CAF-derived exosomes drive trastuzumab resistance in HER2-positive breast cancer via YAP-USP8-HER2 axis activation Authors: Weijia Yu, Li Jiang Published: 2026 ABSTRACT AbstractBackground and objectiveCancer-associated fibroblasts (CAFs), critical constituents of the tumor microenvironment, promote tumor progression and therapeutic resistance through exosome secretion. However, the mechanisms by which CAF-derived exosomes contribute to trastuzumab resistance in HER2-positive breast cancer remain unclear. This study investigated whether CAF-derived exosomes activate the YAP-USP8 axis to stabilize HER2 and induce trastuzumab resistance.MethodsPrimary CAFs and normal fibroblasts (NFs) were isolated from HER2-positive breast cancer specimens, and exosomes were characterized by TEM, nanoparticle tracking analysis, and immunoblotting of exosomal markers. HER2-positive breast cancer cells (HCC1954 and BT-474) were treated with CAF- or NF-derived exosomes in vitro. Mechanistic studies were performed mainly in HCC1954 cells, whereas BT-474 cells were used for validation. Cargo dependency was assessed using RNase/Proteinase K protection assays, and YAP pathway involvement was examined using Verteporfin. In vivo tumor growth was evaluated using xenograft models.ResultsCAF-derived exosomes reduced LATS1 and YAP phosphorylation (p < 0.001), enhanced nuclear YAP accumulation, and upregulated USP8 transcription, leading to HER2 stabilization (P < 0.001). Membrane disruption abolished Hippo suppression, indicating dependence on intravesicular cargo. CAF-exo treatment increased proliferation, elevated trastuzumab IC50, reduced apoptosis, and accelerated tumor growth in vivo (all p < 0.001). These effects were reproduced in BT-474 cells and were reversed by YAP or USP8 silencing. Pharmacological inhibition of YAP similarly restored trastuzumab sensitivity. TCGA-BRCA analysis revealed positive correlations between YAP1 and USP8, and between USP8 and ERBB2, supporting the molecular relevance of this axis.ConclusionThis study identified a novel CAF-exosome-YAP-USP8-HER2 signaling axis that drives trastuzumab insensitivity in HER2-positive breast cancer. Intervening in this pathway may represent a potential treatment approach to counteract microenvironment-induced chemoresistance.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13058-026-02288-7.Keywords: Cancer-associated fibroblasts, Exosomes, YAP, USP8, HER2, Trastuzumab resistance Introduction Breast cancer continues to be the most commonly diagnosed malignancy and a primary contributor to cancer-linked deaths in women globally, with more than 2.3 million new cases and approximately 685,000 fatalities reported each year according to the GLOBOCAN 2020 statistics [1]. Among breast cancer subtypes, approximately 15–20% of patients overexpress human epidermal growth factor receptor 2 (HER2), a membrane-bound tyrosine kinase known to drive aggressive disease phenotypes, frequent relapse, and adverse outcomes [2, 3]. The introduction of HER2-targeted agents, particularly trastuzumab, has markedly enhanced clinical prognosis in patients with HER2-positive breast cancer [4]. However, intrinsic and acquired resistance to trastuzumab still affects up to 30–50% of patients [5], representing a critical barrier to durable treatment efficacy and emphasizing the urgent necessity to uncover the molecular basis of therapeutic resistance [6]. Increasing evidence points to the tumor microenvironment (TME) being an essential driver of drug insensitivity in solid tumors. Among the cellular components of the TME, cancer-associated fibroblasts (CAFs) have emerged as potent regulators in cancer development and therapeutic outcomes [7]. CAFs exert their effects not only through direct cell–cell interactions but also via secretion of paracrine factors, including exosomes—small extracellular vesicles (30–150 nm) that carry bioactive molecules such as RNAs, proteins, and lipids [8]. In breast cancer, CAF-derived exosomes have been implicated in promoting proliferation, invasion, stemness, and chemoresistance; however, their role in modulating HER2-targeted therapy resistance remains largely unexplored [9]. One emerging regulatory axis implicated in cancer progression and drug resistance is the Hippo signaling pathway [10]. Yes-associated protein (YAP), a central pathway regulator, is typically phosphorylated and sequestered in the cytoplasm under normal conditions, where it drives gene transcription linked to cell survival, growth, as well as stemness [11]. Aberrant activation of YAP has been associated with resistance to multiple anticancer agents across tumor types [12]. In HER2-positive breast cancer, emerging evidence has suggested that YAP may crosstalk with receptor tyrosine kinases, although the downstream mechanisms require further clarification [13]. Deubiquitinating enzymes (DUBs), such as ubiquitin-specific protease 8 (USP8), are key regulators of protein turnover by modulating the stability of oncogenic receptors including EGFR and HER2 [14]. USP8 is known to stabilize its substrates by removing polyubiquitin chains, thereby preventing their degradation via the proteasome [15]. While the USP8–HER2 interaction has been suggested in non-breast cancer contexts, whether YAP can transcriptionally regulate USP8 to stabilize HER2 and how this might be influenced by exogenous signals like CAF-derived exosomes are still not completely understood. This research sought to determine whether CAF-derived exosomes contribute to trastuzumab resistance within HER2-positive breast tumors by modulating the Hippo–YAP–USP8 pathway. Using in vitro and in vivo models, molecular and functional assays, and mechanistic analyses, we sought to delineate a novel exosome-mediated regulatory pathway that promotes HER2 stabilization and reduces drug sensitivity, potentially offering new therapeutic targets to overcome resistance in breast cancer. Materials and methods Clinical specimens Primary tumor tissues and adjacent histologically normal tissues were collected from female patients diagnosed with HER2-positive breast cancer who underwent surgery performed at Wujin Hospital Affiliated with Jiangsu University. Tumor and adjacent normal tissues were obtained from a total of n = 6 patients, and CAFs and NFs were isolated from each patient-derived specimen independently. CAFs and corresponding NFs were established and maintained as separate primary cultures from each individual patient. Unless otherwise specified, exosomes used in functional experiments were derived from at least three independent CAF isolates to minimize patient-specific bias. Comparable biological effects were observed across different CAF preparations. Tumor-derived CAFs were extracted, while normal fibroblasts (NFs) were derived from adjacent non-tumorous tissues using