. This study aims to investigate the role of exosomal miR-668-3p from M2 macrophages in modulating DDP resistance, using both in vitro and in vivo models to provide a comprehensive analysis.Materials and methodsThe expression profiles of DDP-resistant GC tissues were assessed through microarray, while immunofluorescence confirmed the uptake of these exosomes by GC cells. The role of miR-668-3p in regulating DDP resistance was explored using CCK8 assays, colony formation, EDU incorporation, and Western blotting. The interaction between miR-668-3p and ETS1 was validated through RIP and RNA pull-down assays. Furthermore, the regulatory role of the miR-668-3p/ETS1/EGFR axis in autophagy and DDP resistance was examined in GC cell lines and a tumor xenograft model.ResultsmiR-668-3p was significantly upregulated in DDP-resistant GC tissues. Exosomes originating from M2 macrophages transfer miR-668-3p to GC cells, enhancing their DDP resistance. Additionally, miR-668-3p was found to bind to ETS1 mRNA, leading to its suppression and a consequent decrease in EGFR expression. This reduction in EGFR expression was closely linked to the activation of autophagy, further augmenting DDP resistance in GC cells.ConclusionM2 macrophage-derived exosomal miR-668-3p promotes DDP resistance in GC cells by targeting the ETS1/EGFR axis, thereby activating the autophagy pathway. Future research should focus on developing targeted inhibition strategies for miR-668-3p to effectively reverse DDP resistance in GC cells, optimizing its potential for clinical application.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12935-025-04140-y. Introduction Gastric cancer (GC) ranks as the fifth most common malignancy globally, with over 968,000 new cases reported annually, accounting for 4.9% of all cancer diagnoses [1]. Notably, the incidence of GC exhibits significant geographic variability, with East Asia-particularly China and Japan-contributing to more than half of the global cases [1, 2]. Cisplatin (DDP), a platinum-based chemotherapeutic, remains a cornerstone for treating advanced GC, particularly in metastatic settings or as an adjuvant therapy following resection to prevent recurrence [3, 4]. However, the development of DDP resistance during disease progression poses a major clinical challenge, as it is associated with a poor prognosis and high relapse rates [5, 6]. Therefore, unraveling the mechanisms behind DDP resistance is essential for enhancing treatment outcomes and guiding the development of innovative therapeutic strategies. DDP resistance mechanisms are complex, encompassing enhanced drug efflux, increased DNA repair capacity, inhibition of apoptotic pathways, and modulation by the tumor microenvironment [6, 7]. Within this microenvironment, tumor-associated macrophages (TAMs), which typically adopt an M2-like phenotype, play a pivotal role in shaping cancer cell responses to therapy, often via exosome secretion [8, 9]. Exosomes, small vesicles facilitating intercellular communication, are enriched with microRNAs (miRNAs) that significantly influence gene expression and have been implicated in chemoresistance [10–13]. Autophagy, a cellular degradation process, exerts a dual role (tumor suppressor/promoter, depending on context) in gastric cancer (GC) and its resistance to chemotherapy like cisplatin (DDP), which is key for developing chemoresistance-overcoming strategies. Studies on CLIC1-mediated autophagy show that overexpressing intracellular chloride channel 1 (CLIC1) in GC cells boosts autophagy, reducing DDP sensitivity; inhibiting CLIC1 or autophagy increases DDP sensitivity, indicating CLIC1-induced autophagy promotes chemoresistance [14]. This aligns with autophagy aiding cancer cell survival under therapeutic stress. Supporting autophagy’s dual role, traditional medicine DSGOST exerts anticancer effects on GC via autophagy activation, but autophagy inhibition raises cell death (showing autophagy’s protective role) [15]. Autophagy-related gene-5 (ATG-5) is upregulated in chemoresistant GC (correlating with poor survival) and linked to enhanced autophagy, further implicating autophagy in chemoresistance [16]. Autophagy’s dual nature is also seen in modulating therapeutic response of GC cells: it acts as a tumor suppressor by maintaining cellular homeostasis and preventing oncogenic transformation, but is upregulated under stress (e.g., chemotherapy) to promote cancer cell survival and drug resistance [17, 18]. This complicates targeting autophagy in cancer therapy, though its inhibition may improve DDP efficacy. Autophagy’s dual role in GC and DDP resistance calls for a nuanced therapeutic approach. Understanding its context-specific functions can help develop targeted therapies to inhibit/modulate autophagy. Future research should identify molecular markers predicting autophagy’s protective role and explore new autophagy-modulating targets in GC 19–20. Substantial evidence indicates that exosomes from tumor-associated macrophages (TAMs) promote DDP resistance in gastric cancer (GC) cells through the delivery of specific miRNAs. For instance, miRNA-21, carried by M2 macrophage-derived exosomes, targets the PTEN gene, thereby activating the PI3K/Akt signaling pathway and enhancing DDP resistance in GC cells [21]. Similarly, exosomal miR-588 from M2 macrophages downregulates CYLD, inhibiting apoptosis while promoting cell proliferation and DDP resistance in GC cells [22]. Additionally, recent findings indicate that TAM-derived exosomes transmit miR-223, further influencing GC cell responses to chemotherapy, underscoring the multifaceted role of miRNAs in DDP resistance [23, 24]. Given these insights, understanding the regulatory mechanisms of miRNAs has become a key research priority, presenting a potential avenue for developing novel therapeutic strategies to overcome DDP resistance in GC. Hyperactivation of the epidermal growth factor receptor (EGFR), a receptor tyrosine kinase that links extracellular signals to cellular homeostasis, is known to initiate pathways contributing to tumor chemoresistance [25, 26]. Moreover, EGFR is a key regulator of several autophagy-related signaling pathways, which can either suppress or promote cancer depending on context [27, 28]. Given the established connection between autophagy activation and DDP resistance in GC cells, it is plausible that EGFR plays a role in DDP resistance by initiating the autophagy pathway. In this study, miR-668-3p was identified as a contributor to DDP resistance in GC, noting its overexpression in DDP-resistant GC tissues and presence in M2 macrophage-derived exosomes. Our results demonstrate that miR-668-3p promotes autophagy in GC cells, likely through the downregulation of EGFR expression. Additionally, ETS Proto-Oncogene 1 (ETS1) was identified as an target gene of miR-668-3p and an upstream transcription factor that regulates EGFR expression. This study highlights miR-668-3p as a potential therapeutic target for reversing DDP resistance by modulating the ETS1/EGFR axis and activating the autophagy pathway, providing valuable insights into overcoming chemoresistance in GC. Materials and methods Clinical sample collection and detection Clinical specimens, including DDP-sensitive GC tissue samples (N = 41) and DDP- resistance GC tissue samples (N = 41), were obtained from patients with GC at Changzheng Hospital, Naval Medical University. Written informed consent was obtained from each patient prior to sample collection, and all specimens were immediately preserved in liquid nitrogen post-surgery for subsequent analysis. This study was conducted in adherence to the Helsinki Declaration and received approval from the Ethics Committee of Changzheng Hospital, Naval Medical University. After surgery, patients were treated with standard DDP-based therapeutic regimens (FP regimen, ECF regimen and its derivative regimens, or DCF regimen). The choice of chemotherapy regimen was based on clinical guidelines and patient-specific factors (e.g., performance status, organ function) following standard clinical practice. Tumors were defined as DDP-resistant if recurrence occurred during or after adjuvant DDP-based chemotherapy following R0 excision. Conversely, tumors were defined as DDP-sensitive if no recurrence was observed during DDP-based therapy; these definitions were consistent with previous literatures [31, 32]. For the DDP-sensitive group, these patients had no evidence of tumor recurrence during the same rigorous follow-up schedule (clinical and imaging evaluations every 3 months for the first 2 years, then every 6 months) for a minimum follow-up period of 24 months after completing DDP-based adjuvant therapy. Primary tumor samples from patients in the sensitive group were also collected before chemotherapy, and no recurrence was observed after completing chemotherapy with a follow-up period of ≥ 24 months. Recurrence was confirmed during the follow-up period after R0 resection and adjuvant DDP-based chemotherapy. Patients were monitored every 3 months for the first 2 years, and every 6 months thereafter. Recurrence was defined as the reappearance of tumor lesions (local or metastatic) confirmed by imaging (CT/PET-CT) and/or biopsy during or after chemotherapy. Tissue samples were collected during initial surgical resection (before chemotherapy). For the DDP-resistant group, samples were from patients who later developed recurrence during adjuvant DDP-based chemotherapy. Primary tumor samples from patients in the resistant group were collected before chemotherapy, and these patients later developed recurrence during