Exosomal Oleic Acid Promotes Lymphangiogenesis and Nodal Metastasis in Cervical Cancer via the AKT/mTOR Pathway

ABSTRACT

Cervical cancer (CC) exhibits a pronounced tropism for regional lymphatic dissemination, a process driven by tumor‐associated lymphangiogenesis. While metabolites within the metastatic niche are increasingly recognized as determinants of organotropic metastasis, the role of exosome‐mediated metabolite transfer in tumor‐lymphatic endothelial cell (LEC) crosstalk remains largely unexplored. Here, we demonstrate that oleic acid (OA) is significantly enriched in both CC lymph node metastases and the peritumoral lymphatic microenvironment. Exosomes derived from highly metastatic CC cells actively package OA in a manner dependent on stearoyl‐CoA desaturase (SCD), the rate‐limiting enzyme of de novo fatty acid synthesis. Upon internalization by LECs, exosomal OA triggers the AKT/mTOR signaling axis, eliciting robust LEC proliferation and endothelial‐to‐mesenchymal transition (EndMT), thereby fostering lymphangiogenesis and nodal colonization. Knockdown of SCD abolishes these pro‐lymphangiogenic effects, a deficit fully reversed by the reconstitution of OA‐loaded exosomes. In vivo, exosomes from SCD‐silenced cells exhibit a severely compromised capacity to drive primary tumor growth, intratumoral lymphangiogenesis, and lymph node metastasis (LNM). Notably, free OA administration exerts substantially weaker effects than its exosomal counterpart, underscoring the superior efficiency of exosome‐mediated metabolite trafficking. Clinically, FASN and SCD expression are significantly upregulated in lymph node‐positive specimens and positively correlate with lymphatic vessel density and p‐AKT levels. Furthermore, circulating exosomal OA levels are significantly elevated in patients with nodal involvement, suggesting its potential as a non‐invasive diagnostic biomarker. Collectively, our findings establish a paradigm wherein tumor‐derived exosomes function as specialized vehicles for intercellular OA transfer, activating the AKT/mTOR pathway to license lymphangiogenic reprogramming. This work identifies exosomal metabolite shuttling as a central node in tumor‐lymphatic communication and proposes targeting OA synthesis or exosomal delivery as a promising therapeutic strategy against CC metastasis.

Cervical cancer exosomes deliver SCD‐dependent oleic acid to LECs, activating AKT/mTOR to drive VEGF‐independent lymphangiogenesis and nodal metastasis, identifying exosomal OA as a therapeutic target.

1. Introduction

Cervical cancer (CC) remains one of the most prevalent malignancies affecting women globally, with a particularly high burden in developing regions. Its elevated mortality is strongly linked to lymph node metastasis (LNM), a key prognostic indicator of disease progression and unfavorable clinical outcomes (Bray et al. 2024, Noguchi et al. 1987, Gien and Covens 2008). This metastatic process is fundamentally driven by tumor‐associated lymphangiogenesis‐the activation of lymphatic endothelial cells (LECs) and the subsequent formation of new lymphatic networks that facilitate tumor dissemination (Li and Li 2015, Karaman and Detmar 2014). While established mediators such as VEGF‐C and its receptor VEGFR‐3 are known predictors of reduced survival (García‐Caballero et al. 2017, Donnem et al. 2010), the multifaceted molecular mechanisms by which tumor cells systematically reprogram LECs remain incompletely understood.

Lipid metabolic reprogramming is a recognized hallmark of the tumor microenvironment (TME) that significantly contributes to malignancy (HR et al. 2023). Using spatial metabolomics, our group previously identified an abnormal accumulation of fatty acid metabolites—including oleic, palmitic, and stearic acids—within CC tissues. Notably, integrated overlay analysis with MSI‐HE stained sections revealed that oleic acid (OA) is preferentially localized within regions enriched with both tumor cells and lymphatic structures (Xu et al. 2025). This mirrors findings in other malignancies where specific phospholipids and bioactive bile acids have been linked to invasion and nodal colonization (Kawashima et al. 2020, Lee et al. 2019). These observations suggest that dysregulated fatty acid metabolism may actively drive lymphangiogenesis. We therefore hypothesize that tumor cells export these reprogrammed metabolites as bioactive signals to remodel the lymphatic niche.

A critical question remains: how are these metabolic signals precisely delivered to the lymphatic niche? Exosomes‐extracellular vesicles (30–150 nm)‐have emerged as pivotal mediators of intercellular communication. By transporting proteins, nucleic acids, and lipids, exosomes functionally reprogram recipient cells to facilitate angiogenesis, lymphangiogenesis, and pre‐metastatic niche formation (Zeng et al. 2018, Mashouri et al. 2019, Steinbichler et al. 2017). Studies indicate that LECs rapidly internalize exosomes and translocate them into the lymphatic lumen, enabling drainage to regional lymph nodes within minutes (García‐Silva et al. 2021, Leary et al. 2022). Furthermore, seminal work has demonstrated that exposure to tumor‐derived small extracellular vesicles (sEVs) triggers profound transcriptional changes in lymph nodes, including the upregulation of genes related to inflammation (TNF‐α), hypoxia (HIF‐1α), and lymphangiogenesis (VEGF) (Hood et al. 2011). For instance, sEVs carrying nerve growth factor receptor (NGFR) can activate MAPK and NF‐κB pathways in LECs, driving a pro‐lymphangiogenic switch (García‐Silva et al. 2021).

Despite extensive research into exosomal proteins (Srinivasan et al. 2016) and microRNAs (Zhou et al. 2021), the potential role of small‐molecule metabolites‐particularly lipid species‐as biologically active exosomal cargo remains largely unexplored. This gap is significant given our previous finding that OA is specifically enriched within peritumoral lymphatic regions (Xu et al. 2025). We hypothesize that these lipid metabolites are selectively packaged into tumor‐derived exosomes, acting as “metabolic messengers” that modulate LEC function.

In this study, we investigate whether CC cells exploit exosomes to deliver OA to LECs to drive lymphangiogenesis. Through functional screening and metabolomic profiling, we identified OA as a critically enriched metabolite in pro‐lymphangiogenic exosomes derived from MS751 cells. Utilizing integrated in vitro and in vivo assays, we demonstrate that exosomal OA activates the AKT/mTOR signaling axis in LECs, promoting vessel formation and metastatic dissemination. Our findings are corroborated by clinical data showing a significant correlation between elevated OA signatures and poor patient prognosis. Collectively, this study delineates a novel metabolic pathway for lymphatic reprogramming, offering new mechanistic insights and potential therapeutic targets for CC metastasis.

2. Materials and Methods

2.1. Terminology Notes

In this study, the term “exosomes” refers to sEVs (30–150 nm) isolated from conditioned media (CM) by differential ultracentrifugation at 100,000 g and characterized by multiple complementary techniques. Nanoparticle tracking analysis (NTA) confirmed the expected size distribution, transmission electron microscopy (TEM) visualized the characteristic cup‐shaped morphology, and Western blot analysis verified the presence of exosomal markers (CD9, CD63, TSG101) and the absence of the negative marker Calnexin, consistent with MISEV guidelines. For applications requiring enhanced purity, size‐exclusion chromatography was employed to complement the isolation procedure.

2.2. Human Sample s

Tumor tissue sections from CC patients were obtained from the Department of Pathology at the Xinjiang Medical University Affiliated Tumor Hospital. The study cohort included 28 patients diagnosed with lymph node metastasis‐negative (LNM−) CC and 67 patients diagnosed with lymph node metastasis‐positive (LNM+) CC. Additionally, serum samples from 8 LNM− and 8 LNM+ patients were collected for exosomal OA quantification. All tissue samples were independently verified by two experienced pathologists to confirm the diagnosis. Written informed consent was obtained from all participants. The study protocol was approved by the Ethics Committee of the Xinjiang Medical University (Approval No. IACUC‐20220304‐01).

2.3. Bioinformatics Analysis

Differential expression of FASN, CD36, FATP5, and SCD between CC (n = 306) and normal control groups (n = 13) was analyzed using the Gene Expression Profiling Interactive Analysis (GEPIA) platform (http://gepia.cancer‐pku.cn/). CC patients were stratified into high‐ and low‐risk groups based on the median expression values of FASN and LYVE1. Survival analysis was performed using Kaplan‐Meier Plotter (http://kmplot.com/analysis/), and statistical differences were assessed using the log‐rank test.

2.4. Cell Lines, Reagents, and Lentiviral Transfection

Human CC cell lines (MS751, HT‐3, HeLa, SiHa, C33A) and human lymphatic endothelial cells (HLECs) were purchased from Wuhan Procell Biotechnology. HLEC‐SV40T cells were obtained from Wuhan Shangen Biotechnology, and the cervical epithelial cell line HcerEpic was purchased from Qingqi (Shanghai) Biotechnology Co., Ltd. All cell lines were authenticated by STR profiling and confirmed negative for mycoplasma and viral contamination. MS751 cells were cultured in MEM supplemented with NEAA; HT‐3 cells were cultured in McCoy's 5A medium; HeLa, SiHa, C33A and HcerEpic cells were cultured in DMEM. All media were supplemented with 10% fetal bovine serum (FBS) and 100 U/mL penicillin/streptomycin. HLECs and HLEC‐SV40T cells were cultured in endothelial cell‐specific medium.