enzymatic digestion and differential adhesion, as previously described [16]. The study protocol was approved by the Institutional Review Board of Wujin Hospital Affiliated with Jiangsu University (Ethics Approval No. 2025-SR-422). Fibroblast characterization The morphological characteristics of primary CAFs and NFs were observed using phase-contrast microscopy (Olympus IX73, Japan). Cells were seeded in 6-well plates and cultured for 24–48 h. Representative spindle-shaped fibroblast-like morphology was documented under 200× magnification. To confirm the mesenchymal phenotype, immunofluorescence staining was performed to detect α-smooth muscle actin (α-SMA) and Vimentin expression. Cells grown on sterile glass coverslips were fixed with 4% paraformaldehyde (20 min), permeabilized with 0.2% Triton X-100 (10 min), and blocked using 5% BSA (1 h). After overnight incubation at 4 °C with α-SMA and Vimentin primary antibodies (1:200 each), Alexa Fluor-labeled secondary antibodies (1:500) were applied for 1 h at ambient temperature. Nuclei were DAPI-counterstained. Fluorescent imaging was acquired with a Leica DMi8 microscope. Fluorescence was quantified using ImageJ (NIH, version 1.53). Exosome extraction and characterization CAFs and NFs were cultured in DMEM containing 10% exosome-depleted FBS (System Biosciences, USA). Exosomes were isolated from conditioned media via differential ultracentrifugation (100,000 × g, 70 min, 4 °C). To assess the morphology and size distribution of exosomal vesicles, samples isolated from CAFs and NFs were first analyzed through transmission electron microscopy (TEM). Exosomal samples underwent fixation in 2% paraformaldehyde and applied onto 200-mesh carbon-coated copper grids for 20 min at ambient temperature. Subsequent to removing excess liquid, grids were negatively counterstained with 2% uranyl acetate (pH 7.0) in darkness for 10 min and air-dried. Images were acquired via a JEOL JEM-1400 TEM (80 kV, Japan) to visualize vesicle size, shape, and bilayer membrane structures. Exosomal size distribution and concentration were further assessed via nanoparticle tracking analysis (NTA) on the NanoSight NS300 system (Malvern Instruments, UK) equipped with a 488-nm laser module and CMOS detector. Exosome suspensions were diluted 1:1000 in sterile phosphate-buffered saline (PBS) and injected into the sample chamber. Measurements were performed at room temperature with a constant flow rate, recording three 60-second videos per sample. The NTA-derived results were presented as the mean particle diameter and size distribution profile. Exosomal marker expression (HSP70, TSG101, and CD63) was confirmed through western blot. Exosome cargo digestion assay To determine whether the Hippo pathway–suppressive effect of CAF-derived exosomes depends on specific vesicular cargo, isolated exosomes were subjected to enzymatic digestion and membrane disruption treatments prior to cell incubation. Briefly, CAF-derived exosomes (30 µg) were incubated with RNase A (100 U/mL) or Proteinase K (100 µg/mL) at 37 °C for 30 min to degrade external RNA or surface proteins, respectively. For membrane disruption, exosomes were treated with 0.04% Triton X-100 for 10 min at room temperature. For combined treatments, exosomes were first incubated with Triton X-100 followed by RNase A or Proteinase K digestion. After enzymatic treatment, RNase and Proteinase K were inactivated by heat treatment (65 °C for 10 min), and exosomes were re-isolated by ultracentrifugation (100,000 × g, 70 min, 4 °C) to remove residual enzymes and detergents. The treated exosomes were then resuspended in PBS and applied to HCC1954 cells for subsequent analyses. Exosome uptake assay To verify uptake of CAF- and NF-derived exosomes by HCC1954 cells, exosomes were diI-labeled (a lipophilic dye, Beyotime, C1036, China) according to the supplier’s instruction. Excess dye was eliminated via ultracentrifugation. Stained vesicles (30 µg/mL) were co-cultured with HCC1954 cells seeded on glass coverslips for 24 h. Following treatment, cells were rinsed with PBS, fixed with 4% paraformaldehyde, and counterstained with DAPI. Cytoplasmic red fluorescence indicated successful internalization, which was observed using a Leica DMi8 microscope. Cell maintenance and treatments in HER2-positive breast cancer cells The HER2-positive human breast cancer cell line HCC1954 and BT-474 were acquired from the American Type Culture Collection and propagated in RPMI-1640 medium (Gibco, USA) enriched with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 °C in a 5% CO2-saturated humidified environment. To investigate the effects of exosomes, gene knockdown, and trastuzumab treatment on HER2-positive breast tumor cells, the following experimental groupings were established. HCC1954 cell treatment: Exosome and trastuzumab treatment groups: (1) PBS group (untreated control); (2) NFs-exo group (30 µg/mL NF-derived exosomes); (3) CAFs-exo group (30 µg/mL CAF-derived exosomes); (4) PBS + Tra group (trastuzumab, 200 µg/mL); (5) NFs-exo + Tra group (NFs-exo plus trastuzumab); (6) CAFs-exo + Tra group (CAFs-exo plus trastuzumab) [17]. Exosome and trastuzumab treatment groups: (1) PBS group (untreated control); (2) NFs-exo group (30 µg/mL NF-derived exosomes); (3) CAFs-exo group (30 µg/mL CAF-derived exosomes); (4) PBS + Tra group (trastuzumab, 200 µg/mL); (5) NFs-exo + Tra group (NFs-exo plus trastuzumab); (6) CAFs-exo + Tra group (CAFs-exo plus trastuzumab) [17]. shRNA knockdown groups: (1) shNC (non-targeting control); (2) shYAP (YAP knockdown); (3) shUSP8 (USP8 knockdown). HCC1954 cells were transduced with shRNA constructs targeting either YAP or USP8 using lentiviral vectors (GeneChem, Shanghai, China). Transduction was performed following the supplier’s instructions, and stable cell lines were established through puromycin (2 µg/mL). shRNA knockdown groups: (1) shNC (non-targeting control); (2) shYAP (YAP knockdown); (3) shUSP8 (USP8 knockdown). HCC1954 cells were transduced with shRNA constructs targeting either YAP or USP8 using lentiviral vectors (GeneChem, Shanghai, China). Transduction was performed following the supplier’s instructions, and stable cell lines were established through puromycin (2 µg/mL). Combined Exosome + shRNA ± Trastuzumab groups: Cells were further divided into combinations of CAFs-exo and shRNA (shNC, shYAP, shUSP8), with or without trastuzumab treatment, to evaluate interactions between CAF-derived exosomes, YAP/USP8 signaling, and drug sensitivity. Combined Exosome + shRNA ± Trastuzumab groups: Cells were further divided into combinations of CAFs-exo and shRNA (shNC, shYAP, shUSP8), with or without trastuzumab treatment, to evaluate interactions between CAF-derived exosomes, YAP/USP8 signaling, and drug sensitivity. BT-474 cells were treated with PBS, NFs-exo (30 µg/mL), or CAFs-exo (30 µg/mL) under identical conditions. Xenograft establishment and treatment Six-week-old BALB/c female nude mice (weighing 22 ± 2 g) were sourced from the Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) and bred in a specific pathogen-free (SPF) environment with unrestricted access to feed and water. All in vivo experiments were approved by the Laboratory Animal Ethics Committee of Guangdong Medical Laboratory Animal Center (Approval No. D202502-7) and were conducted in accordance with national guidelines for the care and use of laboratory animals. HCC1954 cells (4 × 106 cells/150 µL PBS) were subcutaneously injected in the right flank of nude mice. Once tumors approached roughly 100 mm³, animals were randomly assigned to four treatment groups (n = 6 per group): (1) PBS group: Intratumoral injection of PBS every 5 days; (2) CAFs-exo group: Intratumoral injection of CAFs-derived exosomes (30 µg in 100 µL PBS) every 5 days; (3) PBS + Tra group: PBS injection as above + intraperitoneal injection of trastuzumab (20 mg/kg, once per week) for six consecutive weeks; (4) CAFs-exo + Tra group: CAFs-exo injection as above + trastuzumab as above [18, 19]. Tumor volumes were recorded at 3-day intervals using calipers, with volumes determined via the following formula: (length × width2)/2. At the endpoint of treatment, mice were euthanized, and tumor tissues were excised, photographed, and weighed. All measurements were performed in a blinded manner. In vivo tumor immunofluorescence Frozen tumor tissues were embedded in OCT compound and cryosectioned into 8 μm slices. Sections were fixed in pre-chilled acetone (10 min), air-dried, rehydrated with PBS, and blocked with 5% BSA for 1 h at ambient temperature. Slides were then exposed overnight at 4 °C to primary antibodies Ki67 (1:200, ab16667, Abcam) and HER2 (1:200, ab134182, Abcam). Once washed, the tissue sections were treated with fluorophore-labeled secondary antibodies, stained with DAPI, and mounted with antifade mounting medium (Beyotime). Fluorescent signals were visualized using a Leica DMi8 microscope, and fluorescence intensity was quantified using ImageJ (NIH, version 1.53). Western blot To evaluate the levels of exosome-associated and signaling proteins, western blot analysis was performed on exosome preparations, fibroblast lysates, and HCC1954 and BT-474 cells subjected to various treatments. Cells and exosome pellets were homogenized in RIPA buffer (Beyotime, P0013B, China) containing protease/phosphatase inhibitors (Beyotime, P1045), and quantified using a BCA kit (Beyotime, P0012). Proteins (20–30 µg) were subjected to SDS-PAGE and blotted onto PVDF membranes (Millipore, IPVH00010) which were blocked for 1 h at room temperature in 5% milk/TBST and then incubated with the following Abcam primary antibodies overnight at 4 °C: TSG101 (1:1000, ab125011), CD63 (1:1000, ab134045), HSP70 (1:1000, ab5439), LATS1 (1:1000, ab243656), p-LATS1 (1:1000, ab305029), YAP (1:1000, ab205270), p-YAP (1:1000, ab76252), USP8 (1:1000, ab228572), HER2 (1:1000, ab134182), Flag-HER2 (1:1000, ab205921), Ubiquitin (1:1000, ab7780), GAPDH (1:3000, ab9485), Histone H3 (1:3000, ab1791). After washing with TBST (3 × 10 min), membranes were exposed to HRP-labeled secondary antibodies (Abcam) at ambient temperature for 1 h: Goat anti-rabbit (ab6721) and anti-mouse (ab6789), both at 1:3000. Signal detection was achieved with enhanced chemiluminescence (ECL, Thermo Scientific, 32106, USA) and imaged on a Bio-Rad ChemiDoc MP system (USA). Densitometric analysis of band intensity was performed using ImageJ, with normalization to GAPDH or Histone H3. CCK-8 analysis Cell viability assessment was carried out using Dojindo’s Cell Counting Kit-8 (CCK-8, CK04, Japan) according to the supplier’s protocol. For HCC1954 cells, cells (PBS, NFs-Exo, CAFs-exo, CAFs-exo+shNC, CAFs-exo+shYAP, CAFs-exo+shUSP8, and VP, CAFs-exo + VP, PBS + Tra, PBS + Tra+VP, CAFs-exo + Tra, CAFs-exo + Tra+VP) were plated in 96-well plates (3 × 10³/well, 100 µL/well, 10% FBS). For Verteporfin treatment, cells were pretreated with VP (4 µM) for 2 h prior to exosome and/or trastuzumab exposure, and DMSO concentration was kept constant (< 0.1%) in all groups. For time-course analysis, cell viability was monitored at 0, 24, 48, and 72 h by applying 10 µL CCK-8 solution, and after 2 h at 37 °C, the optical density (OD) was detected at 450 nm with a BioTek Epoch reader (USA). To assess trastuzumab sensitivity, HCC1954 cells from the same treatment groups were exposed to increasing concentrations of trastuzumab (0, 0.5, 1, 10, 20, 50, 100, and 1000 µg/mL) for 72 h. Following the addition of CCK-8 solution, the OD was recorded as described above. For BT-474 cells, PBS-, NFs-exo (30 µg/mL)-, and CAFs-exo (30 µg/mL)-treated cells were seeded under identical conditions. Cell viability was assessed at 0, 24, 48, and 72 h. For drug sensitivity analysis, BT-474 cells were treated with increasing concentrations of trastuzumab (0, 0.5, 1, 10, 20, 50, 100, and 1000 µg/mL) in the presence or absence of exosomes for 72 h, followed by CCK-8 measurement as described above. Dose–response curves were generated in GraphPad Prism 9.0, with half-maximal inhibitory concentration (IC50) for trastuzumab obtained through nonlinear regression. Each experiment was repeated independently in triplicate. Bioinformatics analysis Transcriptomic data (RNA-seq) and corresponding clinical information for breast cancer (TCGA-BRCA cohort) were obtained from The Cancer Genome Atlas (TCGA) database. Gene expression levels of YAP1, USP8, and ERBB2 were extracted for downstream analyses. Breast cancer molecular subtypes were defined according to the PAM50 classification provided in the TCGA dataset. HER2-positive cases were defined as tumors classified as the HER2-enriched subtype by PAM50. All remaining molecular subtypes were grouped as non–HER2-enriched for comparative analyses. Spearman correlation coefficients were calculated to evaluate associations between YAP1, USP8, and ERBB2 expression levels in the overall cohort and in the HER2-enriched subgroup. For survival analysis, patients were stratified into high- and low-expression groups based on the median expression value of each gene. Kaplan–Meier overall survival curves were generated, and differences between groups were assessed using the log-rank test. All statistical analyses were performed using R software, with packages including survival and survminer. Clonogenic survival assay Colony formation assays were conducted to assess the long-term proliferative capacity of HCC1954 cells. Treated cells were seeded into 6-well plates at a density of 500 cells per well in RPMI-1640 medium supplemented with 10% FBS. After incubation under standard conditions (37 °C, 5% CO2) for 10–14 days, colonies were