adjuvant DDP chemotherapy, confirming DDP resistance. The clinicopahological characteristics of the GC patients were summarized in Table 1. The DDP-sensitive and DDP-resistant groups were matched by clinicopathological features (e.g., age, stage, gender). There is no significance in age, gender, tumor stage, or differentiation status between DDP-sensitive and DDP-resistant groups. Table 1The clinicopathological characteristics of the GC patientsCharacteristicsPatients (%)DDP-resistanceDDP-sensitive(N = 41)(N = 41) Age (%) < 6023 (56.1)25 (60.9) ≥ 6018 (43.9)16 (39.1) Gender (%) Female20 (48.8)22 (53.7) Male21 (51.2)19 (46.3) Tumor size (cm) < 3.519 (46.3)21 (51.2) ≥ 3.522 (53.7)20 (48.8) Grade Well-moderately17 (41.5)18 (43.9) Poorly-signet24 (58.5)23 (56.1) Clinical stage II14 (34.1)17 (41.5) III27 (65.9)24 (58.5) T classification T1-T219 (46.3)13 (31.7) T3-T422 (53.7)28 (68.3) N classification N012 (29.3)11 (26.8) N1-N329 (70.7)30 (73.2) Table 1The clinicopathological characteristics of the GC patientsCharacteristicsPatients (%)DDP-resistanceDDP-sensitive(N = 41)(N = 41) Age (%) < 6023 (56.1)25 (60.9) ≥ 6018 (43.9)16 (39.1) Gender (%) Female20 (48.8)22 (53.7) Male21 (51.2)19 (46.3) Tumor size (cm) < 3.519 (46.3)21 (51.2) ≥ 3.522 (53.7)20 (48.8) Grade Well-moderately17 (41.5)18 (43.9) Poorly-signet24 (58.5)23 (56.1) Clinical stage II14 (34.1)17 (41.5) III27 (65.9)24 (58.5) T classification T1-T219 (46.3)13 (31.7) T3-T422 (53.7)28 (68.3) N classification N012 (29.3)11 (26.8) N1-N329 (70.7)30 (73.2) The clinicopathological characteristics of the GC patients Cell culture Human GC cell lines, including BGC823 and MKN45, were cultured in Roswell Park Memorial Institute (RPMI)−1640 medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich, St. Louis, MO, USA). Human embryonic kidney 293 T (HEK 293 T, ATCC, Manassas, VA, USA) cells were maintained in Dulbecco’s modified Eagle medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) containing 10% FBS. To establish DDP-resistant BGC823 and MKN45 cells, parental GC cells were exposed to gradually increasing concentrations of DDP (Sigma-Aldrich, St. Louis, MO, USA). In brief, BGC823 and MKN45 cells, during their exponential growth phase, were initially incubated in a medium containing 0.1 µg/ml of DDP. After 14 days of adaptation, the DDP concentration was incrementally increased to 0.2 µg/ml for an additional two weeks of culture. This stepwise DDP escalation continued to 0.4 µg/ml and finally 0.8 µg/ml, with each phase allowing the cells to stabilize and undergo repeated passages. Each concentration was maintained for 14 days. During this period, the drug-containing medium was replaced every 3 days until cells stably grew in 0.8 µg/ml DDP, yielding drug-resistant cell lines. The resulting DDP-resistant cell lines were designated BGC823/DDP and MKN45/DDP, respectively. All cell cultures were maintained at 37 °C in a 5% CO2 incubator (Thermo Fisher Scientific, Waltham, MA, USA), following standard laboratory practices. Cell transfection Cells were seeded in 6-well plates at a density of 1 × 10⁶ until reaching 70–80% confluence. Transfection was performed using Lipofectamine 2000 (Lipo2000, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Briefly, miRNA mimics, inhibitors, or their respective negative controls were diluted in an Opti-MEM medium (Thermo Fisher Scientific, Waltham, MA, USA) and combined with the transfection reagent. The mixture was incubated at room temperature for 15 min, after which it was added to the cells, followed by an additional 48 h of culture prior to downstream assays. Macrophages differentiation First, THP-1 cells (ATCC, Manassas, VA, USA) were cultured in RPMI 1640 medium supplemented with 10% FBS and 1% penicillin-streptomycin. They were maintained at 37 °C in a 5% CO₂ incubator. To obtain M0 macrophages, phorbol 12-myristate 13- acetate (PMA, Sigma-Aldrich, St. Louis, MO, USA) was added to the culture medium at a final concentration of 20 ng/mL. The cells were incubated for 48 h. During this period, THP-1 cells adhered to the culture vessel and were differentiated into M0 macrophages. To obtain M1 macrophages, after M0 macrophages were obtained, they were stimulated with interferon-γ (IFN-γ, Sigma-Aldrich, St. Louis, MO, USA) at a concentration of 20 ng/mL and lipopolysaccharide (LPS, Sigma-Aldrich, St. Louis, MO, USA) at a concentration of 5 µg/mL. The cells were incubated for 48 h. Under the action of IFN-γ and LPS, M0 macrophages were polarized into M1 macrophages. M1 macrophage polarization was confirmed by detecting the expression of CD86, inducible nitric oxide synthase, and tumor necrosis factor-α (TNF-α). To obtain M2 macrophages, interleukin-4 (IL-4, Sigma-Aldrich, St. Louis, MO, USA) and interleukin-13 (IL-13, Sigma-Aldrich, St. Louis, MO, USA) were used. IL-4 and IL-13 were added to the culture medium of M0 macrophages at concentrations of 25 ng/mL each. The cells were incubated for 48 h. The polarization of M2 macrophages was confirmed by detecting CD206 and IL-10. Microarray analysis Total RNA was isolated from cells using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) per the manufacturer’s guidelines. RNA purity (A260/A280 > 1.9, A260/A230 > 2.0) was assessed via NanoDrop spectrophotometer (Thermo Fisher), and integrity (RIN > 8.0) verified with Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA); only qualified samples proceeded. For cDNA synthesis, 100 ng total RNA/sample was reverse transcribed (cDNA synthesis kit, Thermo Fisher). Another 100 ng RNA/sample was dephosphorylated, Cy3-labeled (Agilent miRNA Complete Labeling and Hyb Kit), with labeling efficiency quantified by NanoDrop (> 8.0 pmol Cy3/µl). Labeled samples were hybridized to arrays (55 °C, 20 h, Agilent Hybridization Oven), then washed (Agilent buffers: Buffer 1 at RT, Buffer 2 at 37 °C) to remove non-specific binding. Hybridized arrays were scanned with Agilent G2600D SureScan Scanner (3 μm resolution, 100% PMT gain) and Roche high-resolution scanner. TIFF images were processed via Agilent Feature Extraction Software (v10.7.3.1) for signal quantification. Raw data were preprocessed/analyzed with R (v4.2.1, limma). Unreliable/low-signal probes were filtered out. miRNA arrays were normalized, and differential expression analyzed via empirical Bayes moderated t-test (limma). P-values were adjusted for multiple testing using the Benjamini-Hochberg method to control the False Discovery Rate (FDR). miRNAs with an FDR < 0.05 and |log2FC| > 1 were considered significantly differentially expressed. Real-time quantitative (RT)-qPCR Real-time quantitative PCR (qPCR) was performed using SYBR Green qPCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) and gene-specific primers on a CFX96 real-time PCR detection system (Bio-Rad, USA). All the primer sequences are listed in Supplementary Table S1. The primers were validated by using melting curve analysis to confirm a single amplification product and agarose gel electrophoresis to check for the correct band size. Next, a standard curve was constructed to accurately calculate the primers’ amplification efficiency, ensuring it is between 90% and 110% with good linearity (R² > 0.99). Each reaction mixture contained SYBR Green Master Mix, forward and reverse primers, and a cDNA template, with volumes adjusted according to the manufacturer’s instructions. The thermal cycling conditions were optimized and typically included an initial denaturation step followed by amplification cycles. The fluorescence intensity of SYBR Green, bound to double-stranded DNA, was measured to monitor amplification. Target gene expression levels were quantified using the ΔΔCt method, with GAPDH as the internal control for mRNA and U6 snRNA for miRNA normalization. Cytotoxicity assay and autophagy Inhibition Cells were seeded into 96-well plates at a density of 5,000 cells per well, 24 h prior to treatment. The next day, cells were exposed to a range of DDP concentrations (0.05–8 µg/ml) for 48 h to evaluate dose-dependent effects. After treatment, 10 µl of Cell Counting Kit-8 (CCK-8, Sigma-Aldrich, St. Louis, MO, USA) solution was added to each well and incubated for 2 h at 37 °C. Absorbance at 450 nm, correlating with cell viability, was measured using a microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). The viability data were analyzed using GraphPad Prism software (Version 9.0), with non-linear regression analysis applied to generate dose-response curves and determine half-maximal inhibitory concentration (IC50) values. To investigate the necessity of autophagy in DDP resistance, cells were pre-treated for 2 h with the autophagy inhibitor Chloroquine (CQ, Sigma-Aldrich; 20 µmol/L) or vehicle control (PBS). Following pre-treatment, the cells were co-exposed to CQ and either M2 macrophage-derived exosomes (M2-EXO, 20 µg/mL) or transfected with miR-668-3p mimic (50 nM), in the presence of 1.0 µmol/L DDP for 48 h. Cell viability was then assessed using the CCK-8 assay as described above. Colony formation assay After a 12-hour exposure to 1.0 µmol/L DDP, cells were harvested via trypsinization and seeded into six-well plates at a density of 300 cells per well. The cells were then cultured for an additional two weeks. Colony visualization was achieved by fixing the cells in methanol for 15 min at room temperature. Following fixation, the colonies were stained with crystal violet, which selectively binds to DNA, rendering the colonies visible. The stained plates were gently rinsed with distilled water to remove excess dye and then air-dried. Colony counts were performed using ImageJ (NIH). 