To generate stable gene knockdown models, MS751 cells were transduced with lentiviral particles harboring shRNA targeting human SCD (shSCD), shRNA targeting human RAB27A (shRAB27A), or non‐targeting control shRNA (shNC) at a multiplicity of infection (MOI) of 30. The three independent shRNA target sequences for SCD were: hSCD‐shRNA1: ACGACATTCGCCCTGATATAA; hSCD‐shRNA2: CACATGCTGATCCTCATAATT; hSCD‐shRNA3: GCCATCTTGGCCAGGATTAAA. The three independent shRNA target sequences for RAB27A were: h‐Rab27a‐shRNA1: GTCTGATGGAGATTATGATTA; h‐Rab27a‐shRNA2: CCAGTGTACTTTACCAATATA; h‐Rab27a‐shRNA3: ATAGCACTCGCAGAGAAATAT. The negative control shRNA‐NC sequence was: ACTACCGTTGTTATAGGTGT. Stable polyclonal populations were selected using 2 µg/mL puromycin for 7–10 days. Knockdown efficiencies of SCD and RAB27A were validated by Western blot.

2.5. Cell Culture Treatment

When indicated, CC cells were treated with the following compounds: OA (0‐800 µM, NO. HY‐N1446, MedChemExpress), GW4869 (20 µLof 1 µg/µL, NO. 6823‐69‐4, Merck). After treating CC cells with OA for 48h, the culture medium was removed, and the cells were washed twice with PBS. Subsequently, exosome‐free serum medium was added. After 48 h, the conditioned medium (CM) was collected for exosome extraction. When the CC cells were in good growth condition, they were washed twice with PBS, and exosome‐free serum medium was added. In the experimental group, 20 µL of a 1 µg/µL GW4869 solution was added. After 48 h, the CM was collected.

For signaling pathway experiments, LECs were seeded 1 day before inhibitor and Exo‐OA treatment. Inhibitors were added 60 min before Exo‐OA treatment. The following inhibitors were used Miransertib (NO.ARQ‐092, MedChemExpress, 1 µM). For LECs treated with exosomes for a long time (such as for functional analysis such as EMT marker determination and colony formation), cells were treated with 5 µg/mL exosomes for another 2–7 days after cell inoculation. For short‐term treatment of LECs with exosomes (signal activation assay), cells were treated with 5 µg/mL exosomes for 24 h.

2.6. Isolation and Characterization of Exosomes

Exosomes were isolated from CM using differential ultracentrifugation. CM was centrifuged at 500 × g (10 min) and 12,000 × g (20 min) to remove debris. Exosomes were then pelleted at 100,000 × g for 70 min, washed in PBS, and pelleted again. For high‐purity applications, size‐exclusion chromatography (Exo‐spin) was utilized. Exosome morphology was visualized using TEM (FEI Tecnai T20), and size distribution was determined by Nanoparticle Tracking Analysis (NTA) (NanoFCM U30).

2.7. Exosome Internalization Assay

Exosomes (50 µg) were labeled with PKH26 Red Fluorescent dye (Sigma‐Aldrich) for 5 min. The reaction was stopped with exosome‐free serum, and labeled exosomes were washed with PBS. HLECs were co‐cultured with PKH26‐labeled exosomes for 4 h. Cells were then fixed, permeabilized with 0.2% Triton X‐100, and stained with CoraLite Plus 488‐Phalloidin (actin filaments) and Hoechst 33342 (nuclei) before confocal imaging (Nikon).

2.8. Western Blot

Total protein was isolated from LECs using radioimmunoprecipitation assay (RIPA) lysis buffer (R0010, Solarbio, Beijing, China). The protein concentration was determined with bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA). The protein was quantified to 30 µg by adding the loading buffer and PBS in proportion. Then proteins were subjected to 10% sodium dodecyl sulfate (SDS)‐polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes. The membranes were blocked with 5% non‐fat milk and probed with specific antibodies overnight at 4°C. After washing three times in TBST (containing 0.1% Tween 20), the membranes were probed with anti‐mouse or anti‐rabbit immunoglobulin coupled to horseradish peroxidase for 2 h at room temperature. The signals were visualized with an enhanced chemiluminescence detection kit (NO. BL523aA, Biosharp). The gray value of the target protein was imaged and used to calculate target protein expression. The primary and secondary antibodies used in this study are listed in Table S1.

2.9. Metabolomic Analysis

LC‐MS analysis was performed using a Waters ACQUITY UPLC I‐Class plus coupled with a Thermo QE plus high‐resolution mass spectrometer. Exosomes were treated with a methanol‐acetonitrile protein precipitant, vortexed, and centrifuged. The supernatant was filtered (0.22 µm) and analyzed on an ACQUITY UPLC HSS T3 column. Data were processed using Progenesis QI v3.0. For clinical samples, OA concentrations in serum exosomes were quantified using an OA detection kit (Beyotime) via the absorbance method at 570 nm.

2.10. Determination of an Optimal Free OA Concentration and Validation of Concentration Equivalence between Exosomes and Free OA

Free OA was serially diluted in complete culture medium to final concentrations of 0, 10, 50, 100, 150, 200, 400, and 800 µM. HLEC and HLEC‑SV40T cells were then treated with each concentration in triplicate wells. After 24 h of culture, 10 µL of CCK‑8 reagent was added to each well, followed by incubation for 2 h, and the absorbance was measured at 450 nm.

Exosomes were isolated from the culture supernatant of HCerEpiC cells. Purified exosomes were mixed with FITC‑labeled OA at concentrations of 0, 10, 50, 100, 150, 200, and 250 µM, and then subjected to three cycles of freeze‑thaw loading: freezing in liquid nitrogen for 5 min followed by thawing in a 37°C water bath for 3 min per cycle. After the final thaw, the mixture was immediately placed in liquid nitrogen. Free (unloaded) FITC‑OA was removed by ultracentrifugation, and the exosome pellets were resuspended in PBS.

Equal aliquots of the exosome suspensions were used to measure fluorescence (excitation 495 nm/emission 520 nm). A standard curve for free FITC‑OA (0–250 µM) was generated in parallel. The fluorescence intensity of exosome‑associated FITC‑OA was plotted against the corresponding loading concentration, and the lowest concentration at which the fluorescence intensity reached a plateau was defined as the saturation concentration for loading.

2.11. In Vitro Functional Assays

2.11.1. Tube Formation Assay

LECs were seeded onto 96‐well plates pre‐coated with 60 µL/well of Matrigel (Corning, USA) at a density of 1.5 × 10⁴ cells/well, and treated with CC cell‐derived CM or exosomes (5 µg/mL) for 16 h. Tube formation was imaged under an inverted phase‐contrast microscope (Nikon E400). The total tube length and number of branch points were quantified from three randomly selected fields per well using ImageJ software.

2.11.2. Colony Formation Assay

LECs (100 cells/well) were seeded in 6‐well plates and treated with exosomes (5 µg/mL) for 7–21 days, with medium replaced every 3 days. When visible colonies formed, cells were fixed with 4% paraformaldehyde for 30 min and stained with 0.5% crystal violet (Beijing Solar Biotech, China). Colonies containing >50 cells were counted manually and verified using ImageJ.

2.11.3. Cell Proliferation Assay (CCK‐8)

LECs were seeded in 96‐well plates (2 × 10³ cells/well) and treated with exosomes (5 µg/mL) for 5 days. At 24 and 48 h post‐treatment, 10 µL of CCK‐8 reagent (Proteintech, China; Cat# PF00004) and 90 µL of culture medium were added to each well. After incubation at 37°C for 1.5 h in the dark, absorbance at 450 nm was measured using a Multiskan FC microplate reader (Thermo Fisher Scientific, USA).

2.11.4. Wound Healing (Scratch) Assay

LECs pre‐treated with exosomes for 48 h were seeded into 6‐well plates and cultured to full confluence. A linear scratch was introduced across the monolayer using a sterile 200‐µL pipette tip. After gentle washing with PBS to remove detached cells, the medium was replaced with low‐serum medium (1% FBS) to minimize proliferation effects. Wound closure was photographed at 0 and 48 h under an inverted microscope (Nikon E400). The relative wound width was measured using ImageJ, and the migration rate was calculated as the percentage of wound closure.

2.11.5. Transwell Invasion Assay

Cell invasion was assessed using 24‐well Transwell chambers (8‐µm pore size; Corning, USA) pre‐coated with 50 µL of diluted Matrigel (1:8 in serum‐free medium). LECs (2 × 10⁴ cells/well) were seeded in the upper chamber in 200 µL of serum‐free medium, while 600 µL of complete medium containing 10% FBS was added to the lower chamber. After 24 h of incubation at 37°C, non‐invaded cells on the upper surface were gently removed with a cotton swab. Invaded cells on the lower surface were fixed with 4% paraformaldehyde for 30 min, stained with 0.1% crystal violet for 20 min at room temperature, and rinsed with PBS. The number of invaded cells was counted in five randomly selected fields per insert under an inverted microscope (Nikon E400).

2.12. Nude Mouse Popliteal LNM Model

Female BALB/c nude mice (3–4 weeks old) were injected in the right hind footpad with 5 × 10⁵ MS751 cells. From day 11, 30 µg of isolated exosomes were injected intratumorally every 3 days. On day 28, mice were euthanized via cervical dislocation. Primary tumors and inguinal lymph nodes were harvested for immunohistochemical (IHC) and immunofluorescence (IF) analysis. Animal procedures were approved by the Xinjiang Medical University Animal Ethics Committee (No. IACUC‐20240711‐54).