immobilized in 4% paraformaldehyde (20 min, room temperature) and visualized by 0.1% crystal violet staining (15 min, room temperature). Plates were gently rinsed with PBS and allowed to air-dry. Colonies > 50 cells were enumerated manually using a stereomicroscope. All conditions were tested in triplicate across three independent experiments. TUNEL staining Apoptosis was analyzed via the TUNEL assay (Beyotime, C1086) following the manufacturer’s protocol. Cells were plated on sterile glass coverslips in 24-well plates and treated with the respective treatments for 48 h. Following fixation with 4% paraformaldehyde (25 min) and 0.3% Triton X-100 permeabilization (10 min) at ambient temperature, TUNEL staining was performed using a one-step apoptosis detection kit (Beyotime, C1086, China) as directed by the manufacturer. After DAPI staining (5 min, room temperature), slides were sealed with antifade reagent. Apoptotic cells (green) and total nuclei (blue) were observed under a Leica DMi8 microscope, and five randomly selected fields of view were analyzed. The apoptosis index was defined as the fraction of TUNEL-labeled nuclei among DAPI-positive nuclei. Each experiment was performed in triplicate. Dual-luciferase reporter assay For YAP-USP8 promoter interaction analysis, HCC1954 cells were co-transfected with either a wild-type (WT) or mutant (MUT) USP8 promoter-driven firefly luciferase reporter plasmid (300 ng/well) along with a Renilla luciferase plasmid (50 ng/well) as internal control, using Lipofectamine 3000 (Thermo Fisher Scientific, USA) in 24-well plates. After 48 h of incubation (37 °C, 5% CO2), luciferase levels were detected using the Promega system (E1910, USA) and recorded on a GloMax luminometer (Promega, USA). Firefly luciferase activity was normalized to Renilla luciferase activity. All conditions were tested in triplicate and repeated in three independent experiments. Results were expressed as relative luciferase activity. Co-immunoprecipitation (Co-IP) To verify the interaction between USP8 and HER2 as well as the ubiquitination level of HER2 under various treatment conditions. For interaction analysis, total protein lysates from HCC1954 cells were prepared using IP buffer (Beyotime, P0013, China) containing protease inhibitors. Proteins (1 mg/sample) were incubated overnight at 4 °C with an anti-USP8 antibody (ab228572, Abcam) or normal IgG (negative control, ab172730, Abcam), followed by a 2-hour incubation with Protein A/G magnetic beads (Beyotime, P2117, China). After washing and elution in loading buffer, immunoprecipitates were analyzed via western blotting using anti-HER2 (ab134182, Abcam) to detect protein–protein interaction. For HER2 ubiquitination assays, lysates were immunoprecipitated with anti-Flag (ab205921, Abcam) to pull down Flag-tagged HER2. Ubiquitination was detected using anti-ubiquitin antibody (ab7780, Abcam), and GAPDH served as an internal loading control. Each experiment was repeated independently in triplicate. Statistical analysis Data analysis was carried out in SPSS v23.0 (IBM Corp., Armonk, NY, USA). Results were expressed as mean ± standard deviation (SD) from at least three independent experiments. The unpaired Student’s t-test (two-tailed) was applied for two-group analysis, and one-way analysis of variance (ANOVA) with Tukey’s post hoc test for multiple groups. A p-value < 0.05 was deemed significant. Results Identification and characterization of CAFs and CAF-derived exosomes To confirm the mesenchymal phenotype of isolated fibroblasts and validate the identity and purity of extracted exosomes, a series of morphological, immunofluorescence, ultrastructural, and molecular assays were conducted. As shown in Fig. 1A, CAFs displayed a more elongated, spindle-shaped morphology compared to the flatter, cobblestone-like appearance of NFs. Immunofluorescence analysis revealed significantly higher expression levels of α-SMA and Vimentin in CAFs than in NFs (P < 0.001), indicating successful enrichment of activated fibroblasts (Fig. 1B). Identification and characterization of CAFs and their derived exosomes. A Representative morphological features of CAFs and NFs under phase-contrast imaging (scale = 100 μm). B Immunofluorescence staining with quantification of α-SMA and Vimentin levels in CAFs vs. NFs (scale = 50 μm). C Transmission electron microscopy visualized bilayer-membrane vesicles in CAFs-exo and NFs-exo (scale = 200 nm). D Nanoparticle tracking analysis (NTA) of exosome size distribution from CAFs and NFs. E Immunoblotting of exosomal markers (TSG101, CD63, HSP70) in exosomes and parent cells. F DiI-labeled exosomes internalized by HCC1954 cells visualized under fluorescence microscopy (scale = 50 μm). Results are expressed as mean ± SD (n = 3). ***P < 0.001 vs. NFs Transmission electron microscopy demonstrated the presence of typical round or oval-shaped vesicles with bilayer membrane structures in both CAF-derived and NF-derived exosome preparations (Fig. 1C). Nanoparticle tracking analysis showed that both exosome populations had a typical size distribution ranging between 50 and 200 nm, with the mean particle diameter of CAFs-exo slightly larger than that of NFs-exo (Fig. 1D). Western blotting identified the enrichment of classical exosomal markers, such as TSG101, CD63, as well as HSP70, in CAFs-exo and NFs-exo, while their parental cells expressed minimal levels of these proteins (Fig. 1E). Furthermore, uptake studies using DiI-labeled exosomes showed efficient internalization of both CAFs-exo and NFs-exo into HCC1954 cells, as evidenced by abundant red fluorescent puncta distributed throughout the cytoplasm (Fig. 1F). Collectively, these results confirmed the successful isolation and characterization of functional exosomes from CAFs and NFs. CAF-derived exosomes enhance proliferation and activate YAP signaling in HCC1954 cells To explore the functional effect of CAF-derived exosomes on breast cancer cell behavior, we first evaluated the viability and proliferative capacity of HCC1954 cells under different treatment conditions. According to Fig. 2A, treatment with CAFs-exo significantly enhanced viability at 24, 48, and 72 h compared to PBS controls (P < 0.001), whereas NFs-exo treatment induced only a mild, non-significant increase. Consistently, colony formation assays revealed a significantly greater number of colonies in the CAFs-exo group versus control group (P < 0.001, Fig. 2B), further confirming their growth-promoting effect. CAF-derived exosomes enhance proliferation and activate YAP signaling in HCC1954 cells. A CCK-8 assay evaluating viability across time points after treatment with PBS, NFs-exo, or CAFs-exo. B Representative images and quantification of colony formation under the same conditions. C Dose-response curves of trastuzumab and corresponding