5-Ethynyl-2′-deoxyuridine (EdU) incorporation assay For the EdU proliferation assay, cells were incubated with 5-ethynyl-2′-deoxyuridine (EdU, Thermo Fisher Scientific, Waltham, MA, USA) for 2 h, followed by fixation with 4% paraformaldehyde for 15 min at room temperature. Cells were then incubated with Apollo Dye Solution (Thermo Fisher Scientific, Waltham, MA, USA), which binds specifically to incorporated EdU, allowing for the detection of proliferating cells. The staining protocol was carried out according to the manufacturer’s guidelines to ensure optimal labeling. Hoechst 33,342 (Thermo Fisher Scientific, Waltham, MA, USA) was subsequently applied to stain the nuclei, facilitating the identification and differentiation of individual cells. Proliferating cells, indicated by EdU-Apollo staining, were visualized using a fluorescence microscope, and images were captured for quantification of proliferating cells using image analysis software. Western blotting assay Proteins were isolated from cells using radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, Waltham, MA, USA) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, Bio-Rad, Shanghai, China). The proteins were then transferred to a polyvinylidene difluoride (PVDF, Bio-Rad, Shanghai, China) membrane using a wet transfer system. To block non-specific binding, the membranes were incubated with 5% skimmed milk in Tris-buffered saline with Tween 20 (TBST, Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature. After blocking, the membranes were incubated overnight at 4 °C with appropriately diluted primary antibodies (Sigma-Aldrich, St. Louis, MO, USA), allowing specific binding to the target proteins. Following washes with TBST, the membranes were treated with HRP-conjugated secondary antibodies (Sigma-Aldrich, St. Louis, MO, USA) for 2 h at room temperature. Protein-antibody complexes were visualized using a chemiluminescent detection system (Bio-Rad, Shanghai, China) to detect HRP activity. Quantification and analysis of the protein bands were conducted using ImageJ software. Full, uncropped Western blot images for all key figures are provided in Supplementary Figure S1. Exosome isolation and detection Exosomes were purified from culture supernatants through differential ultracentrifugation. Initially, cell debris and larger vesicles were removed by low-speed centrifugation (500 g, 10 min, 4℃). The supernatant was then centrifuged at 10,000 g for 30 min to pellet remaining large vesicles and apoptotic bodies. Next, the cleared supernatant underwent ultracentrifugation at 100,000 g for 70 min to precipitate exosomes. The exosome pellet was washed in phosphate-buffered saline (PBS, Sigma-Aldrich, St. Louis, MO, USA) and subjected to a second round of ultracentrifugation under the same conditions. The final pellet was reconstituted in PBS. To confirm exosome morphology, transmission electron microscopy (TEM, Thermo Fisher Scientific, Waltham, MA, USA) was used. Small volumes of exosome suspensions were placed on carbon-coated copper grids, stained with uranyl acetate, and air-dried for imaging, which provided detailed visual confirmation of exosome morphology. Nanoparticle tracking analysis (NTA) was conducted to assess exosome size distribution and concentration. The exosome suspension was diluted in PBS, and the NTA instrument (Eppendorf, Hamburg, Germany) tracked the Brownian motion of individual particles using a laser beam, enabling the calculation of size distribution and particle concentration. The detection of PKH67-labeled exosomes in cells was performed using the PKH67 Green Fluorescent Cell Linker Kit (PKH67GL, Sigma-Aldrich, St. Louis, MO, USA) as follows. Briefly, diluent C and PKH67 dye was first added to the extracted exosomes at 37 °C for 5 min before the staining was terminated using bovine serum albumin. Then the exosomes were repurified and incubated with GC cells for 24 h. DAPI were used to stain the nucleus of GC cells. After thorough washing with complete medium to remove unbound dye, the stained cells were fixed, mounted, and visualized under a fluorescence microscope. Quantitative analysis of fluorescence intensity was conducted to evaluate dye uptake and retention, with unstained cells serving as a negative control to establish the background signal (Supplementary Figure S2). Immunofluorescence (IF) Cells were cultured on sterilized coverslips and allowed to adhere for 12–16 h. They were then fixed with 4% paraformaldehyde for 15 min at room temperature and permeabilized with 0.1% Triton X-100 in PBS for 10 min. To block non-specific binding, the cells were incubated with 5% bovine serum albumin (BSA, Thermo Fisher Scientific, Waltham, MA, USA) in PBS (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature. Primary antibodies (Sigma-Aldrich, St. Louis, MO, USA) diluted in the blocking solution were applied, and the cells were incubated overnight at 4 °C. The following day, the coverslips were thoroughly washed with PBST (Sigma-Aldrich, St. Louis, MO, USA) and incubated with fluorescently labeled secondary antibodies (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature in the dark. After washing off excess secondary antibodies, nuclei were counterstained with DAPI (Thermo Fisher Scientific, Waltham, MA, USA). Coverslips were mounted on glass slides with an antifade mounting medium, and images were captured using fluorescence microscopy (Olympus, Tokyo, Japan). ImageJ software was used for subsequent fluorescence intensity quantification. Immunohistochemistry (IHC) All tissue samples used for IHC analysis were initially reviewed and diagnosed independently by two experienced pathologists. Paraffin-embedded tissue sections were first deparaffinized by immersion in xylene (Sigma-Aldrich, St. Louis, MO, USA), followed by graded rehydration using 100% to 70% ethanol (Sigma-Aldrich, St. Louis, MO, USA). Antigen retrieval was performed by heating the sections in citrate buffer (pH 6.0, Sigma-Aldrich, St. Louis, MO, USA) using a microwave. The sections were then treated with 3% hydrogen peroxide (Sigma-Aldrich, St. Louis, MO, USA) for 10 min at room temperature to block endogenous peroxidase activity. Afterward, the sections were blocked with 5% serum (from the same source as the secondary antibody) for 30 min at room temperature. Primary antibodies (Sigma-Aldrich, St. Louis, MO, USA) diluted in the blocking solution were applied, and the slides were incubated overnight at 4 °C. The following day, sections were thoroughly washed and incubated with HRP-conjugated secondary antibodies (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature. Horseradish peroxidase (HRP) activity was visualized using a DAB (3,3’-diaminobenzidine) substrate (Thermo Fisher Scientific, Waltham, MA, USA). After staining, the sections were counterstained with hematoxylin to visualize the nuclei, providing contrast to the DAB staining. The slides were dehydrated through graded ethanol, cleared in xylene, and mounted with a coverslip. Stained slides were examined under a light microscope (Olympus, Tokyo, Japan), and staining intensity and positive staining area were quantified using ImageJ software. Luciferase reporter assay Cells were co-transfected with the pmirGLO dual-luciferase reporter plasmid (Promega, WI, USA), containing either the wild-type or mutant 3’UTR of the ETS1 gene, along with miR-668-3p mimic or inhibitor using Lipo 2000. After a 48 h incubation to allow for gene expression, the cells were lysed, and luciferase activity was measured using the dual-luciferase reporter assay system (Promega, WI, USA). This system quantifies both firefly luciferase (encoded by the pmirGLO vector) and Renilla luciferase (co-expressed as an internal control). Luminescence signals were detected using a microplate luminometer, and the results were expressed as relative luciferase activity (firefly/Renilla ratio). This setup enabled the evaluation of miR-668-3p’s impact on the post-transcriptional regulation of ETS1 via interaction with its 3’UTR. RNA Immunoprecipitation (RIP) Cells were lysed in RIP lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with protease and RNase inhibitors (Thermo Fisher Scientific, Waltham, MA, USA) to maintain protein and RNA integrity. The lysates were incubated overnight at 4 °C with magnetic beads conjugated to an AGO2-specific antibody (Thermo Fisher Scientific, Waltham, MA, USA), along with a control IgG antibody (Thermo Fisher Scientific, Waltham, MA, USA), to facilitate RNA-protein complex immunoprecipitation. After incubation, the beads were thoroughly washed to remove non-specifically bound material. RNA co-precipitated with AGO2 or IgG was extracted using TRIzol reagent and analyzed by qPCR to assess the enrichment of specific target RNAs compared to input controls. RNA pull-down assay Biotinylated miRNA probes were synthesized via in vitro transcription, and DDP-resistant GC cells were lysed to release their contents. The lysates were incubated with the biotin-labeled RNA probes (Thermo Fisher Scientific, Waltham, MA, USA) at 4 °C for 12 h to allow RNA-protein complex formation. Following this incubation, streptavidin-coated magnetic beads (Thermo Fisher Scientific, Waltham, MA, USA) were added to capture the biotin-labeled RNA-protein complexes. After extensive washing to remove non-specifically bound material, the associated RNAs were extracted using a TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). The purified RNAs were analyzed by qPCR to quantify and confirm interactions with the target RNA, with comparisons to control samples to validate interaction