2.13. IF Analysis

Tissue samples are fixed with formalin, then dehydrated, and finally embedded in paraffin. Using a microtome, the paraffin block is cut into sections typically 4 micrometers thick. The sections are dried in an oven at 60°C, followed by treatment with xylene for deparaffinization three times, each for 10 min, to remove the paraffin. The sections are then sequentially immersed in different concentrations of ethanol solution (100%, 85%, 75%), each for 5 min, and finally rinsed with distilled water. According to the requirements of the antibody (citrate buffer or Tris‐EDTA buffer for section pretreatment), the sections are heated in a 95°C water bath for 30 min. A blocking solution (NO.ZLI‐9056, ZSGB‐BIO, Beijing, China) is added and incubated at room temperature for 30 min. A diluted specific primary antibody is applied to the sections and incubated overnight at 4°C. The next day, the sections are washed three times with PBS, and a fluorescently labeled secondary antibody corresponding to the primary antibody is added and incubated under dark conditions (37°C for 1 h). The next day, wash the slices with PBS three times, then add the corresponding fluorescently labeled secondary antibody to the primary antibody, and incubate in the dark (37°C, 1 h). Wash the slices with PBS buffer three times, then incubate with 1 mg/mL Hoechst 33342 dye for 10 min and wash with PBS. Seal the slices with a mounting medium containing antifade agent (NO. G1401‐5ML, Servicebio, Wuhan, China) to prevent fluorescence signal decay. Observe the slices under a fluorescence microscope (Zeiss Airyscan2, ZEISS, Germany) capture images, and analyze the results. The primary and secondary antibodies used in this study are listed in Table S1

2.14. IHC Analysis

Paraffin sections were immersed in fresh xylene for 10 min × 3 times; after removing excess liquid, they were placed in absolute ethanol for 3 min × 3 times; after removing excess liquid, they were placed in 95% ethanol for 3 min × 2 times; after removing excess liquid, they were placed in 75% ethanol for 3 min × 2 times; rinsed with distilled water for 1 min, and placed in PBS buffer. Microwave heating was used for antigen retrieval, and an endogenous peroxidase blocking agent was added, followed by incubation at room temperature for 10 min; PBS buffer was rinsed for 3 min × 3 times. 100 µL or an appropriate amount of primary antibody was added, and incubated at 37°C for 60 min; PBS buffer was rinsed for 3 min × 3 times. 100 µL or an appropriate amount of enzyme‐labeled goat anti‐mouse/rabbit IgG polymer (NO. PV‐6000, ZSGB‐BIO) was added, and incubated at 37°C for 20 min; PBS buffer was rinsed for 3 min × 3 times. An appropriate amount of freshly prepared DAB chromogenic solution was added, and incubated at room temperature for 3 min, followed by rinsing with tap water, then hematoxylin staining was performed for 20 s, followed by rinsing with tap water for 15 min. Finally, dehydration, clearing, and neutral resin mounting were performed. The primary antibodies used in this study are listed in Table S1

2.15. Statistical Analysis

Data are expressed as mean ± SD from at least three independent experiments. Statistical significance was analyzed using GraphPad Prism 9 and SPSS 26.0. Comparisons between two groups were performed using unpaired t‐tests; multiple groups were compared via one‐way ANOVA with Tukey's post hoc test. Survival was analyzed using Kaplan‐Meier curves and the log‐rank test. P < 0.05 was considered statistically significant.

3. Results

3.1. CC‐Derived Exosomes Potently Enhance the Lymphangiogenic Function of LECs

To assess the influence of CC cells on LEC activity, we utilized a comparative model system comprising HPV‐positive, highly metastatic MS751 cells; HPV‐negative, highly metastatic HT‐3 cells; HPV‐negative, low metastatic C33A cells; and normal cervical epithelial cells (HCerEpiC). CM were harvested from these cells and evaluated for their biological effects on LECs (Figure 1A). Tube formation assays revealed that both MS751‐CM and HT‐3‐CM significantly augmented the tubulogenic capacity of HLECs, leading to networks with greater density and branching complexity relative to HCerEpiC‐CM (Figure 1B, C ). Consistent results were obtained in both HLEC and HLEC‑SV40T cells (Figure S1A). In contrast, C33A‑CM exerted only a very weak effect on LECs tube formation, with no statistical difference compared to HCerEpiC‑CM. Notably, CM from the HPV‐positive MS751 line induced the most potent stimulatory effect among all tested cell lines. This trend was further confirmed by subsequent functional analyses: MS751‑CM and HT‐3‑CM both enhanced the proliferative activity (Figure 1D, Figure S1B), colony formation (Figure S1C‑E), as well as invasion and migration (Fig. S1F–H) of HLEC and HLEC‑SV40T cells, with MS751‑CM consistently eliciting the strongest response across all measured parameters (Figure 1B‑D, Figure S1A‑H). C33A‑CM showed no significant promoting effect on any of the above functional parameters. These findings indicate that CC cells, and specifically MS751 cells, promote LECs activation through secreted factors. Due to its dominant phenotypic profile, the MS751 cell line was selected for further investigation into the mechanistic basis of its pro‐lymphangiogenic activity.

FIGURE 1.

Cervical cancer cells activate lymphatic endothelial cells by secreting exosomes with pro lymphangiogenic activity

(A) Experimental scheme. Conditioned medium (CM) was collected from MS751, HT‐3, C33A, and normal cervical epithelial cells (HCerEpiC) and used to treat primary human lymphatic endothelial cells (HLECs) or immortalized HLEC‐SV40T cells. (B‐C) Quantitative analysis of tube formation (tube density and number of branch points) in HLECs treated with the indicated CM for 8 h. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.01 vs. HCerEpiC‐CM; ##P < 0.01, ###P < 0.001 vs. C33A‐CM; &&P < 0.01, &&&P < 0.001 vs. HT‐3‐CM. (D) Proliferation of lymphatic endothelial cells treated with the indicated CM for 24 and 48 h, assessed by CCK‐8 assay. Data are presented as mean ± SD from three independent experiments; two‐way ANOVA. Statistical annotations are the same as in (B‐C). (E) Transmission electron microscopy image showing the typical cup‐shaped morphology of cell‐derived exosomes. Scale bar, 500 nm. (F)Confocal microscopy images. PKH26‐labeled exosomes (red) were internalized into the cytoplasm of HLECs after 4 h of co‐incubation. Nuclei were counterstained with Hoechst 33342 (blue). Scale bar, 10 µm.(G‐H) Tube formation, invasion, and migration abilities of HLECs under different treatments. Groups: CM‐MS751 + DMSO, CM‐MS751 + GW4869 (exosome release inhibitor), intact CM‐MS751, and exosome‐depleted conditioned medium (CM‐Exo‐depleted). (G) Upper panels: representative images and quantitative analysis of tube formation. Lower panels: quantitative analysis of Transwell invasion and migration assays. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. ***P < 0.001 vs. CM‐MS751 + DMSO; ###P < 0.001 vs. CM‐MS751. Scale bar, 100 µm. (H) Quantitative results of the invasion and migration assays (see panel G for details). (I‐J) HLECs treated with equal protein amounts of exosomes derived from MS751, HT‐3, C33A, and HCerEpiC cells. (I) Representative images and quantitative analysis of tube formation. (J) Representative images and quantitative analysis of invasion ability. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. **P < 0.01, ***P < 0.001 vs. HCerEpiC‐Exo; #P < 0.05, ##P < 0.01, ###P < 0.001 vs. C33A‐Exo; &P < 0.05, &&&P < 0.001 vs. HT‐3‐Exo. Scale bar, 100 µm. (K‐L) Immortalized HLEC‐SV40T cells treated with equal protein amounts of exosomes from the same four cell lines. (K) Representative images and quantitative analysis of tube formation. (L) Representative images and quantitative analysis of invasion ability. Statistical annotations are the same as in (I‐J). Scale bar, 100 µm.

We next focused on exosomes—key intercellular communicators with a diameter of ∼30–150 nm—which mediate TME remodeling by shuttling bioactive cargo such as proteins, lipids, and nucleic acids. Exosomes were isolated from the CM of MS751, HT‐3, C33A, and HCerEpiC cells and rigorously characterized. TEM confirmed the presence of vesicles with classic cup‐shaped morphology (Figure 1E, Figure S1I). Nanoparticle tracking analysis indicated a primary size distribution peak between 100 and 130 nm (Figure S1J). Western blot analysis verified the expression of canonical exosomal markers (CD9, CD63, CD81, Tsg101, Hsp70) and the absence of the endoplasmic reticulum contaminant Calnexin (Figure S1K), collectively affirming the successful isolation of high‐purity exosome preparations. Importantly, quantitative assessment revealed no significant difference in exosome secretion levels among the four cell types (Figure S1J), suggesting that the potent activity of MS751‐CM is linked to its exosomal cargo composition rather than quantity. Furthermore, confocal microscopy demonstrated efficient internalization of fluorescently labeled MS751‐derived exosomes by both HLEC and HLEC‑SV40T cells (Figure 1F, Figure S1I‑M), providing direct visual evidence for exosome uptake and potential functional modulation.