IC50 values. D TUNEL staining showing apoptotic cells (green) and nuclei (blue) in each groups. E Immunoblotting of total and phosphorylated LATS1 and YAP in HCC1954 cells. F Immunoblotting of nuclear and cytoplasmic YAP. Data are shown as mean ± SD (n = 3). ***P < 0.001 vs. PBS; ###P < 0.001 vs. NFs-exo; &&&P < 0.001 vs. CAFs-exo To assess trastuzumab sensitivity, dose–response curves demonstrated that CAFs-exo-treated cells exhibited a markedly higher IC50 (799.7 µg/mL) than PBS (429.1 µg/mL) and NFs-exo (368 µg/mL) controls (Fig. 2C), suggesting reduced sensitivity to trastuzumab and potential induction of drug resistance. TUNEL staining further showed that CAFs-exo treatment alone significantly suppressed apoptosis compared to PBS (P < 0.001), whereas both PBS + Tra and NFs-exo + Tra treatments significantly elevated apoptotic rates. Notably, CAFs-exo + Tra treatment only partially restored apoptosis compared with CAFs-exo alone (P < 0.001), and the apoptotic rate remained considerably reduced compared with the PBS + Tra group (Fig. 2D). These findings indicated that CAF-derived exosomes attenuated trastuzumab-induced apoptosis and may protect breast cancer cells from drug-mediated cytotoxicity. Mechanistically, Western blotting showed that CAFs-exo exposure notably reduced the phosphorylated expression of LATS1 and YAP without altering their total protein levels (P < 0.001, Fig. 2E), indicating inactivation of the Hippo pathway. Subcellular fractionation showed increased nuclear YAP accumulation and decreased cytoplasmic YAP levels upon CAFs-exo exposure versus control treatment (P < 0.001, Fig. 2F), suggesting YAP activation through nuclear translocation. Collectively, these data demonstrated that CAF-derived exosomes promoted proliferation, suppressed apoptosis, induced trastuzumab resistance, and activated YAP signaling in HER2-positive breast cancer cells. Identification of CAF-exosomal cargo responsible for Hippo pathway suppression To explore the specific exosomal cargo responsible for Hippo pathway suppression, CAF-derived exosomes were pretreated with RNase A, Proteinase K, Triton X-100, or their combinations prior to incubation with HCC1954 cells. Western blot analysis demonstrated that RNase A treatment alone did not significantly reverse the CAF-exo–induced reduction in p-LATS1 and p-YAP levels. Similarly, Proteinase K treatment alone partially attenuated the effect. However, disruption of exosomal membrane integrity by Triton X-100, particularly in combination with RNase A or Proteinase K, significantly restored LATS1 and YAP phosphorylation (P < 0.001), indicating that intact vesicular structure and intravesicular cargo are required for Hippo pathway suppression (Fig. 3A, B). Intravesicular cargo of CAF-derived exosomes mediates Hippo pathway suppression in HCC1954 cells. A Western blot analysis of LATS1, p-LATS1, YAP, and p-YAP in HCC1954 cells treated with CAF-derived exosomes pretreated with Triton X-100, RNase A, Proteinase K, or their combinations. B Quantification of relative protein expression levels of p-LATS1/LATS1 and p-YAP/YAP normalized to GAPDH. Data are presented as mean ± SD (n = 3). ***P < 0.001 vs. CAFs-exo group; ###P < 0.001 vs. CAFs-exo+Triton group These findings suggest that the inhibitory effect of CAF-derived exosomes on the Hippo pathway is primarily mediated by encapsulated cargo rather than surface-bound proteins, implicating intravesicular RNA and/or protein components as potential upstream regulators. CAF-derived exosomes promote HER2 stabilization via activation of the YAP–USP8 signaling axis To elucidate how CAF-derived exosomes contribute to trastuzumab resistance at the molecular level, we explored the regulation of HER2 stability through the YAP–USP8 axis. Immunoblotting analysis revealed a marked upregulation of both USP8 and HER2 proteins in HCC1954 cells upon CAFs-exo treatment, compared to PBS and NFs-exo controls (P < 0.001), indicating a potential role of CAF-derived exosomes in promoting HER2 stabilization (Fig. 4A). CAF-released exosomes initiated the YAP–USP8 axis and reduced HER2 ubiquitination in HCC1954 cells. A, B Immunoblotting of USP8, HER2, and YAP pathway proteins in HCC1954 cells treated with CAFs-exo, NFs-exo, or PBS (A), or transfected with shYAP vs. shNC (B). C Luciferase reporter assay assessing the transcriptional activation of the USP8 promoter (WT or MUT) under different treatment conditions. D, E HER2 protein expression following YAP or USP8 knockdown. F Co-immunoprecipitation of USP8 and HER2 in HCC1954 cells. G Ubiquitination levels of HER2 in shNC vs. shUSP8 cells; Flag-HER2 used as transfection control. H, I HER2 ubiquitination (H) and total HER2 levels (I) in HCC1954 cells co-treated with CAFs-exo and shRNA constructs. Outcomes are presented as mean ± SD (n = 3). ***P < 0.001 vs. PBS, shNC; ###P < 0.001 vs. CAFs-exo+shNC To clarify the regulatory relationship between YAP and USP8, we silenced YAP in HCC1954 cells using shRNA. YAP knockdown significantly suppressed YAP and p-YAP expression and concomitantly reduced USP8 protein levels (P < 0.001), suggesting that USP8 may act as a downstream effector of YAP (Fig. 4B). Consistently, luciferase reporter assay showed that CAFs-exo treatment significantly enhanced USP8 promoter activity in cells transfected with the wild-type (WT) USP8 promoter, whereas this effect was abolished by shYAP or mutation of the YAP-responsive element (P < 0.001, Fig. 4C). These findings support a direct transcriptional regulation of USP8 by YAP. Next, we investigated whether the YAP–USP8 axis influences HER2 protein stability. Silencing YAP or USP8 in HCC1954 cells significantly reduced HER2 protein levels relative to control cells (P < 0.001, Fig. 4D, E). Co-IP assays further revealed a physical relationship linking USP8 to HER2 (Fig. 4F), suggesting that USP8 may stabilize HER2 via direct binding. To investigate whether USP8 modulates HER2 ubiquitination, we evaluated the levels of HER2-Ub in shUSP8 and control cells. Silencing USP8 led to a notable increase in HER2 ubiquitination without affecting Flag-HER2 expression (P < 0.001), indicating enhanced degradation of HER2 in the absence of USP8-mediated deubiquitination (Fig. 4G). Finally, we examined whether this regulatory axis mediates the effect of CAF-derived exosomes on HER2 stability. CAFs-exo treatment markedly decreased HER2 ubiquitination and increased HER2 protein levels, whereas silencing YAP or USP8 reversed these effects (P < 0.001, Fig. 4H, I). These findings collectively confirmed that CAF-derived exosomes promoted HER2 stabilization by activating the YAP–USP8 axis, thereby contributing to HER2-positive breast cancer cell insensitivity to trastuzumab. Clinical relevance of the YAP-USP8-HER2 axis in TCGA-BRCA To assess the clinical relevance of the YAP-USP8-HER2 