specificity. Autophagy flux assay The autophagy flux assay was performed using mCherry-GFP-LC3 reporter system (catalog no. C3011, Beyotime, Shanghai, China) and Bafilomycin A1 (catalog no. 5.08409, Sigma-Aldrich, St. Louis, MO, USA) treatment. Briefly, Cells were plated on coverslips and transfected with the plasmid using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA) for 6 h. Following this transfection period, the cells were subjected to either control conditions or TRF intervention for 48 h. Coverslips were processed through sequential steps of fixation with paraformaldehyde and permeabilization with Triton X-100 (No. P0096; Beyotime, Shanghai, China). Nuclear staining was performed using Hoechst 33,342 (No. H3570; Thermo Fisher Scientific, Waltham, MA, USA). After mounting with an antifade medium (ProLong Diamond Antifade Mountant, catalog no. P36970; Thermo Fisher Scientific, Waltham, MA, USA), the stained samples were visualized using a confocal laser scanning microscope (Nikon A1R HD25; Nikon, Tokyo, Japan). Tumor xenograft experiment Female and male BALB/c nude mice (five mice/group) were subcutaneously injected with either unmodified BGC823 cells or BGC823 cells overexpressing ETS1 (ETS1-OE) or EGFR (EGFR-OE). Equal numbers of male and female mice were randomly assigned to ensure consistent gender ratios in each group. Once the tumors reached approximately 100 mm3, the mice were randomly divided into six groups. Each group received intraperitoneal injections of DDP (20 mg/kg) twice a week and intertumoral injections of 10 µg exosomes derived from M2 macrophages every two days. The exosomes were either untreated, EXO-depleted, or miR-668-3p-depleted. Tumor sizes were measured weekly, and at week 5, the mice were euthanized by CO2 asphyxiation. Tumors were excised and weighed. The detailed treatment schedule was as follows: Subcutaneously inject all mice with their respective cell lines (Unmodified, ETS1-OE, or EGFR-OE). Measure and record initial body weights. Days 1-~14: Monitor mice daily for health and welfare. Measure tumor dimensions (length and width) with calipers 2–3 times per week. Calculate tumor volume using the formula: Volume (mm³) = (Length × Width²)/2. Once tumors reach ~ 100 mm³: Record this date for each mouse. Randomly assign mice within each cell-line cohort to the three exosome treatment groups. DDP (20 mg/kg, i.p.) was administered twice a week. Exosomes (10 µg, intertumoral) was administered every two days. This results in 3–4 injections per week, depending on the week. Caliper measurements and calculate volume was performed at least twice weekly coinciding with exosome injections. Record mouse weights at the same time. On Day 35 (or if a tumor exceeds a predefined ethical size limit, e.g., 1000 mm³, earlier), the mice were euthanized by CO2 asphyxiation. Final tumor measurement and mouse weighing was performed. Tumors were excised and weighed individually. Tumors were used for further analysis (snap-freezing in liquid nitrogen for qT-PCR analysis or fixation for IHC analysis). This study adhered to the Laboratory Animal Guideline for Ethical Review of Animal Welfare of the People’s Republic of China and received approval from the Animal Ethics Committee of Changzheng Hospital, Naval Medical University. Statistical analysis Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism software. Comparisons between two groups were made using the Student’s t-test. For experiments designed to test a pre-specified, limited number of hypotheses (e.g., qPCR validation of candidate miRNAs or proteins from Western blot), comparisons between two groups were made using a two-tailed Student’s t-test without multiple comparison correction, as these were confirmatory analyses guided by prior results from the microarray or preliminary data. For comparisons across more than two groups, one-way ANOVA followed by Tukey’s post-hoc test was applied. Data are presented as mean ± standard deviation (SD). Survival curves were plotted using the Kaplan-Meier method and analyzed with the log-rank test. The correlation between miR-668-3p, ETS1, and EGFR expression levels was assessed through linear regression analysis. A p-value of less than 0.05 was considered statistically significant for all analyses unless otherwise specified (e.g., the FDR threshold for microarray). The sample size (n) for each experiment is provided in the corresponding figure legend. Results Overexpression of miR-668-3p in DDP-resistant GC tissues significantly reduces patient survival Given the mounting evidence implicating miRNAs in GC DDP resistance [33], a comprehensive compilation of miRNAs associated with the survival of patients with GC was extracted from the UALCAN database (https://ualcan.path.uab.edu/). Subsequently, miRNA levels from this list were examined via microarray in five DDP-sensitive and five DDP-resistant GC tissues. The analysis revealed a pronounced upregulation of miR-668-3p in the resistant tissues (Fig. 1A). Additionally, five highly expressed miRNAs in resistant tissues underwent further validation by qPCR across two groups of GC tissues, confirming elevated miR-668-3p levels in the resistant samples (Fig. 1B). qPCR analysis of miR-668-3p expression in 41 DDP-sensitive and 41 DDP-resistant GC tissues reaffirmed its significant overexpression in resistant tissues (Fig. 1C). Moreover, UALCAN database analysis demonstrated that elevated miR-668-3p expression was linked to a markedly reduced survival probability in patients with GC compared to those with low or medium expression (Fig. 1D). Collectively, these results indicate that miR-668-3p overexpression in DDP-resistant GC tissues contributes significantly to reduced patient survival. Fig. 1Overexpression of miR-668-3p in DDP-resistant GC tissues significantly reduces patient survival. (A) Heatmap illustrating the microarray data of miRNAs associated with patient survival in DDP-sensitive and DDP-resistant GC tissues (n = 5). (B) Relative levels of five miRNAs in DDP-sensitive and DDP-resistant GC tissues (n = 5, *p < 0.05 between indicated groups; **p < 0.01 between indicated groups). (C) qPCR verification of relative miR-668-3p levels in DDP-sensitive and DDP-resistant GC tissues (n = 41, **p < 0.01 between indicated groups). (D) Kaplan-Meier survival analysis of patients with high (n = 96) and low/medium (n = 285) miR-668-3P expression from the UALCAN database (p = 0.0047) Overexpression of miR-668-3p in DDP-resistant GC tissues significantly reduces patient survival. (A) Heatmap illustrating the microarray data of miRNAs associated with patient survival in DDP-sensitive and DDP-resistant GC tissues (n = 5). (B) Relative levels of five miRNAs in DDP-sensitive and DDP-resistant GC tissues (n = 5, *p < 0.05 between indicated groups; **p < 0.01 between indicated groups). (C) qPCR verification of relative miR-668-3p levels in DDP-sensitive and DDP-resistant GC tissues (n = 41, **p < 0.01 between indicated groups). (D) Kaplan-Meier survival analysis of patients with high (n = 96) and low/medium (n = 285) miR-668-3P expression from the UALCAN database (p = 0.0047) MiR-668-3p enhances DDP resistance in GC cells To assess the cytotoxic effects of varying DDP concentrations on two standard GC cell lines (BGC823 and MKN45) and their DDP-resistant counterparts (BGC823/DDP and MKN45/DDP), a CCK8 assay was conducted. The results showed that both cell viability and IC50 values were significantly higher in the resistant lines than in the parent cells (Fig. 2A-B). Additionally, qPCR confirmed a marked increase in miR-668-3p expression in resistant GC cell lines compared to their non-resistant counterparts (Fig. 2C). Fig. 2MiR-668-3p enhances DDP resistance in GC cells. (A) Cell viability and (B) IC50 values of normal cell lines (BGC823, MKN45) and DDP-resistant cell lines (BGC823/DDP, MKN45/DDP) across gradient concentrations of DDP, as measured by the CCK8 assay. (C) qPCR analysis of miR-668-3p expression in BGC823, MKN45, BGC823/DDP, and MKN45/DDP cells. (D, F) Cell viability and (E, G) IC50 values of BGC823 and MKN45 cells treated with miR-668-3p inhibitor or mimic in gradient DDP concentrations. (H) Western blot analysis of autophagy-related proteins in BGC823 cells treated with miR-668-3p inhibitor or mimic in 1.0 µmol/L DDP for 12 h. Colony formation (I-K) and Edu incorporation (L-M) assays for BGC823 and MKN45 cells treated with miR-668-3p inhibitor or mimic in 1.0 µmol/L DDP for 12 h. ** p < 0.01between indicated groups MiR-668-3p enhances DDP resistance in GC cells. (A) Cell viability and (B) IC50 values of normal cell lines (BGC823, MKN45) and DDP-resistant cell lines (BGC823/DDP, MKN45/DDP) across gradient concentrations of DDP, as measured by the CCK8 assay. (C) qPCR analysis of miR-668-3p expression in BGC823, MKN45, BGC823/DDP, and MKN45/DDP cells. (D, F) Cell viability and (E, G) IC50 values of BGC823 and MKN45 cells treated with miR-668-3p inhibitor or mimic in gradient DDP concentrations. (H) Western blot analysis of autophagy-related proteins in BGC823 cells treated with miR-668-3p inhibitor or mimic in 1.0 µmol/L DDP for 12 h. Colony formation (I-K) and Edu incorporation (L-M) assays for BGC823 and MKN45 cells treated with miR-668-3p inhibitor or mimic in 1.0 µmol/L DDP for 12 h. ** p < 0.01between indicated groups To further elucidate the role of miR-668-3p in DDP resistance, BGC823 and MKN45 cells were treated with either a miR-668-3p inhibitor or a mimic, followed by DDP exposure at varying concentrations. Inhibition of miR-668-3p notably enhanced DDP sensitivity, while its overexpression reinforced resistance (Fig. 2D-G). Western blot analysis of autophagy-related proteins suggested that miR-668-3p promotes autophagy activation in BGC823 cells treated with DDP (Fig. 2H). Additionally, colony formation and Edu incorporation assays revealed that miR-668-3p significantly increased both the number of colonies (Fig. 2I-K) and DNA replication (Fig. 2L-M) in GC cells exposed to DDP. Together, these results underscore the role of miR-668-3p in enhancing DDP resistance in GC cells. M2 macrophage-derived exosomes promote DDP resistance in GC The role of exosomes as pivotal mediators of communication between cancer cells is gaining increasing attention. These vesicles carry specific miRNAs, potentially contributing to the development of DDP resistance in GC 14–17. TEM analysis of venous blood samples from patients with both normal and DDP-resistant GC confirmed the presence of exosomes exhibiting characteristic morphology and size in both groups (Fig. 3A). In a cohort of 35 patients with GC, the proportions of M0, M1, and M2 macrophages in venous blood were quantified, and patients were categorized into low and high groups based on these levels. Kaplan-Meier survival analysis revealed a significant association between elevated M2 macrophage levels and decreased patient survival, while no significant correlation was found for M0 and M1 macrophages (Fig. 3B-D). Fig. 3M2 macrophage-derived exosomes promote DDP resistance in GC. (A) TEM images and particle size measurements of exosomes from the venous blood of normal individuals and DDP-resistant patients. Kaplan-Meier survival analysis correlating the numbers of (B) M0, (C) M1, and (D) M2 macrophages with overall survival in 35 patients with GC (**p < 0.01 between indicated groups). (E) Workflow for culturing GC cells in macrophage-derived medium. (F) Protein levels of exosomal markers in MDMs, M0-CM, M1-CM, and M2-CM. (G) TEM images of exosomes in M0-CM, M1-CM, and M2-CM. (H) PKH67 levels in BGC823 cells after 6 h of treatment with M0-CM, M1-CM, and M2-CM, observed via immunofluorescence. Scale bar = 25 μm. TEM: transmission electron microscopy M2 macrophage-derived exosomes promote DDP resistance in GC. (A) TEM images and particle size measurements of exosomes from the venous blood of normal individuals and DDP-resistant patients. Kaplan-Meier survival analysis correlating the numbers of (B) M0, (C) M1, and (D) M2 macrophages with overall survival in 35 patients with GC (**p < 0.01 between indicated groups). (E) Workflow for culturing GC cells in macrophage-derived medium. (F) Protein levels of exosomal markers in MDMs, M0-CM, M1-CM, and M2-CM. (G) TEM images of exosomes in M0-CM, M1-CM, and M2-CM. (H) PKH67 levels in BGC823 cells after 6 h of treatment with M0-CM, M1-CM, and M2-CM, observed via immunofluorescence. Scale bar = 25 μm. TEM: transmission electron microscopy GC cells were then cultured in media conditioned by macrophages differentiated from PMA-treated THP-1 cells (Figure E). Analysis of exosomal markers in culture media from monocyte-derived macrophages (MDMs), including M0, M1, and M2 macrophages (M0-CM, M1-CM, and M2-CM), revealed higher exosomal protein levels in M2-CM compared to the other macrophage-conditioned media (Figure F). Similarly, TEM analysis demonstrated a greater prevalence of exosomes in M2-CM than in M0-CM or M1-CM (Figure G). Moreover, immunofluorescence detection of PKH67, an exosomal marker, in BGC823 cells treated for six hours with M0, M1, or M2 macrophage-conditioned media showed a significant increase in PKH67-positive signals in cells treated with M2-CM, indicating efficient internalization of M2 macrophage-derived exosomes by GC cells (Figure H). These findings suggest that M2 macrophage-derived exosomes are critical for enhancing DDP resistance in GC. M2 macrophage-derived exosomes enhance DDP resistance in GC associated with miR-668-3p To explore the impact of M2 macrophage-derived exosomes on DDP resistance in GC, BGC823 and MKN45 cells were treated with M2-CM, exosomes isolated from M2-CM (M2-EXO), and M2-CM depleted of exosomes (M2-CM Depleted of EXO). The viability and IC50 values of GC cells under these conditions were evaluated across varying DDP concentrations. The data revealed that both cell viability and IC50 values were significantly elevated in cells treated with M2-CM and M2-EXO compared to the control group, whereas no significant difference was observed between the M2-CM Depleted of EXO and control groups (Fig. 4A-D). Next, the number of cell clones (Fig. 4E-G) and the proportion of cells in the synthesis phase (Fig. 4H-I) were assessed after 12 h of DDP exposure across the three treatment conditions. The results confirmed that both metrics were markedly higher in the M2-CM and M2-EXO groups than in the control and M2-CM Depleted of EXO groups. Fig. 4M2 macrophage-derived exosomes promote DDP resistance in GC through miR-668-3p. Cell viability (A, C) and IC50 values (B, D) of BGC823 and MKN45 cells exposed to M2-CM, M2-EXO, and M2-CM without EXO at varying concentrations of DDP. The proliferation of BGC823 and MKN45 cells treated with M2-CM, M2-EXO, and M2-CM without EXO in 1.0 µmol/L DDP for 12 h was assessed by colony formation (E-G) and Edu incorporation assays (H-I). (J) The relative expression of miR-668-3p in BGC823 cells following treatment with M2-CM, M2-EXO, and M2-CM without EXO. Cell viability (K-L) and colony formation (M-N) of BGC823 and MKN45 cells treated with M2-EXO or M2-EXO combined with a miR-668-3p inhibitor in the presence of DDP. (O) The mRNA levels of miR-668-3p in M0 macrophages, M1 macrophages and M2 macrophages. ** p < 0.01 between indicated groups; n.s. not significant between indicated groups M2 macrophage-derived exosomes promote DDP resistance in GC through miR-668-3p. Cell viability (A, C) and IC50 values (B, D) of BGC823 and MKN45 cells exposed to M2-CM, M2-EXO, and M2-CM without EXO at varying concentrations of DDP. The proliferation of BGC823 and MKN45 cells treated with M2-CM, M2-EXO, and M2-CM without EXO in 1.0 µmol/L DDP for 12 h was assessed by colony formation (E-G) and Edu incorporation assays (H-I). (J) The relative expression of miR-668-3p in BGC823 cells following treatment with M2-CM, M2-EXO, and M2-CM without EXO. Cell viability (K-L) and colony formation (M-N) of BGC823 and MKN45 cells treated with M2-EXO or M2-EXO combined with a miR-668-3p inhibitor in the presence of DDP. (O) The mRNA levels of miR-668-3p in M0 macrophages, M1 macrophages and M2 macrophages. ** p < 0.01 between indicated groups; n.s. not significant between indicated groups Additionally, miR-668-3p levels in BGC823 cells were measured following the three treatments, showing a significant increase in miR-668-3p expression in the M2-CM and M2-EXO groups compared to the control and M2-CM Depleted of EXO groups (Fig. 4J). To further investigate the role of miR-668-3p, BGC823 and MKN45 cells were treated with M2-EXO in combination with a miR-668-3p inhibitor. Upon subsequent DDP treatment, both cell viability (Fig. 4K-L) and clone formation (Fig. 4M-N) were significantly reduced in the co-treated group, reaching levels comparable to the control group. This evidence indicates that inhibiting miR-668-3p effectively reversed the DDP resistance induced by M2-EXO. These results collectively demonstrate that M2 macrophage-derived exosomes enhance DDP resistance in GC cells, primarily through the upregulation of miR-668-3p. The role of miR-668-3p, the transcription factor ETS1, and the downstream Targe gene EGFR in GC development Previous studies have underscored the critical role of EGFR-mediated autophagy in driving therapeutic resistance [19, 20]. In our investigation, EGFR mRNA expression was significantly lower in DDP-resistant GC tissues compared to DDP-sensitive tissues (Fig. 5A). Receiver operating characteristic (ROC) analysis yielded an area under the curve (AUC) of 0.9209 (p < 0.0001), demonstrating a high level of accuracy in distinguishing DDP-resistant GC tissues based on EGFR expression (Fig. 5B). Moreover, stratification of 35 patients with GC into low and high EGFR expression groups revealed that diminished EGFR mRNA levels were strongly associated with poorer survival outcomes (Fig. 5C). A negative correlation was also identified between EGFR mRNA and miR-668-3p levels (p < 0.001, Fig. 5D). Fig. 5The role of miR-668-3p, the transcription factor ETS1 and the downstream targe gene EGFR in GC development. (A) The relative mRNA level of EGFR in DDP-sensitive and DDP-resistant GC tissues (n = 41). (B) The ROC analysis to evaluate the diagnostic value of EGRF in DDP resistance (AUC = 0.9209, P < 0.0001). (C) Kaplan-Meier analysis of the correlation between EGFR expression and overall survival in 35 GC patients. (D) The negative correlation of EGFR expression and miR-668-3p level (p < 0.0001). (E) The intersection of proteins with binding sites for miR-668-3p and TFs potentially regulating EGFR expression. (F) Heatmap of microarray data of the intersection in DDP-sensitive and DDP-resistant GC tissues (n = 5). (G) The relative mRNA level of ETS1 in DDP-sensitive and DDP-resistant GC tissues (n = 41). (H) The ROC analysis to evaluate the diagnostic value of ETS1 in DDP resistance (AUC = 0.9358, P < 0.0001). (I) The positive correlation of EGFR expression and ETS1 level (p < 0.0001). (J) The negative correlation of ETS1 expression and miR-668-3p level (p < 0.0001). (K) The motif sequence for ETS1 obtained from the JASPAR database. (L-M) The expression of EGFR and ETS1 in DDP-sensitive and DDP-resistant GC tissues observed by immunohistochemistry. ** p < 0.01 between