To definitively establish the necessity of exosomes in the pro‐lymphangiogenic activity of MS751‐CM, we employed two complementary strategies: pharmacological inhibition (GW4869) and genetic blockade (Rab27a knockdown). First, MS751 cells were treated with GW4869; a schematic diagram of the experimental workflow is shown in Figure S2A. NTA revealed that GW4869 reduced the number of exosome particles in the CM by approximately 53% (Figure S2C). In parallel functional experiments, compared with the vehicle control group (CM‑MS751+DMSO), the GW4869‑treated group (CM‑MS751+GW4869) significantly attenuated the ability of MS751‑CM to stimulate HLEC tube formation (Figure 1G), proliferation (Figure S2B, D), as well as invasion and migration (Figure 1H, Figure S2E). Similarly, physical depletion of exosomes by ultracentrifugation (CM‑Exo‑depleted) resulted in a comparable loss of pro‑lymphangiogenic activity relative to complete CM (CM‑MS751) (Figure 1G‑H, Figure S2B‑E). Second, we generated a stable Rab27a‑knockdown MS751 cell line (shRAB27A, Figure S2F). Electron microscopy and NTA confirmed that Rab27a knockdown markedly reduced exosome secretion (Figure S2G‑H). Exosomes isolated from shRAB27A and control cells were verified by Western blotting to express canonical exosomal markers (CD9, CD63, Tsg101, etc.), confirming successful exosome isolation (Figure S2I ). Incubation of HLEC and HLEC‑SV40T cells with shRAB27A‑derived exosomes significantly inhibited their tube formation, invasion, and migration capacities (Figure S2J‑K). These consistent results obtained from independent approaches (GW4869 treatment, ultracentrifugation depletion, and Rab27a knockdown) demonstrate that exosome secretion is essential for the pro‑lymphangiogenic activity of MS751‑CM. Finally, we directly compared the functional potency of purified exosomes derived from different cell sources. HLEC and HLEC‑SV40T cells were treated with equal protein amounts of exosomes derived from MS751 (MS751‑Exo), HT‑3 (HT‑3‑Exo), C33A (C33A‑Exo), or HCerEpiC (HCerEpiC‑Exo) cells. The results showed that MS751‑Exo exerted the most robust enhancement of LEC tube formation (Fig. 1K‑L, S2L‑M), outperforming C33A‑Exo, HT‑3‑Exo, and HCerEpiC‑Exo. Taken together, these data demonstrate that exosomes secreted by MS751 cells possess intrinsically potent pro‑lymphangiogenic activity and directly enhance the functional capacity of LECs.

3.2. Metabolomic Profiling Identifies OA as a Key Differential Metabolite Enriched in MS751 Exosomes

Since exosomes derived from CC cells (MS751‑Exo and C33A‑Exo) and normal epithelial cells (HCerEpiC‑Exo) exhibited comparable physical characteristics (Figure S1J), the distinct functional impact on LECs was postulated to originate from differences in their molecular cargo. To systematically delineate the metabolic profiles of exosomes from different origins, untargeted metabolomic analysis was performed on MS751‑Exo, C33A‑Exo, and HCerEpiC‑Exo. The quality of the metabolomic data was assessed by principal component analysis (PCA), which revealed tight clustering within groups and clear separation among the three groups, indicating high data reliability and intergroup comparability (Figure 2A). Typical total ion chromatograms obtained by liquid chromatography‑mass spectrometry exhibited well‑resolved peaks with satisfactory separation, confirming the stability and reproducibility of the analytical system (Figure 2B). Based on these data, the metabolite profiles of the three exosome types were compared. Compared with C33A‑Exo, a total of 136 significantly altered metabolites were identified in MS751‑Exo, among which 29 were upregulated and 107 were downregulated (Figure 2C, E). KEGG pathway enrichment analysis of these differentially abundant metabolites revealed that lipid metabolism‑related pathways were significantly enriched (Figure 2D). These results indicate that, compared with exosomes derived from low‑metastatic C33A cells, exosomes derived from highly metastatic MS751 cells exhibit a more pronounced reprogramming and enrichment of fatty acid‑associated metabolites, suggesting a potential functional role in the pro‑lymphangiogenic activity of tumor‑derived exosomes.

FIGURE 2.

Metabolomic Profiling Identifies Oleic Acid as a Discriminatory Metabolite Enriched in MS751‐Derived Exosomes

(A) Principal component analysis (PCA) score plot demonstrating clear separation of metabolic profiles among the MS751‐Exo, C33A‐Exo, and HCerEpiC‐Exo groups, indicating strong intergroup comparability. (B) Representative total ion chromatogram obtained by liquid chromatography‐mass spectrometry (LC‐MS). (C) Volcano plots of differential metabolites. Left: MS751‐Exo vs. C33A‐Exo; right: MS751‐Exo vs. HCerEpiC‐Exo. Red indicates significantly upregulated metabolites, blue indicates significantly downregulated metabolites (criteria: fold change ≥ 2 and P < 0.05, as determined by appropriate statistical test—please specify if different). (D) KEGG pathway enrichment analysis showing the top 10 most significantly enriched pathways among the differential metabolites. Lipid metabolism‐related pathways are predominantly represented. (E‐F) Differential analysis of fatty acid subclasses. (E) Volcano plot of fatty acid metabolites (comparison as in C). (F) Comparison of signal intensities (or relative abundance) of selected fatty acids: oleic acid (OA), phosphatidylethanolamine (PE), palmitic acid (PA), and stearic acid (SA). Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 (specific pairwise comparisons should be indicated—adjust as needed). (G) Absolute quantification of oleic acid in exosomes from the three indicated cell lines, measured using a fatty acid assay kit. The oleic acid content in MS751‐Exo is significantly higher than that in HCerEpiC‐Exo and C33A‐Exo. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. **P < 0.01 vs. HCerEpiC‐Exo; ###P < 0.001 vs. C33A‐Exo.

Within these significantly enriched fatty acid metabolism pathways, we further examined differentially expressed fatty acid molecules. Volcano plot analysis revealed that OA was among the most markedly upregulated metabolites in MS751‑Exo, whereas stearic acid (SA) and PE showed relatively modest increases (Figure 2F). To quantitatively validate these differences, specific fatty acid quantification assays were performed on MS751‑Exo, C33A‑Exo, and HCerEpiC‑Exo samples. The results confirmed that OA levels were significantly elevated in MS751‑Exo compared to HCerEpiC‑Exo (Figure 2G ); moreover, OA levels were also significantly higher in MS751‑Exo than in C33A‑Exo. As the most abundant monounsaturated fatty acid in humans, OA serves as a key precursor for phospholipid synthesis in cell membranes and participates in energy storage and signal transduction. Given its markedly elevated abundance in MS751‑Exo, OA was identified as a critical differential metabolite in tumor‑derived exosomes and was selected as the core focus for subsequent functional mechanistic studies.

3.3. Exosome‐Delivered OA Functions as a Critical Effector in Lymphangiogenesis

To establish the intrinsic bioactivity of OA, HLEC and HLEC‑SV40T cells were treated with increasing concentrations of free OA within the non‑cytotoxic range (0–150 µM) determined by CCK‑8 assay (Figure S3A). The results showed that OA significantly and dose‑dependently enhanced tube formation (Figure 3A), proliferation (Figure 3B, Figure S3B), migration (Figure 3C), and invasion (Figure S3C) in both cell types, confirming OA as a potent pro‑lymphangiogenic metabolite. To investigate whether OA serves as a key functional component conferring exosomal activity, normal cervical epithelial cell‑derived exosomes (HCerEpiC‐Exo) were mixed with 50 µM FITC‑labeled OA (FITC‐OA) and loaded using either repeated freeze‑thaw cycles or conventional co‑incubation. The loading efficiencies of the two methods showed no significant difference (Figure S3D). Therefore, the freeze‑thaw method was selected for subsequent experiments. To further optimize loading conditions, exosomes were mixed with various concentrations (0–250 µM) of FITC‐OA and subjected to repeated freeze‑thaw cycles. By measuring and comparing the fluorescence intensity in the exosomal fractions, a plateau was observed when the loading concentration reached 150 µM, indicating the saturating loading concentration (Figure 3D).

FIGURE 3.

Exosomal delivery of oleic acid promotes lymphangiogenesis and its serum level positively correlates with lymph node metastasis

(A–C) Free oleic acid (OA) enhances lymphatic endothelial cell functions in a dose‐dependent manner. (A) Representative images and quantitative analysis of tube formation in HLECs and HLEC‐SV40T cells treated with increasing concentrations of OA (0–400 µM). (B) Proliferation of HLECs and HLEC‐SV40T cells assessed by CCK‐8 assay after OA treatment. (C) Representative images and quantitative analysis of Transwell invasion assay. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 vs. 0 µM; #P < 0.05, ##P < 0.01, ###P < 0.001 vs. 15 µM; &P < 0.05, &&P < 0.01 vs. 30 µM. Scale bar, 100 µm (applies to A and C). Determination of the saturation concentration for OA loading into exosomes. HCerEpiC‐derived exosomes were loaded with increasing concentrations of FITC‐OA (0–250 µM). Fluorescence intensity of the exosomal fractions was measured, identifying 150 µM as the saturation loading concentration (the lowest concentration at which fluorescence reaches a plateau). (E–H) Functional comparison of OA‐loaded exosomes versus free OA and tumor‐derived exosomes. (E) Representative images and quantitative analysis of tube formation in HLECs and HLEC‐SV40T cells treated with PBS‐Exo (control), OA‐Exo (OA‐loaded HCerEpiC exosomes), Tumor‐Exo (MS751‐derived exosomes), or free FITC‐OA (at equivalent concentration). (F) Quantitative results of tube formation (branch points and total tube length) for the same treatments (details as in E). (G) Representative images and quantitative analysis of HLEC invasion after the indicated treatments. (H) Quantitative analysis of HLEC migration after the indicated treatments. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 vs. PBS‐Exo; ##P < 0.01, ###P < 0.001 vs. free FITC‐OA. Scale bar, 100 µm (applies to E and G). (I–J) Effects of exosomes derived from SCD‐1‐knockdown cells. (I) Quantitative analysis of tube formation in HLECs and HLEC‐SV40T cells treated with shNC‐Exo (control), shSCD‐1‐Exo (exosomes from SCD‐1‐knockdown MS751 cells), or OA‐Exo (shSCD‐1) — that is, OA‐loaded exosomes from SCD‐1‐knockdown cells. (J) Quantitative analysis of invasion ability of HLECs and HLEC‐SV40T cells under the same treatment conditions. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 vs. PBS group; ##P < 0.01, ###P < 0.001 vs. shNC‐Exo; &P < 0.05, &&P < 0.01, &&&P < 0.001 vs. shSCD‐1‐Exo. Scale bar, 100 µm (applies to representative images, if included). (K–M) Analysis of serum exosomes from cervical cancer patients with or without lymph node metastasis (LNM). (K) Western blot analysis of serum exosomal markers CD9, TSG101, and CD81 in LNM‐positive (LNM+) and LNM‐negative (LNM‐) patients. (L) Nanoparticle tracking analysis showing size distribution of serum exosomes from LNM+ and LNM‐ patients. (M) Targeted metabolomic quantification of oleic acid (OA) content in serum exosomes and exosome‐depleted serum supernatant from the two patient groups (n  =  8 per group). Data are presented as mean ± SD; Student's t‐test. ***P < 0.001; ns, not significant.