signaling axis, we analyzed transcriptomic data from the TCGA-BRCA cohort. In the overall cohort, YAP1 expression showed a positive correlation with USP8 (Spearman r = 0.15, p = 0.00058), and USP8 expression was positively associated with ERBB2 (r = 0.26, p = 9.4 × 10⁻¹⁰) (Fig. 5A). Notably, within the HER2-positive subgroup, the correlation between YAP1 and USP8 was further strengthened (r = 0.39, p = 0.0029), whereas the association between USP8 and ERBB2 was not statistically significant (r = 0.02, p = 0.87). Kaplan–Meier survival analysis revealed that neither USP8 nor YAP1 expression was significantly associated with overall survival in the TCGA-BRCA cohort (YAP1: log-rank p = 0.17; USP8: log-rank p = 0.11) (Fig. 5B). Although a trend toward poorer survival was observed in the high-expression groups, these differences did not reach statistical significance. These findings support the clinical relevance of the YAP–USP8 axis in breast cancer. Clinical correlation of the YAP-USP8-HER2 axis in the TCGA-BRCA cohort. A Spearman correlation analyses showing the associations between YAP1 and USP8, and between USP8 and ERBB2 in the overall TCGA-BRCA cohort (left two panels) and in the HER2-positive subgroup (right two panels). Correlation coefficients (r) and p-values are indicated. B Kaplan–Meier overall survival curves stratified by USP8 expression (left) and YAP1 expression (right). Log-rank p-values are shown CAF-derived exosomes enhance HCC1954 cell proliferation and reduce trastuzumab sensitivity via YAP/USP8 signaling To better elucidate the function of CAF-derived exosomes in promoting trastuzumab resistance, a series of in vitro functional assays were performed in HCC1954 cells under different treatment conditions. As shown in Fig. 6A, the CAFs-exo group exhibited markedly enhanced viability relative to the PBS control at 24, 48, and 72 h (P < 0.001). Knockdown of YAP or USP8 (CAFs-exo+shYAP and CAFs-exo+shUSP8) significantly attenuated the exosome-induced increase in viability (P < 0.001), indicating that both YAP and USP8 contribute to CAF-exosome-mediated proliferation enhancement. CAF-derived exosomes enhance proliferation and trastuzumab resistance in HCC1954 cells via YAP–USP8 signaling. A CCK-8 assay of cell viability over 72 h under indicated treatments. B Colony‑forming images and corresponding quantitative data. C Trastuzumab sensitivity curves and IC50 values. D TUNEL staining and quantification of apoptotic cells (Scale = 50 μm). Results are expressed as mean ± SD (n = 3). ***P < 0.001 vs. PBS; ###P < 0.001 vs. CAFs-exo+shNC Consistent results were observed in the colony assay (Fig. 6B), where colony counts were markedly elevated in the CAFs-exo and CAFs-exo+shNC groups (P < 0.001), while colony numbers were significantly reduced in the CAFs-exo+shYAP and CAFs-exo+shUSP8 groups (P < 0.001). To assess drug sensitivity, dose–response curves for trastuzumab were generated (Fig. 6C). The IC50 value for trastuzumab was significantly elevated in the CAFs-exo group (747.9 µg/mL) compared to the PBS group (427.1 µg/mL), indicating reduced drug sensitivity. Notably, knockdown of YAP or USP8 partially restored trastuzumab sensitivity, as reflected by decreased IC50 values in the CAFs-exo+shYAP (562.3 µg/mL) and CAFs-exo+shUSP8 (492.1 µg/mL) groups. TUNEL staining results (Fig. 6D) confirmed that trastuzumab treatment significantly increased apoptosis in HCC1954 cells compared to PBS (P < 0.001). However, CAF-derived exosomes attenuated this effect, as shown by a marked reduction in apoptosis in the CAFs-exo + Tra group (P < 0.001). Similarly, CAFs-exo+shNC + Tra–treated cells exhibited low apoptotic rates. Notably, knockdown of YAP or USP8 in the presence of CAF-exosomes and trastuzumab significantly restored apoptosis levels compared to the CAFs-exo+shNC + Tra group (P < 0.001). Although apoptosis in the shUSP8 group was slightly higher than in the shYAP group, the difference was not statistically significant. Collectively, these data demonstrated that CAF-derived exosomes promoted HCC1954 cellular survival, growth, and trastuzumab resistance in a YAP–USP8-dependent manner. Pharmacological inhibition of YAP reverses CAF-exosome–induced trastuzumab resistance To further validate the functional importance of YAP activation in CAF-exosome–induced trastuzumab resistance, the YAP–TEAD inhibitor Verteporfin (VP) was employed as a pharmacological rescue strategy in HCC1954 cells. CCK-8 assays demonstrated that Verteporfin significantly attenuated CAF-exo–induced cell proliferation (P < 0.001, Fig. 7A). Moreover, in the presence of trastuzumab, VP treatment markedly restored drug sensitivity in CAF-exo–treated cells, as evidenced by reduced viability compared with the CAFs-exo + Tra group (P < 0.001). Consistently, western blot analysis revealed that Verteporfin treatment increased p-YAP levels while reducing USP8 and HER2 protein expression in CAF-exo–treated cells (Fig. 7B, C), indicating suppression of YAP transcriptional activity and downstream HER2 stabilization. Pharmacological inhibition of YAP reverses CAF-exosome–induced trastuzumab resistance in HCC1954 cells. A CCK-8 assay showing cell viability at 0, 24, 48, and 72 h in HCC1954 cells treated with PBS, VP, CAFs-exo, CAFs-exo + VP, PBS + Tra, PBS + Tra+VP, CAFs-exo + Tra, and CAFs-exo + Tra+VP. B Western blot analysis of YAP, p-YAP, USP8, and HER2 protein levels in HCC1954 cells under the indicated treatments. GAPDH served as a loading control. C Quantification of relative protein expression levels of p-YAP/YAP, USP8, and HER2. Data are presented as mean ± SD (n = 3). ***P < 0.001 vs. PBS; ###P < 0.001 vs. CAFs-exo; &&&P < 0.001 vs. CAFs-exo + Tra Collectively, these findings confirm that pharmacological inhibition of YAP effectively reverses CAF-exosome–mediated HER2 stabilization and trastuzumab resistance, further supporting the critical role of the YAP–USP8 axis in this process. Validation of CAF-derived exosome–induced trastuzumab resistance in BT-474 cells To further determine whether the pro-resistance effects of CAF-derived exosomes were reproducible in additional HER2-positive breast cancer models, parallel experiments were conducted in BT-474 cells. Consistent with observations in HCC1954 cells, CCK-8 assays demonstrated that CAF-derived exosomes significantly enhanced BT-474 cell viability at 24, 48, and 72 h compared with PBS controls (P < 0.001), whereas NFs-exo exerted only modest effects (Fig. 8A). Dose–response analysis further revealed that CAF-exo treatment markedly increased the IC50 value of trastuzumab in BT-474 cells relative to PBS and NFs-exo groups (Fig. 8B), indicating reduced drug sensitivity and confirming that CAF-derived exosomes promote trastuzumab resistance in this additional HER2-positive model. Mechanistically, western blot