indicated groups The role of miR-668-3p, the transcription factor ETS1 and the downstream targe gene EGFR in GC development. (A) The relative mRNA level of EGFR in DDP-sensitive and DDP-resistant GC tissues (n = 41). (B) The ROC analysis to evaluate the diagnostic value of EGRF in DDP resistance (AUC = 0.9209, P < 0.0001). (C) Kaplan-Meier analysis of the correlation between EGFR expression and overall survival in 35 GC patients. (D) The negative correlation of EGFR expression and miR-668-3p level (p < 0.0001). (E) The intersection of proteins with binding sites for miR-668-3p and TFs potentially regulating EGFR expression. (F) Heatmap of microarray data of the intersection in DDP-sensitive and DDP-resistant GC tissues (n = 5). (G) The relative mRNA level of ETS1 in DDP-sensitive and DDP-resistant GC tissues (n = 41). (H) The ROC analysis to evaluate the diagnostic value of ETS1 in DDP resistance (AUC = 0.9358, P < 0.0001). (I) The positive correlation of EGFR expression and ETS1 level (p < 0.0001). (J) The negative correlation of ETS1 expression and miR-668-3p level (p < 0.0001). (K) The motif sequence for ETS1 obtained from the JASPAR database. (L-M) The expression of EGFR and ETS1 in DDP-sensitive and DDP-resistant GC tissues observed by immunohistochemistry. ** p < 0.01 between indicated groups To explore potential transcription factors regulating EGFR via miR-668-3p, the ENCORI (https://rnasysu.com/encori/) and JASPAR (https://jaspar.elixir.no/) databases were employed. These analyses screened 1035 proteins with miR-668-3p binding sites and identified 189 possible transcription factors that could regulate EGFR expression. The intersection of these datasets yielded 12 candidate proteins (Fig. 5E). Among them, ETS1 mRNA levels were markedly lower in DDP-resistant GC tissues, as confirmed by microarray and further validated by qPCR (Fig. 5F-G). ROC analysis of ETS1 expression produced an AUC of 0.9358 (p < 0.0001), indicating its robust capacity to differentiate DDP-resistant GC tissues (Fig. 5H). Furthermore, ETS1 expression was positively correlated with EGFR and inversely correlated with miR-668-3p in the venous blood of 35 patients with GC (p < 0.001, Fig. 5I-J). The JASPAR database provided the ETS1 motif sequence, which could predict potential DNA binding sites for this transcription factor (Fig. 5K). Immunohistochemistry further confirmed a significant reduction in both EGFR and ETS1 expression in DDP-resistant GC tissues (Fig. 5L-M). These results suggest that reduced expression of EGFR and ETS1, potentially regulated by miR-668-3p, plays a critical role in DDP resistance in GC, identifying potential therapeutic targets to counteract this resistance. Validation of ETS1 expression regulation by miR-668-3p To investigate the regulatory interaction between miR-668-3p and ETS1 mRNA, potential binding sites were predicted using the ENCORI database, leading to the construction of three mutant ETS1 mRNA sequences (ETS1 MUTE1-3) (Fig. 6A). The pmirGLO plasmids containing the 3’UTR regions of both wild-type and mutant ETS1 genes were transfected into 293 T cells, and ETS1 expression was assessed by measuring cellular luciferase activity. After treatment with a miR-668-3p mimic, luciferase activity significantly decreased in the WT group, while MUTE1 and MUTE3 groups exhibited higher activity compared to the WT. In contrast, no significant change was observed in the MUTE2 group relative to WT (Fig. 6B). As expected, treatment with a miR-668-3p inhibitor yielded the opposite results, increasing luciferase activity in the WT and MUTE2 groups, while decreasing it in the MUTE1 and MUTE3 groups (Fig. 6C). These results suggest that miR-668-3p binds to ETS1 mRNA through the MUTE1 sequence. Fig. 6Validation of ETS1 expression regulation by miR-668-3p. (A) Prediction of the binding site and construction of the MUTE site of miR-668-3P with ETS1 mRNA using the ENCORI database. The Luciferase activity in 293T cells loaded with the pmirGLO plasmid ligated with the 3’UTR region of the wild-type and mutant ETS1 genes after treatment with miR-668-3p mimic (B) or inhibitor (C). (D-F) The binding of miR-668-3p to ETS1 detected by RIP assay. (G) The binding of miR-668-3p to ETS1 detected by RNA-pull down assay. The protein (H) and mRNA(I) levels of ETS1 in the cytoplasm and nucleus of BGC823 cells treated with miR-668-3p mimic or inhibitor. ** p < 0.01 between indicated groups Validation of ETS1 expression regulation by miR-668-3p. (A) Prediction of the binding site and construction of the MUTE site of miR-668-3P with ETS1 mRNA using the ENCORI database. The Luciferase activity in 293T cells loaded with the pmirGLO plasmid ligated with the 3’UTR region of the wild-type and mutant ETS1 genes after treatment with miR-668-3p mimic (B) or inhibitor (C). (D-F) The binding of miR-668-3p to ETS1 detected by RIP assay. (G) The binding of miR-668-3p to ETS1 detected by RNA-pull down assay. The protein (H) and mRNA(I) levels of ETS1 in the cytoplasm and nucleus of BGC823 cells treated with miR-668-3p mimic or inhibitor. ** p < 0.01 between indicated groups Further validation was carried out using RNA immunoprecipitation (RIP) assays in BGC823 and MKN45 cell lines, showing significantly higher levels of miR-668-3p and ETS1 mRNA within the Ago2 complex (Fig. 6D-F). Additionally, ETS1 mRNA was successfully pulled down using a wild-type miR-668-3p probe (miR-668-3p-biotin-wt) but not with a mutant probe (miR-668-3p-biotin-MUTE), further confirming the interaction between miR-668-3p and ETS1 mRNA (Fig. 6G). Finally, western blot and qPCR analyses demonstrated that while treatment with miR-668-3p inhibitors or mimics did not alter ETS1 mRNA levels, it significantly affected ETS1 protein levels in both the cytoplasm and nucleus (Fig. 6H-I). This indicates that miR-668-3p directly binds to the 3’UTR region of ETS1 mRNA and inhibits its translation in the cytoplasm, leading to reduced ETS1 protein levels in both the cytoplasm and nucleus. Collectively, these results confirm that miR-668-3p regulates the transcription factor ETS1 by directly interacting with its mRNA, modulating ETS1 protein production. miR668-3p promotes DDP resistance in GC cells via ETS1/EGFR To explore the influence of miR-668-3p and ETS1 on EGFR protein levels, a series of experiments were conducted on BGC823 cells with modified expression of these factors. An increase in miR-668-3p expression led to a significant reduction in EGFR protein levels, while simultaneous ETS1 overexpression restored EGFR levels. Conversely, inhibiting miR-668-3p resulted in a notable increase in EGFR protein levels, an effect that was reversed by ETS1 knockdown (Fig. 7A). These observations were further supported by measurements of EGFR mRNA levels, which mirrored the protein level changes (Fig. 7B). Next, BGC823 cells, including those overexpressing ETS1 or EGFR (OE), were treated with a miR-668-3p mimic and evaluated for DDP resistance. Cells in the miR-668-3p mimic group exhibited significantly higher cell viability (Fig. 7C), IC50 values (Fig. 7D), and colony formation (Fig. 7E) upon DDP treatment. Importantly, these effects were notably mitigated by overexpression of ETS1 or EGFR. Fig. 7miR668-3p promotes DDP resistance in GC cells via ETS1/EGFR. The protein (A-B) and mRNA(B) levels of EGFR in BGC823 cells treated with miR-668-3p mimic, miR-668-3p mimic with ETS1-OE, miR-668-3p inhibitor and miR-668-3p inhibitor with siETS1. ** p < 0.01 between indicated groups. The cell viability (C), IC50 values (D) and clone formation (E) of BGC823 treated with miR-668-3p mimic, miR-668-3p mimic with ETS1-OE, miR-668-3p inhibitor and miR-668-3p inhibitor with siETS1 in DDP. ** p < 0.01 between indicated groups. The protein (F-G) and mRNA(G) levels of EGFR in BGC823 cells treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE. @@p < 0.01 compared to Control; ##p < 0.01 compared to M2-CM. (H) The levels of autophagy-related proteins in BGC823 cells treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE in DDP. (I) The level of LC3-II in BGC823 and MKN45 treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE in DDF observed by immunofluorescence miR668-3p promotes DDP resistance in GC cells via ETS1/EGFR. The protein (A-B) and mRNA(B) levels of EGFR in BGC823 cells treated with miR-668-3p mimic, miR-668-3p mimic with ETS1-OE, miR-668-3p inhibitor and miR-668-3p inhibitor with siETS1. ** p < 0.01 between indicated groups. The cell viability (C), IC50 values (D) and clone formation (E) of BGC823 treated with miR-668-3p mimic, miR-668-3p mimic with ETS1-OE, miR-668-3p inhibitor and miR-668-3p inhibitor with siETS1 in DDP. ** p < 0.01 between indicated groups. The protein (F-G) and mRNA(G) levels of EGFR in BGC823 cells treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE. @@p < 0.01 compared to Control; ##p < 0.01 compared to M2-CM. (H) The levels of autophagy-related proteins in BGC823 cells treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE in DDP. (I) The level of LC3-II in BGC823 and MKN45 treated with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM with ETS1-OE and M2-CM with EGFR-OE in DDF observed by immunofluorescence Further experiments involved treating BGC823 cells with M2-CM, M2-CM depleted of EXO, M2-CM depleted of miR-668-3p, M2-CM combined with ETS1-OE, and M2-CM combined with EGFR-OE. Western blot analysis revealed that culturing with M2-CM led to a marked decrease in EGFR expression, while no significant changes in EGFR levels were observed in the groups treated with M2-CM depleted of EXO or miR-668-3p. However, both ETS1-OE and EGFR-OE