The following control groups were included: untreated exosomes (PBS‐Exo, i.e., HCerEpiC‐Exo incubated with PBS), positive control exosomes (Tumor‐Exo, i.e., MS751‐Exo), and free FITC‐OA treatment (150 µM free OA directly added to the culture medium). Functional assays demonstrated that, compared with the negative control PBS‐Exo, OA‐Exo treatment significantly enhanced the proliferative activity (Figure S3E), tube formation capacity (Figure 3E‑F, Figure S3F), as well as migration and invasion abilities (Figure 3G‑H, Figure S3G‑H) of HLEC and HLEC‑SV40T cells. The pro‑lymphangiogenic effect of OA‐Exo was comparable to that of the positive control Tumor‐Exo. Notably, although free OA treatment also partially enhanced LEC functions, its effect was weaker than that of OA‐Exo (Figure 3E‑H, Figure S3E‑H). This comparison clearly indicates that exosomes, as delivery vehicles, significantly improve the bioavailability and functional efficacy of OA, thereby activating lymphangiogenesis more efficiently.

To elucidate the necessity of OA in tumor exosome‑mediated lymphangiogenesis, we knocked down stearoyl‑CoA desaturase (SCD), the key rate‑limiting enzyme in OA synthesis (Figure S3I). SCD knockdown (shSCD‑1, shSCD‑2) significantly reduced OA levels in MS751 cells and their secreted exosomes (shSCD‑1‑Exo and shSCD‑2‑Exo) (Figure S3J, L). Accordingly, compared with control exosomes (shNC‑Exo), treatment with shSCD‑1‑Exo resulted in comprehensive functional impairment of HLEC and HLEC‑SV40T cells, as evidenced by reduced tubulogenic capacity and branch points (Figure 3I) and significantly diminished invasion ability (Figure 3J, Figure S3K). These results demonstrate that endogenous OA synthesis in tumor cells is essential for the pro‑lymphatic endothelial function of their secreted exosomes. To confirm that this functional loss is specifically attributable to OA reduction, rescue experiments were performed. Exosomes were collected from SCD‑knockdown MS751 cells cultured in medium exogenously supplemented with OA, referred to as OA‑Exo (shSCD‑1) (Figure S3M). Metabolic quantification confirmed restored OA levels in these exosomes (Figure S3L). Consistently, OA‑Exo (shSCD‑1) treatment effectively rescued the lymphangiogenic capacity (Figure 3I), invasion ability (Figure 3J, Figure S3J), and migration ability (Figure 3N) of HLEC and HLEC‑SV40T cells. This rescue experiment directly demonstrates that the functional deficits of exosomes resulting from SCD knockdown are mediated by reduced OA content.

Collectively, these in vitro experiments establish that exosome‐delivered OA is a key effector molecule in promoting lymphangiogenesis. To validate this finding at the clinical level, serum samples were collected from CC patients with positive (LNM+) or negative (LNM−) LNM, and exosomes were isolated from these samples. Isolated exosomes were characterized by Western blot, which confirmed the expression of canonical markers CD9, TSG101, and CD81 (Figure 3K), and nanoparticle tracking analysis revealed a predominant size distribution between 43 and 100 nm (Figure 3L), consistent with exosomal features. Importantly, quantitative analysis of OA in serum exosomes from the two patient groups revealed significantly higher OA content in exosomes from LNM+ patients compared with those from LNM− patients (Figure 3M). This difference was specific, as free OA levels in the corresponding serum supernatants (post‐exosome depletion) did not differ significantly between the two groups (Figure 3M). These findings directly demonstrate that enrichment of OA in circulating exosomes is positively correlated with LNM status in CC.

3.4. CC Exosomes Activate the AKT/mTOR Pathway in LECs through Exosome‐Delivered OA

To directly trace exosomal delivery of OA, a dual‐color fluorescence colocalization system was established. FITC‐OA was loaded into exosomes via repeated freeze–thaw cycles, and exosomal membranes were simultaneously labeled with PKH26, generating dual‐labeled exosomes (Exo‐PKH26/FITC‐OA). Following a 4‐h co‐incubation with LECs, confocal microscopy revealed abundant red punctate signals corresponding to exosomes and widely distributed green OA signals within the cytoplasm of LECs treated with Exo‐PKH26/FITC‐OA. Critically, merged images exhibited prominent yellow colocalization puncta (Figure 4A), directly demonstrating that exosomes deliver OA into recipient cells. Rigorous control experiments excluded nonspecific signals: Exo‐PKH26 (membrane labeling only) displayed only red signals without green fluorescence, whereas free FITC‐OA treatment yielded only faint, diffuse green background (Figure S4A). Together, these results confirm that CC‐derived exosomes serve as functional vehicles that efficiently and specifically deliver OA into LECs. To investigate the downstream mechanism of MS751‐Exo, we first analyzed the metabolomics data. KEGG pathway enrichment analysis showed that the differentially expressed lipid metabolites in MS751‐Exo were significantly enriched in the “mTOR signaling pathway” (Figure 4B), suggesting potential activation of this pathway. To validate this finding and assess its specificity, LECs were treated with PBS, C33A‑Exo, or MS751‑Exo, and signaling pathway changes were examined within 60 min. Western blot analysis revealed that only MS751‑Exo significantly upregulated the phosphorylation levels of AKT (Ser473), mTOR (Ser2448), and its downstream effector S6K (Thr389), whereas C33A‑Exo and PBS showed no significant differences (Figure 4C, Figure S4B). Moreover, because the ERK pathway is also commonly involved in metabolic regulation, we examined the MAPK pathway (ERK1/2 phosphorylation) and found no differences among the three groups (Figure 4C, Figure S4B). These results indicate that MS751‑Exo does not broadly activate signaling pathways but specifically targets the AKT/mTOR/S6K axis.

FIGURE 4.

Cervical cancer‐derived exosomes activate the AKT/mTOR signaling pathway in LECs by delivering oleic acid

(A) Dual‐color fluorescence co‐localization tracing of exosome‐delivered OA. Exosome membranes were labeled with PKH26 (red), and FITC‐OA (green) was loaded into the same exosomes to generate Exo‐PKH26/FITC‐OA. After 4 h of co‐incubation with LECs, pronounced co‐localization of red and green signals (yellow) was observed in the cytoplasm, confirming that exosomes deliver OA into LECs. Scale bar, 30 µm. (B) KEGG pathway enrichment analysis of differentially expressed metabolites in MS751‐Exo, showing significant enrichment of the mTOR signaling pathway. (C) Western blot analysis of phosphorylation levels of AKT, mTOR, S6K, and ERK1/2 in HLECs and HLEC‐SV40T cells treated with PBS, C33A‐Exo, or MS751‐Exo. (D) Western blot analysis of phosphorylation levels of AKT, mTOR, and S6K in HLECs and HLEC‐SV40T cells treated with exosomes loaded with increasing concentrations of OA (0, 50, 150, 200 µM). Data are presented as mean ± SD from three independent experiments; two‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 vs. 0 µM; #P < 0.05, ##P < 0.01, ###P < 0.001 vs. 50 µM; &P < 0.05, &&P < 0.01, &&&P < 0.001 vs. 150 µM. (E) Western blot analysis of AKT/mTOR/S6K phosphorylation and expression of downstream proteins (Cyclin D1, N‐cadherin, E‐cadherin, LYVE1, and PDPN) in HLECs and HLEC‐SV40T cells treated with Exo‐shNC (control), Exo‐shSCD‐1 (exosomes from SCD‐1‐knockdown MS751 cells), or OA‐Exo(shSCD‐1) (OA‐loaded exosomes derived from SCD‐1‐knockdown cells). (F) Representative images of tube formation in HLECs and HLEC‐SV40T cells treated with PBS, C33A‐Exo, MS751‐Exo, or MS751‐Exo combined with the AKT inhibitor Miransertib (MS751‐Exo + Miransertib). (G–H) Migration and invasion assays in HLECs and HLEC‐SV40T cells under the same treatment conditions as in (F). (G) Representative images and quantitative analysis of migration. (H) Representative images and quantitative analysis of invasion. Data are presented as mean ± SD from three independent experiments; one‐way ANOVA. ***P < 0.001 vs. PBS; ###P < 0.001 vs. C33A‐Exo; &&&P < 0.001 vs. MS751‐Exo. Scale bar, 100 µm (applies to F–H).