analysis showed that CAF-exo treatment reduced the phosphorylation levels of LATS1 and YAP without significantly altering total LATS1 or YAP expression (P < 0.001), suggesting suppression of Hippo pathway activity (Fig. 8C, D). Moreover, CAF-exo exposure led to significant upregulation of USP8 and HER2 protein levels in BT-474 cells (P < 0.001), consistent with enhanced HER2 stabilization. CAF-derived exosomes promote proliferation, activate YAP signaling, and induce trastuzumab resistance in BT-474 cells. A CCK-8 assay evaluating cell viability at indicated time points. B Dose–response curves of trastuzumab and corresponding IC50 values. C Immunoblotting of total and phosphorylated LATS1 and YAP, USP8, and HER2 proteins. D Quantification of protein expression levels normalized to GAPDH. Data are expressed as mean ± SD (n = 3). ***P < 0.001 vs. PBS; ###P < 0.001 vs. NFs-exo or NFs-exo-Tra Taken together, these findings indicate that CAF-derived exosomes activate the YAP-USP8-HER2 signaling axis and attenuate trastuzumab sensitivity not only in HCC1954 cells but also in BT-474 cells, thereby supporting the generalizability of this resistance mechanism across HER2-positive breast cancer models. CAF-derived exosomes attenuate trastuzumab efficacy by promoting tumor growth and HER2 stability in vivo To validate the in vivo role of CAF-derived exosomes in mediating trastuzumab resistance, we established a xenograft model using HER2-positive HCC1954 breast cancer cells and applied different treatment regimens. As shown in Fig. 9A, gross tumor morphology revealed that tumors from the CAFs-exo group exhibited visibly larger volumes compared to the PBS group, whereas tumors in the PBS + Tra group were markedly smaller. Notably, tumors in the CAFs-exo + Tra group were substantially larger than those in the PBS + Tra group, indicating that CAF-derived exosomes compromised the antitumor efficacy of trastuzumab. CAF-derived exosomes promote tumor growth and HER2 stabilization, attenuating trastuzumab efficacy in vivo. A Xenograft tumor specimens imaged post-excision. B, C Tumor volume progression (B) and endpoint weights (C). D Immunofluorescence staining of Ki67 and HER2 in tumor tissues; bar graphs show quantification of signal intensity. Scale = 50 μm. Data were reported as mean ± SD (n = 3). **P < 0.01, ***P < 0.001 vs. PBS group; ###P < 0.001 vs. CAFs-exo group Quantitative measurements of tumor volume over a six-week period (Fig. 9B) confirmed these observations. Tumor growth in the CAFs-exo group was significantly accelerated compared to PBS controls from week 3 onwards (P < 0.001). Trastuzumab treatment alone (PBS + Tra) resulted in a robust inhibition of tumor growth (P < 0.001). However, the growth-inhibitory effect was partially reversed in the CAFs-exo + Tra group (P < 0.001), suggesting that CAF-derived exosomes mediate a protective effect against trastuzumab. Consistently, tumor weight measurements at the end of treatment (Fig. 9C) revealed a significant increase in the CAFs-exo group (P < 0.001), and a dramatic decrease in the PBS + Tra group (P = 0.0015). The CAFs-exo + Tra group exhibited intermediate tumor weights, significantly higher than PBS + Tra (P < 0.001), supporting the notion that CAF-derived exosomes diminish trastuzumab sensitivity in vivo. To further examine tumor proliferation and HER2 expression, immunofluorescence staining for Ki67 and HER2 was performed on xenograft tumor sections (Fig. 9D). The CAFs-exo group showed markedly elevated Ki67 and HER2 expression levels relative to the PBS group (P < 0.001), indicating enhanced proliferative activity and HER2 stabilization. In contrast, the PBS + Tra group exhibited the lowest fluorescence intensities for both markers (P < 0.001), while the CAFs-exo + Tra group revealed partially restored Ki67 and HER2 levels (P < 0.001), further confirming the impact of CAF-derived exosomes on supporting tumor growth and HER2 expression under therapeutic pressure. Overall, the above findings indicate that CAF-derived exosomes compromise the antitumor effect of trastuzumab by sustaining HER2 protein levels and enhancing tumor proliferation, thereby promoting resistance to trastuzumab in vivo in HER2-positive breast cancer. Based on the above findings, a schematic working model summarizing the CAF-exo–YAP-USP8-HER2 regulatory axis is presented in Fig. 10. Schematic model illustrating the CAF-exo–YAP-USP8-HER2 signaling axis mediating trastuzumab resistance in HER2-positive breast cancer Discussion This study aimed to explore whether and how exosomes derived from CAFs contribute to trastuzumab resistance in HER2-positive breast cancer. Our results demonstrate that CAF-derived exosomes promote HER2 protein stabilization and tumor cell resistance to trastuzumab by inactivating the Hippo pathway, enhancing YAP nuclear translocation, and upregulating the transcription of the deubiquitinase USP8. This finding identifies a novel regulatory axis—CAF-exo → YAP → USP8 → HER2—that expands our understanding of the mechanisms underlying HER2-targeted therapy resistance. Previous studies have shown that CAF-derived exosomes influence tumor progression through various non-coding RNAs, such as miRNAs, lncRNAs, or circRNAs [20]. However, whether CAF-exosomes are involved in modulating protein stability or receptor tyrosine kinase pathway activity has been less studied. In this work, we observed that CAF-exo treatment markedly lowered LATS1 and YAP phosphorylation, suggesting that CAF-exosomes may suppress Hippo pathway activation. Subcellular fractionation further revealed enhanced nuclear localization of YAP in CAF-exo–treated cells, a phenomenon consistent with classical models of Hippo pathway inactivation. Furthermore, our RNase and Proteinase K protection assays demonstrated that disruption of exosomal membrane integrity abolished the Hippo-suppressive effect, indicating that intact intravesicular cargo, rather than surface-associated components, mediates YAP activation. Although it remains unclear whether specific upstream factors within CAF-exo directly modulate LATS1 or YAP phosphorylation, our data support the view that CAF-exosomes enhance YAP transcriptional activity by alleviating Hippo signaling constraints. YAP is increasingly recognized for its regulatory effects on the ubiquitin–proteasome system [21]. Several studies have shown that YAP can transcriptionally activate members of the USP family, including USP1 and USP9X [22]. Our luciferase reporter assay confirmed that YAP directly activates the USP8 promoter, and YAP knockdown resulted in a notable reduction in USP8 levels, indicating a transcriptional regulatory relationship. USP8 is a well-characterized deubiquitinase known to stabilize RTKs such as EGFR and MET by removing polyubiquitin chains [23]. In our study, CAF-exo treatment