treatments significantly elevated EGFR expression (Fig. 7F). mRNA analysis results were consistent with these findings (Fig. 7G). Western blot was used to detect the phosphorylation level of EGFR (p-EGFR Y1068) and the changes in key proteins of the downstream PI3K/AKT/mTOR pathway (e.g., p-AKT, p-mTOR) after M2-CM treatment. The results showed that although the total EGFR protein level decreased, the levels of p-EGFR and its downstream p-AKT and p-mTOR also decreased accordingly, indicating that the EGFR signaling pathway was inhibited overall (Supplementary Figure S3). This is consistent with autophagy activation (increased LC3-II/I ratio and decreased p62). Moreover, a decrease in p62, an increase in beclin1, and a higher LC3-II/LC3-I ratio-markers of enhanced autophagic activity—were observed exclusively in the M2-CM group, with no similar changes in the other four groups (Fig. 7H). Immunofluorescence analysis confirmed these results, showing a significant increase in autophagy, indicated by LC3-II expression, in the M2-CM group alone (Fig. 7I). In addition, the autophagy flux assay was performed using mCherry-GFP-LC3 reporter system and Bafilomycin A1 treatment. Consistent with the autophagy-related marker changes, the results confirmed that the autophagic flux were significantly enhanced in the M2-CM group, without similar changes in the other four groups (Supplementary Figure S4A-B). Bafilomycin A1 or chloroquine treatment significantly inhibited the autophagic flux in the M2-CM group. To definitively establish whether autophagy activation is a necessary mechanism for the miR-668-3p-mediated DDP resistance, we employed the autophagy inhibitor Chloroquine (CQ). As shown in Supplementary Figure S4C, treatment with either M2-EXO or miR-668-3p mimic significantly enhanced the viability of GC cells under DDP treatment, confirming their role in promoting chemoresistance. Crucially, the addition of CQ largely abolished this protective effect. The viability of cells in both the “M2-EXO + CQ” and “miR-668-3p mimic + CQ” groups was significantly reduced to a level comparable to the control group. These results provide direct functional evidence that autophagy activation is essential for miR-668-3p to confer DDP resistance in GC cells. In conclusion, miR-668-3p promotes DDP resistance in GC cells by enhancing autophagic activity through the ETS1/EGFR pathway. Tumor xenograft experiment in nude mice confirm that miR-668-3p in M2-CM enhances DDP resistance in GC cells via ETS1/EGFR To validate the in vivo role of miR-668-3p, ETS1, and EGFR in mediating DDP resistance in GC, tumor xenograft experiments were performed using BALB/c nude mice. The mice were subcutaneously injected with BGC823 cells subjected to various treatments, and tumor growth rate and weight were measured over time. The results demonstrated that M2-CM significantly promoted tumor growth, while M2-CM depletion of exosomes or miR-668-3p did not significantly affect tumor progression. Notably, overexpression of ETS1 or EGFR effectively counteracted the tumor-promoting effects induced by M2-CM (Fig. 8A-C). Moreover, M2-CM exposure led to a pronounced increase in miR-668-3p levels within tumor tissues, an effect that was not observed in the other treatment groups (Fig. 8D). Immunohistochemical analysis of tumor tissues revealed that M2-CM treatment resulted in increased tumor proliferation, as indicated by elevated Ki-67 staining, alongside enhanced autophagic activity, marked by increased Beclin1 and LC3 staining and reduced p62 staining. Additionally, M2-CM treatment significantly lowered ETS1 and EGFR expression levels, while no significant changes in these markers were noted in the other groups (Fig. 8E). These results confirm that miR-668-3p, delivered via M2-CM, drives DDP resistance in GC cells by modulating the ETS1/EGFR axis, thereby promoting autophagic activity in vivo. Fig. 8Tumor xenograft experiment in nude mice confirm that miR-668-3p in M2-CM enhances DDP resistance in GC cells via ETS1/EGFR. (A-C) The tumor growth rate and weight of the subcutaneous xenograft tumor model in nude mice. ** p < 0.01 between indicated groups. @@p < 0.01 compared to Control; ##p < 0.01 compared to M2-CM. (D) The level of miR-668-3p in various tumor tissues examined by qPCR. (E) The protein levels of tumor growth marker, autophagy-related proteins, ETS1 and EGFR in various tumor tissues visualized by immunohistochemistry Tumor xenograft experiment in nude mice confirm that miR-668-3p in M2-CM enhances DDP resistance in GC cells via ETS1/EGFR. (A-C) The tumor growth rate and weight of the subcutaneous xenograft tumor model in nude mice. ** p < 0.01 between indicated groups. @@p < 0.01 compared to Control; ##p < 0.01 compared to M2-CM. (D) The level of miR-668-3p in various tumor tissues examined by qPCR. (E) The protein levels of tumor growth marker, autophagy-related proteins, ETS1 and EGFR in various tumor tissues visualized by immunohistochemistry Discussion GC, the fifth leading cause of cancer-related mortality globally [1], is often asymptomatic in its early stages, leading to late diagnoses or distant metastasis [34, 35]. Resistance to DDP is a key factor driving poor treatment outcomes and accelerating disease progression [5, 6]. Although numerous strategies have been developed to overcome this resistance, the complexity and multifactorial nature of the underlying mechanisms limit the efficacy of current therapies, which are often associated with significant adverse effects [6, 7]. This underscores the need for a deeper understanding of the biological pathways underlying DDP resistance and for the development of novel therapeutic strategies with strong clinical potential. miRNA-targeted therapies hold promise for enhancing the efficacy of DDP by modulating multiple genes and pathways [36–38]. Extensive research suggests miRNAs play a key role in DDP resistance by regulating drug efflux proteins, apoptosis pathways, DNA repair, and the cell cycle [7, 12, 13, 39]. For instance, miR-21 has been shown to target PTEN, activating the PI3K/AKT pathway and upregulating ABC transporters—drug efflux proteins that contribute to DDP resistance [40, 41]. Similarly, miR-20a has been found to induce DDP resistance by downregulating EGR2 and CYLD in GC cells [42, 43]. This study identified a marked upregulation of miR-668-3p in DDP-resistant GC tissues and cell lines, correlating with reduced patient survival. Previous studies have reported strong miR-668-3p expression across various tumor types. In GC, miR-668-3p promotes cancer cell growth by regulating SOCS3 [45], and similarly drives cell replication and tumor expansion in hepatocellular carcinoma by suppressing BTG2 46. However, its role in drug resistance was previously unexplored. Our findings reveal that elevated miR-668-3p expression in GC cells enhances their growth and proliferation under DDP treatment, suggesting its involvement in the development of DDP resistance. MiR-668-3p is also involved in other caner types. It is crucial in cancer biology, especially non-small cell lung cancer (NSCLC) [46]. Additionally, the circZNF707/miR-668-3p/PFKM axis promotes NSCLC progression via enhancing glycolysis, proliferation, and invasion, indicating miR-668-3p as a therapeutic target [47].​ In colorectal cancer (CRC), the circTMEM59/miR-668-3p/ID4 axis inhibits CRC cell growth and metastasis, supporting miR-668-3p’s tumor-suppressive role [48]. In gastric cancer, the circ_0003356/miR-668-3p/SOCS3 axis suppresses tumor growth and enhances apoptosis, further confirming this role [44]. These show miR-668-3p’s context-dependent tumor-promoting or -suppressive effects in different cancers.​ MiR-668-3p also acts in hepatocellular carcinoma (HCC) [49]. In summary, miR-668-3p has versatile potential in cancer diagnosis, prognosis, and therapy. Its dual role (tumor suppressor/promoter) based on cancer type and molecular context makes it a key target for research and clinical applications, with potential to improve precision cancer management and personalized medicine. The pro-resistance effect of miR-668-3p in this study depends on delivery by M2 macrophage-derived exosomes, and high infiltration of M2-type TAMs in the GC tumor microenvironment may be the key context for its pro-tumor role. In other cancers (e.g., colorectal cancer), without continuous delivery by M2 macrophages, miR-668-3p may exert an anti-tumor effect by targeting different downstream molecules (e.g., ID4), which also explains the tissue-specificity of its function. Exosomes are established mediators of intercellular communication, facilitating the transfer of proteins, lipids, and nucleic acids [10, 50]. TAMs, particularly those of the M2-like phenotype, have been shown to secrete exosomes containing miRNAs, lncRNAs, and proteins, which contribute to drug resistance in tumor cells [8, 9, 14–16, 51, 52]. Consistent with this, our findings revealed a significant association between a high proportion of M2 macrophages and reduced survival rates in patients with GC. Additionally, M2 macrophages were found to secrete more exosomes than M0 and M1 macrophages. Treatment of GC cells with either M2 macrophage-derived culture medium or exosomes notably elevated miR-668-3p levels, promoting DDP resistance. Furthermore, reintroducing miR-668-3p into GC cell lines following exposure to M2-derived exosomes restored DDP sensitivity, suggesting that M2 macrophage-derived exosomes promote DDP resistance through the delivery of miR-668-3p. Autophagy, a process involved in the degradation and recycling of cellular components, plays a