Next, we evaluated the effect of different OA loading amounts on AKT signaling activation in LECs. Exosomes carrying various concentrations of OA (0, 50, 150, 200 µM) were prepared and used to treat HLEC and HLEC‑SV40T cells. Western blot results showed that, compared with the 0 µM group, 50 µM and 150 µM OA dose‑dependently increased the phosphorylation levels of AKT, mTOR, and S6K; the 200 µM treatment slightly decreased the signal compared with 150 µM, but remained significantly higher than the 0 µM control (Figure 4D). These findings demonstrate that exosome‑carried OA activates the AKT/mTOR/S6K signaling pathway in a dose‑dependent manner, with the strongest effect near the saturating loading concentration of 150 µM. We then sought to determine whether tumor exosome‑induced AKT/mTOR pathway activation is specifically dependent on their carried OA. Loss‑of‑function and rescue models were established in homologous MS751 cells. First, exosomes with significantly reduced OA content were isolated from SCD‑knockdown MS751 cells (Exo‑shSCD‑1). Western blot analysis revealed that, compared with control exosomes (Exo‑shNC), Exo‑shSCD‑1 exhibited a markedly impaired ability to activate AKT/mTOR/S6K phosphorylation in HLEC and HLEC‑SV40T cells (Figure 4E, Figure S4C). To directly test the sufficiency of OA, exogenous OA was loaded into Exo‑shSCD‑1 to generate OA‑rescued exosomes (OA‑Exo‑(shSCD‑1)). Notably, OA rescue fully restored the ability to activate the AKT/mTOR pathway (Figure 4E, S4B). Downstream protein analysis showed that in the Exo‑shSCD‑1 group, the upregulation of CyclinD1, N‑cadherin, and the lymphangiogenic markers LYVE1 and PDPN was suppressed, and the downregulation of E‑cadherin was reversed; these changes were reversed in the OA‑Exo‑(shSCD‑1) group (Figure 4E, Figure S4C ). These results demonstrate that exosome‑delivered OA specifically regulates a pro‑lymphangiogenic protein network through activation of the AKT/mTOR pathway.

We further confirmed that activation of the AKT/mTOR pathway and its downstream effects are functionally essential using pharmacological inhibition. HLECs were pretreated with the specific AKT inhibitor Miransertib before the addition of MS751‑Exo (Miransertib + MS751‑Exo). At the molecular level, Miransertib completely abrogated MS751‑Exo‑induced AKT/mTOR/S6K phosphorylation (Figure S4D‑E). More importantly, the expression of downstream functional proteins was also suppressed: compared with the DMSO + MS751‑Exo group, the Miransertib + MS751‑Exo group showed significantly inhibited upregulation of CyclinD1 and N‑cadherin, and the downregulation of E‑cadherin was reversed (Figure S4D‑E). In tube formation assays, MS751‑Exo significantly promoted the network‑forming capacity of HLEC and HLEC‑SV40T cells compared with the PBS and C33A‑Exo control groups; however, the network‑forming capacity of the Miransertib + MS751‑Exo group was markedly impaired, as evidenced by substantially reduced total tubulogenic capacity and branch points (Figure 4F, Figure S4F). In invasion assays, Miransertib co‑treatment also significantly suppressed the enhanced cell motility induced by MS751‑Exo (Figure 4G), and consistent results were observed in migration assays (Figure 4H, Figure S4G). In parallel experiments, the C33A‑Exo treatment group showed no significant differences from the PBS group in any of the morphological parameters (Figure 4F‑H, Figure S4F‑G).

Collectively, these findings demonstrate that activation of the AKT/mTOR/S6K pathway and the consequent regulation of key functional downstream proteins such as CyclinD1, N‑cadherin, LYVE1, and PDPN constitute an indispensable molecular mechanism underlying MS751‑Exo‑mediated pro‑lymphangiogenic effects, and that exosome‑carried OA is the upstream key effector molecule specifically activating this pathway.

3.5. Exosomal Delivery of OA Drives Tumor‐Associated Lymphangiogenesis and LNM in Vivo

To validate the functional role of exosomal OA in vivo, a nude mouse footpad model of CC lymphatic metastasis was established. Six experimental groups were included (see Figure 5A for detailed group design): G1 (vehicle control: shNC cells with intratumoral PBS), G2 (positive control: shNC cells with intratumoral Exo‐shNC), G3/G4 (loss‑of‑function: shSCD‑1 or shSCD‑2 cells with corresponding Exo‐shSCD), G5 (rescue: shSCD‑1 cells with OA‑rescued Exo), and G6 (delivery control: shSCD‑1 cells with free OA). From day 11 onward, the corresponding exosomes (100 µg) or control vehicle were administered via intratumoral injection every 3 days. Throughout the experiment, mouse body weight and footpad orthotopic tumor volume was monitored regularly. Body weights remained stable across all groups (Figure S5A), and tumor growth curves are shown in Figure 5B.

FIGURE 5.

In vivo experiments demonstrate that exosomal delivery of oleic acid drives tumor lymphangiogenesis and lymph node metastasis

(A) Schematic diagram of the experimental design using a footpad cervical cancer lymphatic metastasis model in nude mice. Six groups were established: G1: shNC cells + PBS control, G2: shNC cells + shNC‐Exo, G3: shSCD‐1 cells + corresponding shSCD‐1‐Exo, G4: shSCD‐2 cells + corresponding shSCD‐2‐Exo,G5: shSCD‐1 cells + OA‐Exo(shSCD‐1) (i.e., OA‐loaded exosomes derived from shSCD‐1 cells),G6: shSCD‐1 cells + free OA.Starting from day 11, the corresponding exosomes (100 µg) or control vehicle were administered via intratumoral injection every 3 days. (B) Growth curves of orthotopic tumors in each group over time. (C) Representative images of resected orthotopic tumors at the endpoint (upper panel) and quantitative analysis of tumor volume (lower panel). Data are presented as mean ± SD from eight mice per group; one‐way ANOVA. (D–E) Assessment of metastatic burden in draining lymph nodes. (D) Representative H&E staining images of lymph nodes. (E) Representative images of CEA immunohistochemistry (left) and quantitative analysis of CEA‐positive area (right). Data are presented as mean ± SD (n  =  8 mice per group); one‐way ANOVA. **P < 0.01, ***P < 0.001 vs. G1; ###P < 0.001 vs. G3; &P < 0.05 vs. G6. (F) Statistical analysis of lymph node metastasis incidence across the six groups. Overall comparison was performed using Pearson's chi‐square test; pairwise comparisons were performed using Fisher's exact test. **P < 0.01. (G) LYVE1 immunohistochemical staining of orthotopic tumors (left) and quantitative analysis of lymphatic vessel density (right). Data are presented as mean ± SD (n  =  8 mice per group); one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 vs. G1; ##P < 0.01, ###P < 0.001 vs. G3; &P < 0.05 vs. G6. Scale bars: 50 µm, 100 µm, and 200 µm as indicated in the respective panels.

At the experimental endpoint, tumor and lymph node phenotypes were collected and analyzed. Images of resected orthotopic tumors (Figure 5C) revealed that, compared with the G1 group, administration of wild‐type exosomes (G2) significantly increased tumor volume. In contrast, injection of exosomes derived from SCD‐knockdown cells (G3, G4) partially suppressed tumor growth. OA replenishment into defective exosomes (G5) fully restored their pro‐tumor growth effect, whereas direct injection of free OA (G6) exerted only a weak tumor‐promoting effect. Photographs of orthotopic tumors on day 17 (mid‐intervention) visually reflected these differences (Figure S5B). Of note, no statistically significant differences in draining lymph node (inguinal) volume were observed among the groups (Figure S5C), indicating that exosome treatment did not induce physical lymph node enlargement. Therefore, histopathological methods were employed to directly assess tumor metastasis within lymph nodes. H&E staining revealed that lymph node architecture in the G2 group was extensively disrupted by massive tumor cell colonies, whereas the G1 group exhibited normal nodal structure (Figure 5D). Quantitative analysis of CEA immunohistochemistry further confirmed that the lymph node metastatic burden was significantly higher in the G2 group than in the G1 group (Figure 5E). Compared with the G2 group, G3/G4 groups displayed markedly reduced tumor cell infiltration and CEA‐positive area (Figure 5E), demonstrating that reducing exosomal OA suppresses metastasis. Importantly, OA replenishment (G5) completely rescued the LNM phenotype (Figure 5E). To further verify whether this metastatic effect is dependent on exosomal OA, mice bearing SCD‐knockdown cells and injected with their derived exosomes (G3, G4) exhibited markedly reduced tumor cell infiltration in lymph node H&E sections, and quantitative CEA analysis also revealed significantly decreased metastatic burden compared with the G2 group (Figure 5D, E ). These results demonstrate that reducing exosomal OA content effectively impairs its pro‐metastatic function (Figure 5F). Metastasis is closely associated with lymphangiogenesis. IHC analysis of orthotopic tumors revealed that the density of LYVE1‐positive lymphatic vessels within and around the tumor tissue was significantly higher in the G2 group than in the G1 group (Figure 5G). In contrast, lymphangiogenesis was suppressed in the G3/G4 groups. More importantly, reloading OA into exosomes derived from SCD‐knockdown cells prior to injection (G5) simultaneously rescued the lymphangiogenesis phenotype (Figure 5G) and LNM rate (Figure 5F), restoring them to levels comparable to those of the G2 group.

Mechanistic analysis of tumor tissues revealed markedly enhanced p‐AKT signaling in the G2 group (Figure S5D). Consistent with the pro‐proliferative and pro‐metastatic phenotypes observed in this group, the expression of the proliferation marker Ki‐67 and the mesenchymal marker N‐cadherin was upregulated (Figure S5E, F), whereas expression of the epithelial marker E‐cadherin was downregulated (Figure S5G). These molecular changes induced by MS751‐Exo were substantially attenuated in the G3/G4 groups but re‐emerged in the G5 group, in which exosomal OA was restored (Figure S5D‐G). These results confirm in vivo that exosomal OA‐activated AKT/mTOR signaling and its downstream proliferative and EMT processes are key mechanisms driving tumor lymphangiogenesis and metastasis.