decreased HER2 ubiquitination and increased HER2 protein expression, and these effects were reversed by silencing either USP8 or YAP. These outcomes collectively indicate that HER2 may act as a downstream substrate of USP8, and its stabilization is likely mediated through a YAP–USP8 signaling axis. Consistent with our mechanistic findings, transcriptomic analysis of the TCGA-BRCA cohort revealed a positive correlation between USP8 and ERBB2 expression in the overall dataset, further supporting the molecular relevance of this regulatory axis in human breast cancer samples. Functionally, CAF-exo enhanced HCC1954 cell proliferation, colony formation, and IC50 of trastuzumab, while significantly reducing trastuzumab-induced apoptosis. Notably, similar pro-proliferative effects and reduced trastuzumab sensitivity were also observed in BT-474 cells, supporting the reproducibility of these findings in an additional HER2-positive model. These resistance-associated phenotypes were mitigated by knockdown of YAP or USP8, indicating that the survival advantage conferred by CAF-exosomes is dependent on this axis. Importantly, pharmacological inhibition of YAP using Verteporfin similarly reversed CAF-exosome–induced proliferation and restored trastuzumab sensitivity, further confirming the functional dependency of this resistance mechanism on YAP activation. In vivo, CAF-exo–induced tumor acceleration and its ability to impair trastuzumab efficacy were validated in a xenograft model, further supporting its role in facilitating therapeutic resistance within the tumor microenvironment. Compared with previous reports describing CAF-exosome–mediated drug resistance via PI3K/AKT or Notch pathways through non-coding RNAs such as miR-221/222 [24], this study presents a distinct mechanism wherein CAF-exosomes activate YAP transcription to induce USP8 expression, which in turn stabilizes HER2 protein. This not only expands the known functional scope of YAP in the tumor microenvironment but also provides the first evidence linking USP8-mediated HER2 stabilization to CAF-exosomal signaling. Although we did not directly compare CAF-exosome–induced resistance with established trastuzumab-resistant cell lines, several phenotypic and molecular features observed in our model resemble those reported in acquired resistance systems, including elevated IC50 values, sustained HER2 signaling, and enhanced YAP activation [12, 13]. These similarities suggest that CAF-derived exosomes may recapitulate key aspects of clinically relevant resistance through microenvironment-driven mechanisms. Notably, although YAP1 mRNA levels did not significantly predict overall survival in the TCGA cohort, YAP activity is predominantly regulated at the post-translational level through phosphorylation and subcellular localization. Therefore, transcriptomic datasets may not fully reflect the functional activation status of YAP signaling observed in our mechanistic experiments. Future studies directly comparing CAF-exo–treated cells with established trastuzumab-resistant derivatives would further clarify the extent of mechanistic convergence between microenvironment-mediated and tumor cell–intrinsic resistance pathways. Despite establishing a plausible mechanistic model, our investigation is subject to some limitations. First, although key findings were validated in two HER2-positive breast cancer cell lines (HCC1954 and BT-474), the mechanistic investigations were primarily conducted in HCC1954 cells. Therefore, further validation in additional HER2-positive models, including trastuzumab-resistant derivatives and patient-derived systems, would strengthen the translational relevance of this mechanism. Furthermore, the study did not directly compare CAF-exosome–induced resistance with established trastuzumab-resistant models. Such comparative analyses would help determine whether microenvironment-driven resistance converges mechanistically with intrinsic or acquired tumor cell resistance. Second, though the CAFs and exosomes analyzed were obtained from primary breast cancer tissues, patient-to-patient heterogeneity in CAF phenotypes and exosomal cargo may influence their effects. Moreover, although our cargo digestion assays suggest that Hippo pathway suppression depends on intravesicular components, the precise molecular species (e.g., specific miRNAs or proteins) remain unidentified, highlighting the need for comprehensive transcriptomic and proteomic profiling. Finally, the study focused on tumor cell-autonomous mechanisms, and the impact of CAF-exo on other tumor microenvironmental components, such as immune or endothelial cells, warrants further exploration to evaluate the translational potential of targeting this axis. In conclusion, our findings reveal a novel CAF-exo–YAP-USP8-HER2 axis that facilitates trastuzumab resistance in HER2-positive breast cancer. Targeting this regulatory pathway may represent a promising strategy to overcome microenvironment-induced resistance and improve the efficacy of HER2-targeted therapies. Conclusion This research reveals that CAF-derived exosomes promote trastuzumab resistance in HER2-positive breast cancer by inactivating the Hippo pathway, enhancing YAP-mediated transcription of USP8, and stabilizing HER2 protein through inhibition of its ubiquitination. These findings uncover a previously unrecognized CAF-exosome–YAP-USP8-HER2 regulatory axis that functionally links the tumor microenvironment with receptor stability and therapeutic resistance. Targeting this axis may provide a novel strategy for bypassing resistance to HER2-directed therapies and enhancing treatment results in HER2-positive breast cancer patients. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Author contributions WJY and LJ contributions to conception and design; LJ been involved in drafting the manuscript and revising it critically for important intellectual content; WJY made substantial contributions to acquisition of data; and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This study did not receive any funding in any form. Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate This study involving human participants was reviewed and approved by the Ethics Committee of Wujin Hospital Affiliated with Jiangsu University (Approval No. 2025-SR-422). All procedures were conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines. All animal experiments were approved by the Laboratory Animal Ethics Committee of Guangdong Medical Laboratory Animal Center (Approval No. D202502-7) and were conducted in accordance with national guidelines for the care and use of laboratory animals. Consent for publication All authors approved this manuscript. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement No datasets were generated or analysed during the current study.