dual role in cancer, functioning as both a tumor suppressor and promoter during malignant transformation [53, 54]. In early tumor development, autophagy has a tumor-inhibiting effect, facilitating the removal of harmful cellular elements or inducing autophagic cell death. However, under therapeutic stress, autophagy serves as a protective mechanism, aiding in tumor growth and survival [27, 55]. Recent studies highlight the pivotal role of autophagy activation in mediating DDP resistance in GC cells [29]. For instance, circular RNA circCUL2 has been shown to enhance DDP resistance through miR-142-3p/ROCK2-mediated autophagy [30], while CLIC1 modulates DDP sensitivity by activating autophagy [14]. In line with these findings, our study also detected elevated miR-668-3p expression enhancing autophagy in GC cells. Most importantly, our study provides direct causal evidence linking miR-668-3p to autophagy-mediated resistance. The key finding that pharmacological inhibition of autophagy by Chloroquine completely reversed the DDP resistance induced by both M2-EXO and miR-668-3p mimic underscores that autophagy is not merely a correlative phenomenon but is functionally essential for this specific resistance mechanism. This critical experiment solidifies the proposed pathway, in which exosomal miR-668-3p from M2 macrophages targets ETS1 to downregulate EGFR. This downregulation, in turn, is both necessary and sufficient to trigger a pro-survival autophagic response, ultimately leading to chemotherapy failure. The role of ETS1 in mediating cisplatin resistance has been a focal point of recent research, particularly in the context of various cancers. ETS1, a transcription factor, has been implicated in promoting resistance to cisplatin, a widely used chemotherapeutic agent, through various molecular pathways. The study of ETS1’s involvement in cisplatin resistance is crucial for developing targeted therapies to overcome this resistance and improve treatment outcomes [56–58]. However, a growing body of evidence indicates that the function of ETS1 is highly context-dependent, and it may exert an inhibitory role in certain tumor types or under therapeutic stress. For instance, in various cancers, the expression of ETS1 was significantly decreased in cancer tissues, such as adrenocortical carcinoma, bladder urothelial carcinoma, breast invasive carcinoma, cervical squamous cell carcinoma and endocervical adenocarcinoma, kidney chromophobe, kidney renal papillary cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, ovarian serous cystadenocarcinoma, thyroid carcinoma, uterine corpus endometrial carcinoma, uterine carcinosarcoma [59]. These results indicated that ETS1 may suppress tumor growth under some circumstances. In the present study, we found that ETS1 expression was significantly downregulated in DDP-resistant GC tissues, and its expression level was positively correlated with patient survival. These findings imply that ETS1 may act as a tumor suppressor in the context of DDP resistance. EGFR typically inhibits autophagy through the PI3K/Akt/mTOR pathway [27]. The results indicated a reduction in EGFR expression in DDP-resistant GC tissues. The phosphorylation level of EGFR as well as the downstream PI3K/AKT/mTOR signaling were also correspondingly attenuated, which was consistent with the autophagy activation phenotype. Results showed that after M2-CM treatment, not only the total EGFR protein level decreased, but also the phosphorylated form (p-EGFR Y1068) and the phosphorylation levels of downstream PI3K/Akt/mTOR pathway molecules (p-Akt, p-mTOR) significantly decreased. Since this pathway is a classical inhibitory pathway of autophagy, EGFR downregulation promotes autophagy by relieving the activation of the PI3K/Akt/mTOR pathway, ultimately enhancing DDP resistance. An intriguing finding of our study is the significant reduction of EGFR expression in DDP-resistant GC tissues, which stands in contrast to a substantial body of literature linking EGFR overexpression to tumor progression and chemoresistance in various cancers [60, 61]. This apparent discrepancy underscores the complex, context-dependent nature of oncogenic signaling. We propose several non-mutually exclusive hypotheses to explain this phenomenon. First, under the sustained stress of DDP chemotherapy, cancer cells may undergo signal reprogramming. The downregulation of a potent mitogenic driver like EGFR could represent an adaptive strategy to decelerate cell proliferation and promote a pro-survival, slow-cycling state, for which autophagy is a key facilitator. Second, and most pertinent to our mechanistic model, is the well-established role of the EGFR/PI3K/Akt/mTOR axis as a key negative regulator of autophagy [27]. Our data show that the downregulation of EGFR is coupled with a concomitant attenuation of its downstream phosphorylated Akt and mTOR. This suggests that the miR-668-3p/ETS1 axis does not merely lower EGFR levels but actively derepresses the autophagy pathway. In this specific context, the loss of EGFR function is not a loss of oncogenic activity but rather a switch to an alternative survival mechanism. This aligns with emerging views that EGFR can modulate autophagy and cell fate through both canonical and non-canonical pathways, and its role can switch under therapeutic stress [27, 28]. Finally, this unique mechanism is likely orchestrated by the specific tumor microenvironment, particularly the influence of M2 macrophages. The transfer of exosomal miR-668-3p imposes an external regulatory layer that overrides the cancer cell’s intrinsic signaling state, creating a specific niche where the conventional pro-tumorigenic role of EGFR is supplanted by its role as a gatekeeper of autophagy-dependent survival. Through screening transcription factors that interact with miR-668-3p and regulate EGFR, ETS1 was identified as being downregulated in DDP-resistant GC cells. As a key member of the ETS family of transcription factors, ETS1 plays critical roles in stem cell development, cellular senescence, apoptosis, and tumorigenesis [56]. Previous studies have shown that ETS1 overexpression promotes tumor progression and resistance to sorafenib in hepatocellular carcinoma [57, 58]. Similar variations in the function of the ETS1/EGFR pathway have been observed across different cancers. Further in vivo and in vitro analyses revealed that miR-668-3p directly binds to ETS1 mRNA, inhibiting ETS1 protein synthesis and nuclear translocation, which in turn reduces EGFR transcription. This EGFR suppression enhances autophagy, thereby contributing to DDP resistance in GC cells. Although microRNAs can regulate hundreds of targets, our study provides compelling evidence that the ETS1/EGFR axis is a primary and functionally dominant pathway through which exosomal miR-668-3p promotes DDP resistance. This conclusion is supported by our initial bioinformatics screening, which identified ETS1 as a key node linking miR-668-3p to EGFR regulation, and was rigorously validated by a series of functional experiments. The fact that restoring ETS1 or EGFR expression was sufficient to reverse the resistance induced by miR-668-3p or M2 macrophage-derived exosomes underscores the centrality of this axis. Although other predicted targets of miR-668-3p (e.g., SOCS3 or BTG2 45, 46, 51) may have auxiliary roles in different contexts, their effects in our model appear subordinate to the ETS1/EGFR/autophagy pathway. Future comprehensive analyses, such as Ago2-CLIP or transcriptomic profiling, could further elucidate the full network of genes co-regulated by miR-668-3p in this specific context. Limitations and conclusion This study demonstrates that M2 macrophage-derived exosomal miR-668-3p promotes DDP resistance in GC via the ETS1/EGFR-autophagy axis; however, several limitations should be noted. First, it mainly used established GC cell lines (BGC823 and MKN45) and a cell line-derived xenograft model, which, while powerful for mechanistic dissection, cannot fully replicate the complexity and heterogeneity of the human tumor microenvironment (TME). Future studies should validate these findings in patient-derived organoids (PDOs) or xenografts (PDXs), which better preserve the original tumor’s characteristics and have higher clinical predictive value. Second, while miR-668-3p targets ETS1 to downregulate EGFR and activate autophagy (a survival mechanism here), the precise molecular switch governing autophagy’s dual role (survival vs. cell death) in GC needs further study. Third, the identified miR-668-3p/ETS1/EGFR axis was not explored for crosstalk with other known resistance mechanisms (e.g., other exosomal miRNAs, DNA repair pathways). Drug resistance is likely multifactorial, and this axis may work in concert with other pathways. In summary, despite these limitations, this study reveals a novel exosome-mediated mechanism of DDP resistance: exosomes from M2 macrophages transfer miR-668-3p to GC cells, where it binds to ETS1 mRNA, inhibiting its translation and leading to the downregulation of EGFR. This reduction in EGFR expression enhances pro-survival autophagy, thereby promoting DDP resistance. This work not only highlights miR-668-3p as a promising therapeutic target but also paves the way for developing novel strategies to overcome chemoresistance in GC, with the clear next step being validation in more clinically relevant models. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3. Supplementary Material 4. Supplementary Material 4. Supplementary Material 5. Supplementary Material 5. Supplementary Material 6. Supplementary Material 6.