Finally, the efficiency of exosome‐mediated delivery versus free molecule delivery was compared. Although direct intratumoral injection of an equivalent amount of free OA (G6) exhibited a trend toward promoting metastasis and lymphangiogenesis, its effects were significantly weaker than those observed in the G5 group, in which an equivalent amount of OA was delivered via exosomes (Figure 5G, Figure S5D–G). This directly demonstrates that exosomes, as delivery vehicles, are essential for OA to exert its maximal pro‐metastatic function in vivo. Collectively, these in vivo experiments confirm that CC cell‐derived exosomes drive tumor‐associated lymphangiogenesis and promote LNM through efficient delivery of OA.

3.6. High Expression of OA Synthases in Tumor Tissue Correlates with Lymphangiogenesis, AKT Activation, and Poor Prognosis

To investigate the clinical relevance of the OA synthesis pathway in CC lymphatic metastasis, the expression profile of fatty acid synthase (FASN) was first analyzed. Western blot analysis revealed that FASN protein levels were significantly higher in CC tissues from patients with lymph node metastasis (LNM+) than in those from LNM−patients (Figure 6A). At the mRNA level, FASN expression was positively correlated with the lymphatic endothelial marker LYVE1 (Figure 6B). This finding was further validated through analysis of a CC cohort from the public GEPIA database. Compared with normal cervical tissues, CC tissues exhibited significantly upregulated mRNA expression of both FASN and stearoyl‐CoA desaturase (SCD), the direct enzyme responsible for OA synthesis (Figure 6C). Moreover, high expression of both genes was associated with advanced clinical stage (Figure 6D). In contrast, the expression of fatty acid transport proteins FATP5 and CD36, which mediate extracellular OA uptake, was not significantly altered in cancerous tissues (Figure 6C), suggesting that cancer cells may rely more on de novo synthesis than on exogenous uptake for OA acquisition. To further validate the above findings at the cellular level, we examined the expression of proteins involved in fatty acid synthesis and uptake in different CC cell lines. Compared with normal cervical epithelial cells (HCerEpiC) and low‑metastatic C33A cells, the highly metastatic MS751 cells showed significantly elevated expression levels of FASN and SCD (Figure 6E). In contrast, the expression of the fatty acid transport proteins FATP5 and CD36 exhibited no consistent differences among the cell lines (Figure 6E). These results indicate that endogenous fatty acid synthesis is the primary cause of OA enrichment in MS751‑derived exosomes. Collectively, these results indicate that key components of the fatty acid synthesis pathway (FASN/SCD) are upregulated in CC and correlate with lymphangiogenesis marker expression and disease progression, supporting a potential role for lipid metabolic reprogramming in lymphatic metastasis.

FIGURE 6.

Oleic acid synthesis enzyme expression positively correlates with lymphangiogenesis and lymph node metastasis in tumor tissues

(A) Western blot analysis of FASN and LYVE1 protein expression levels in lymph node metastasis‐positive (LNM+) versus metastasis‐negative (LNM−) cervical cancer tissues, with corresponding quantitative analysis. Data are presented as mean ± SD (n  =  8 per group); Student's t‐test. ***P < 0.001 vs. LNM−. (B) Correlation analysis between FASN mRNA expression and LYVE1 mRNA expression in cervical cancer tissues. Pearson correlation coefficient (r) and P value are shown. (C) GEPIA database analysis of FASN, SCD, FATP5, and CD36 mRNA expression levels in cervical cancer tissues compared to normal cervical tissues. *P < 0.05, **P < 0.01 (as indicated in the original figure; please verify significance levels). (D) GEPIA database analysis of clinical stage distribution in patients stratified by high versus low expression of FASN and SCD. High expression of both genes is significantly associated with a higher clinical stage. (E) Western blot analysis of FASN, SCD, FATP5, CD36, and GAPDH (loading control) expression in normal cervical epithelial cells (HCerEpiC) and cervical cancer cell lines (C33A and MS751). Data are presented as mean ± SD from three independent experiments; two‐way ANOVA. **P < 0.01, ***P < 0.001 vs. HCerEpiC; ##P < 0.01, ###P < 0.001 vs. C33A. (F) Representative immunofluorescence dual‐staining images of SCD (red) and LYVE1 (green) in patient cervical cancer specimens. Co‐localization appears in yellow. Scale bar as indicated (e.g., 50 µm). (G) Representative immunofluorescence dual‐staining images of LYVE1 (green) and p‐AKT (red) in patient specimens. Scale bar as indicated. (H) Comparison of lymph node metastasis (LNM) rates stratified by high versus low expression of SCD or p‐AKT. Statistical analysis was performed using Fisher's exact test or chi‐square test (as appropriate). *P < 0.05, **P < 0.01 (please specify exact test). (I) Scatter plot showing the correlation between SCD expression and p‐AKT expression in cervical cancer tissues. Spearman's rank correlation coefficient (ρ) and P value are shown.

To validate this association in situ at the tissue level, IF analysis was performed on patient specimens. Dual staining for SCD and LYVE1 revealed that tumor regions with high SCD expression, particularly at the tumor–stromal interface, exhibited a marked increase in surrounding LYVE1‐positive lymphatic vessel density (Figure 6F), indicating spatial colocalization between areas of active tumor lipid synthesis and lymphatic hyperplasia. Critically, in adjacent sections subjected to dual staining for LYVE1 and p‐AKT, these LYVE1‐positive LECs displayed clear intracellular p‐AKT activation signals (Figure 6G). To determine the clinical significance of these findings, quantitative analysis of the aforementioned markers was performed. Patients were stratified into high‐ and low‐expression groups based on IF intensity. Results demonstrated that patients with high SCD expression or high p‐AKT expression (Figure 6G) exhibited significantly higher rates of lymph node metastasis (LNM+) than their low‐expression counterparts. Furthermore, SCD expression and p‐AKT expression were significantly positively correlated in tissue specimens (Figure 6H). Collectively, integrating multiple lines of evidence, this study reveals that upregulation of the fatty acid synthesis pathway (FASN/SCD) in CC cells is spatially and statistically concordant with increased density of lymphatic vessels (LYVE1+) in the TME, activation of the AKT/mTOR pathway (p‐AKT+) in LECs, and elevated rates of LNM. These findings provide clinical and histopathological support for the central hypothesis that tumor cell‐autonomous OA synthesis drives lymphangiogenesis and lymphatic metastasis.

4. Discussion

LNM is a pivotal determinant of prognosis and therapeutic decision‐making in CC (Pinto et al. 2022, Jeong et al. 2020). Consequently, elucidating the mechanisms underlying LNM is essential for developing targeted interventions. In solid tumors, lymphangiogenesis—a process involving the sprouting of new lymphatic vessels from pre‐existing ones or the differentiation of endothelial progenitor cells into LECs—constitutes a critical step facilitating metastatic spread (Ji 2006, Dieterich et al. 2022). This process is classically driven by key lymphangiogenic factors, notably VEGF‐C/D and FGF, which activate VEGFR‐3 on LECs to trigger downstream signaling cascades that promote LECs proliferation, migration, and new vessel formation (Cao et al. 2012, Banerjee et al. 2023). However, clinical evidence indicates that a subset of LNM cases presents with low VEGF expression, suggesting the involvement of alternative, VEGF‐independent pro‐lymphangiogenic pathways (Chen et al. 2020, Roy et al. 2021, He et al. 2025). In this context, lipid metabolic reprogramming has emerged as a hallmark of cancer metabolism, significantly contributing to TME remodeling and metastatic progression (Martin‐Perez et al. 2022, Corn et al. 2020). In CC, elevated expression of lipid metabolism enzymes such as FASN and FABP5 correlates with LNM (Du et al. 2022, Zhang et al. 2020); yet, their precise mechanistic roles and potential involvement in non‐canonical lymphangiogenesis remain poorly defined.

Tumor‐derived exosomes have garnered recognition as crucial mediators of intercellular communication, capable of systemically delivering bioactive molecules—including proteins, nucleic acids, and metabolites—to reprogram recipient cell function (Wu et al. 2019, Hoshino et al. 2015). Prompted by this paradigm, our initial experiments revealed that CM from HPV‐positive MS751 CC cells induced more robust lymphangiogenesis in LECs compared to CM from HT‐3 cells. Considering that tumor CM contains both soluble factors and vesicular components, and given the established role of exosomes in long‐distance signaling (Meldolesi 2018, Wortzel et al. 2019), we specifically investigated their contribution. Our data demonstrated that exosomes isolated from MS751 cells significantly enhanced LEC activity. Critically, inhibition of exosome secretion using GW4869, as well as genetic knockdown of RAB27A, substantially attenuated the lymphangiogenic potential of MS751‐CM, as well as LEC migration and invasion, thereby directly establishing the necessity of exosomes in this process. A key finding of our study is that while exosomes derived from MS751 cells and normal controls exhibited comparable physical characteristics (e.g., quantity and size), their biological capacity to promote lymphangiogenesis was markedly distinct. This functional divergence strongly implies that tumor cells endow exosomes with specialized activity not through quantitative changes but via selective cargo loading. This concept is supported by prior studies: for instance, García‐Silva et al. reported that in melanoma, sEVs carrying NGFR activated a lymphangiogenic program through MAPK and NF‐κB pathways, leading to VEGFR3 phosphorylation and increased LYVE1+ vessels (García‐Silva et al. 2021). Beyond proteins, exosomal non‐coding RNAs also contribute to lymphangiogenesis. In Cervical squamous cell carcinoma, exosomal miR‐221–3p was shown to promote lymph node lymphangiogenesis and metastasis by activating ERK/AKT signaling and downregulating vasohibin‐1 (VASH1) (Zhou et al. 2022). Similarly, in bladder cancer, sEV‐contained lncRNA LNMAT2 recruited hnRNPA2B1 to induce H3K4me3 modification at the PROX1 promoter—a master regulator of LEC differentiation—thereby activating lymphangiogenesis (Chen et al. 2020).

In summary, tumor‐derived exosomes can effectively prime draining lymph nodes for metastasis by delivering pro‐lymphangiogenic signals, such as specific proteins (e.g., NGFR, CXCR4) and long non‐coding RNAs, thereby contributing to pre‐metastatic niche formation. However, within the context of profoundly reprogrammed tumor metabolism, a critical question remains: can abundant tumor‐specific metabolites also be packaged into exosomes and play an independent, decisive role in remodeling the lymph node microenvironment? To address this, we employed untargeted metabolomics and identified a specific enrichment of the monounsaturated fatty acid OA in exosomes derived from CC cells (MS751‐Exo). While dietary OA intake has been linked to enhanced LNM in CC (Yang et al. 2018), our study is the first to associate this phenomenon with tumor‐derived exosomes, positing exosomal delivery as a potential mechanism for OA's pro‐metastatic function. Although OA's roles in cancer are context‐dependent and often centered on tumor cell‐autonomous effects (Zhang et al. 2024, Shen et al. 2017, Menendez et al. 2005), its capacity to regulate lymphangiogenesis in LECs via exosomal delivery was unknown. Our gain‐of‐function experiments, using OA‐loaded exosomes (OA‐Exo) from normal cells, demonstrated that exosome‐delivered OA potently enhanced LEC tube formation, proliferation, migration, and invasion—effects comparable to those of native tumor exosomes. Furthermore, OA‐Exo treatment upregulated the proliferation marker Cyclin D1 and induced epithelial‐mesenchymal transition (EMT)‐like marker shifts (increased N‐cadherin, decreased E‐cadherin) in LECs. These findings not only anchor OA's pro‐metastatic effect to a specific exosomal delivery mechanism but also rigorously demonstrate that exosomal OA is both necessary and sufficient to reprogram LECs into an activated, pro‐lymphangiogenic state.

The functions of fatty acids in cancer are diverse. Saturated fatty acids like palmitic acid (PA) promote tumor proliferation, for instance, via β‐catenin pathway activation (Zhang et al. 2023), or induce immunosuppression through Akt/mTOR‐mediated signaling (Sun et al. 2024). The role of OA is equally complex and tumor‐type specific, ranging from inhibiting the Her‐2/neu oncogene in breast cancer (Menendez et al. 2005) to influencing cancer cell migration and microenvironment crosstalk[39]. This heterogeneity underscores the importance of defining its function within a specific context. In CC lymphangiogenesis, our KEGG enrichment analysis pointed to the mTOR pathway as a potential target of OA‐enriched exosomes (OA‐Exo). The AKT/mTOR pathway is a central regulator of cell growth, metabolism, and survival, and is a common signaling hub targeted by various exosomal cargos (e.g., miRNAs) to reprogram recipient cells (Li et al. 2021, Huang et al. 2024, Xue et al. 2024). However, whether a specific metabolite like OA can directly drive lymphangiogenesis through this pathway was unclear. Our molecular analyses confirmed that OA‐Exo efficiently induces phosphorylation and activation of the AKT/mTOR pathway in LECs. Crucially, pharmacological inhibition of AKT with Miransertib not only blocked this pathway activation but also completely reversed all pro‐lymphangiogenic phenotypes induced by OA‐Exo in LECs, including proliferation, migration, tube formation, and EMT marker changes. Collectively, these gain‐of‐function and loss‐of‐function experiments establish that activation of the AKT/mTOR pathway is the necessary and sufficient downstream mechanism through which exosomal OA drives functional reprogramming of LECs to promote lymphangiogenesis.

Building upon these mechanistic insights, our findings hold significant translational implications. Exosomes, as natural endogenous nanovesicles obtainable non‐invasively from biofluids, are promising liquid biopsy biomarkers and therapeutic vehicles. Clinically, we observed that serum exosomal OA levels were significantly elevated in CC patients with lymph node metastasis (LNM+). Moreover, tumor tissue expression of key OA biosynthetic enzymes (FASN/SCD) correlated directly with lymphatic vessel density, metastatic risk, and poor prognosis. These associations position circulating exosomal OA and tissue FASN/SCD as potential novel biomarkers for predicting lymphatic spread. Therapeutically, exosomes offer unique advantages as drug delivery systems due to their biocompatibility, low immunogenicity, and inherent targeting potential, which can be further enhanced through engineering (Liang et al. 2021, Kimiz‐Gebologlu and Oncel 2022, Chen et al. 2024). Our in vivo experiments provide a direct rationale for targeting this pathway: in a CC model, exosomes from MS751 cells (MS751‐Exo) promoted LNM and peritumoral lymphangiogenesis. Crucially, this effect was strictly dependent on exosomal OA, as evidenced by the diminished activity of exosomes from SCD‐knockdown cells (Exo‐shSCD) and the full restoration of function upon OA reloading. Notably, exosome‐delivered OA exhibited a far stronger pro‐metastatic effect than an equivalent amount of free OA, underscoring that exosomes are not passive carriers but essential for maximizing OA's biological efficacy. This underscores the potential of strategies aimed at inhibiting exosomal OA loading or delivery, or leveraging engineered exosomes to deliver OA‐pathway inhibitors, as novel interventions. It should be noted that although OA‑rescued exosomes (shSCD1+OA‐Exo) restored functional activity in vitro, in vivo effects were slightly weaker than those of control exosomes (Exo‐shNC). This discrepancy may be attributed to two factors. On the one hand, in the control Exo‐shNC group (G2), tumor cells possess high endogenous OA and are also supplemented with exosomal OA, whereas in the rescue group (G5), endogenous OA synthesis is knocked down and the sole reliance on exogenous exosomal delivery may not fully achieve the cumulative OA levels observed in the control group. On the other hand, unlike direct treatment in vitro, in vivo factors such as biodistribution, half‑life, and metabolic competition within the TME may affect delivery efficiency. Quantitative measurement of OA content in tumor tissues or optimization of exosome delivery strategies would help validate this interpretation.

Despite these insights, several questions remain. The precise molecular target through which OA initiates AKT/mTOR signaling‐whether via a surface receptor or direct intracellular modulation‐remains to be elucidated. Additionally, the specific transporters (such as FABPs) responsible for the active packaging of OA into exosomes during biogenesis require further investigation. Future studies with larger patient cohorts are warranted to determine whether exosomal OA content is positively correlated with the extent of lymph node metastatic burden.

In conclusion, this study delineates a novel metabolism‐vesicle‐signaling axis that drives LNM in CC. We demonstrate that CC cells undergo lipid metabolic reprogramming, resulting in the selective enrichment of OA within secreted exosomes. These exosomes function as specialized transcellular carriers that deliver OA to LECs, where it activates the AKT/mTOR pathway to license a pro‐lymphangiogenic program. Our work identifies a VEGF‐independent mechanism of tumor‐induced lymphangiogenesis and highlights exosomal OA and its biosynthetic enzymes (FASN/SCD) as potent diagnostic biomarkers and therapeutic targets. By blocking exosomal OA delivery or inhibiting the SCD‐mediated synthesis pathway, clinicians may gain a new strategic perspective for combating lymphatic dissemination and improving outcomes for patients with CC.

Author Contributions

Jinqiu Li: methodology, investigation, writing – original draft, data curation. Xiaojing Huang: formal analysis, data curation, visualization. Guzhalinuer Abulizi: resources, investigation, writing – review and editing. Ayshamgul Hasim: conceptualization, supervision, funding acquisition, writing – review and editing, project administration.

Funding

This work was supported by the Graduate Student Innovation Program of Xinjiang Autonomous Region (Grant No. XJ2024G173) and the National Natural Science Foundation of China (Grant No. 82260516).

Ethics Approval and Consent to Participate

All procedures involving human participants were performed in accordance with the ethical standards of the Declaration of Helsinki. Written informed consent was obtained from all individual participants included in the study. The study protocol was reviewed and approved by the Ethics Committee of the Xinjiang Medical University (Approval No. IACUC‐20220304‐01).

All animal experiments were conducted in strict accordance with the protocols approved by the Animal Ethics Committee of Xinjiang Medical University (Approval No. IACUC‐20240711‐54).

Consent for Publication

Not applicable.

Patient Consent for Publication

Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.

Conflicts of Interest

The authors declare that they have no competing interests.

Supporting information

Supporting Figure S1. Exosome characterization, supplementary cell functional assays, and conditioned medium depletion controls.

Supporting Figure S2. Validation of exosome secretion inhibition (GW4869 and RAB27A knockdown) efficiency and functional comparison of exosomes from different sources.

Supporting Figure S3. Supplementary functional assays of free OA, optimization of exosomal drug loading, validation of SCD knockdown, and quality control of rescue experiments.

Supporting Figure S4. Tracking controls for exosomal OA delivery, time course of AKT/mTOR pathway activation, and functional validation of the inhibitor.

Supporting Figure S5. Orthotopic tumor growth dynamics and molecular pathological evidence of exosomal OA delivery promoting metastasis

Supporting Information: jev270328‐sup‐0006‐S1‐S5.docx

Supporting Table S1. Information of primary antibodies used in the present study.

Data Availability Statement

The bioinformatics results presented in this study were derived from the following public resources: GEPIA (http://gepia.cancer‐pku.cn/) and the Kaplan‐Meier Plotter database (http://kmplot.com/analysis/). All primary data generated or analyzed during this study are included in this published article [and its supplementary information files]. Additional datasets, including the raw metabolomics files, are available from the corresponding author upon reasonable request.