Bacterial extracellular vesicles from Streptococcus mutans drive OSCC progression through β-catenin signaling

Abstract

Oral squamous cell carcinoma (OSCC) is an aggressive malignancy in which microbial dysbiosis is increasingly recognized as a contributing factor, yet the pathogenic potential of bacterial extracellular vesicles (EVs) remains poorly defined. To determine whether Streptococcus mutans–derived EVs (SmEVs) influence OSCC progression, we isolated and characterized SmEVs from ATCC 35,668 and examined their effects in CAL-27 cells and a nude-mouse xenograft model. SmEVs were readily internalized by OSCC cells, enhancing their proliferative, migratory, and invasive capacities in vitro. Transcriptomic analysis revealed that SmEV exposure reshaped the CAL-27 gene expression landscape, prominently enriching the Wnt/β-catenin signaling pathway, accompanied by increased activation of β-catenin, TCF7, and FOSL1/Fra-1 in both cultured cells and xenograft tumors. Functional loss-of-experiments further demonstrated that β-catenin is critically required for these effects, as its silencing abrogated SmEV-induced proliferation, migration, and activation of the TCF7/Fra-1 axis. DIA proteomics identified a broad array of SmEV proteins, and in silico docking highlighted several abundant, cancer-associated candidates with strong predicted interactions with β-catenin. In vivo, co-administration of SmEVs accelerated tumor growth and reinforced activation of the β-catenin axis. Collectively, these findings provide the first evidence that S. mutans–derived EVs promote OSCC progression through β-catenin–driven transcriptional programs and identify specific vesicular proteins as candidate mediators, offering new insights into microbe-derived factors as potential diagnostic or therapeutic targets in OSCC.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-48919-z.

Introduction

Oral squamous cell carcinoma (OSCC), arising from the oral mucosal epithelium, is the predominant form of head and neck cancer and accounts for more than 90% of oral malignancies^1,2. In 2020, approximately 377,713 new cases of OSCC were reported worldwide, with a 5-year overall survival rate of about 60–70%^3,4. Tobacco use, alcohol consumption, betel-quid chewing, and poor oral hygiene are major risk factors for oral cancer⁵. OSCC frequently causes disfigurement and functional impairments—such as speech and swallowing difficulties and persistent pain—which substantially compromise patients’ psychological well-being and quality of life⁶. Despite recent advances in screening and targeted therapies, OSCC incidence and mortality remain high, and early detection and treatment remain major clinical challenges^7,8. Therefore, delineating molecular mechanisms that drive OSCC proliferation and metastasis may provide new opportunities for prevention and therapy.

Extracellular vesicles (EVs) are bilayered, spherical membrane structures with diameters of ~ 20–400 nm that carry components derived from the parent bacterium (including proteins, nucleic acids, lipids, and metabolites) and thereby modulate host cell activities such as protein synthesis, cellular metabolism, and intercellular communication⁹. In 2009, Staphylococcus aureus was reported to naturally release EVs into the extracellular environment, refuting the prior assumption that Gram-positive bacteria could not produce EVs because of their thick cell wall¹⁰. Owing to their modifiability and efficient cargo-loading capacity, EVs have been extensively investigated as drug-delivery vehicles. However, EVs may also exert deleterious effects: virulence factors packaged within EVs can damage host cells, resist proteolytic degradation, evade immune surveillance, facilitate long-distance dissemination, and contribute to antibiotic resistance^11,12. Multiple studies indicate that bacterial EVs can exert both pro- and anti-tumor effects^12,13. For example, an acetyltransferase from Akkermansia muciniphila has been shown to reprogram the tumor microenvironment and attenuate colorectal tumorigenesis¹⁴, while EVs from Lactobacillus reuteri promote cancer cell death and exhibit antitumor activity¹⁵. Our previous work demonstrated that EVs from Lacticaseibacillus paracasei PC-H1 induce apoptosis in colorectal cancer cells via the PDK1/AKT/Bcl-2 pathway¹⁶. By contrast, Fusobacterium nucleatum-derived EVs promote intratumoral colonization by F. nucleatum and significantly accelerate colorectal cancer progression¹⁷. Collectively, these findings indicate that bacterial EVs mediate interspecies communication and have diverse effects on host biology.

Growing evidence links OSCC with alterations in the oral microbiota, including enrichment of pathogens such as Porphyromonas gingivalis, Fusobacterium nucleatum, and members of the genus Streptococcus¹⁸. Chronic inflammation triggered by bacterial infection can promote cellular proliferation, oncogene activation, and angiogenesis, thereby facilitating oral carcinogenesis^18,19. Compared with healthy oral mucosa, taxa such as Saccharibacteria (TM7), members of the family Flavobacteriaceae, and Vibrionaceae have been found to be significantly enriched in OSCC cohorts²⁰. Oral biofilm microbiome analyses have revealed marked alterations in bacterial community composition in OSCC, including significant changes in the relative abundance of Streptococcus mutans (S. mutans), and higher S. mutans abundance has been associated with advanced clinical stage and poor disease control²¹. S. mutans is a Gram-positive cariogenic bacterium that forms robust biofilms on tooth surfaces and efficiently ferments diverse carbohydrates into organic acids, lowering local pH and promoting dental caries^22,23. To date, however, the effects of SmEVs on OSCC have not been reported. Based on the emerging roles of bacterial EVs in cancer and the enrichment of S. mutans in OSCC, we hypothesized that EVs derived from S. mutans might promote OSCC progression by modulating β-catenin signaling. In this study, we therefore investigated the effects of SmEVs on OSCC growth and metastasis and explored the underlying molecular mechanisms, with a particular focus on Wnt/β-catenin activation.

Materials and methods

S. mutans culture and SmEVs isolation

Streptococcus mutans ATCC 35,668 was obtained from ATCC (USA). Briefly, the strain was inoculated into TSB broth and cultured overnight at 37 °C. Bacterial cultures were centrifuged at 4 °C, 4,500 × g for 15 min. The resulting supernatant was passed through a 0.45 μm filter and then subjected to ultracentrifugation at 120,000 × g for 1 h at 4 °C. The pellet was resuspended in PBS and ultracentrifuged again at 120,000 × g for 1 h at 4 °C to obtain purified SmEVs. The concentration of extracted SmEVs was measured at 280 nm using a Nanodrop 2000 spectrophotometer (Thermos, USA).

Transmission electron microscopy (TEM)

SmEVs were fixed with 2.5% glutaraldehyde and then dropped onto 300-mesh copper grids. Negative staining was performed with 2% uranyl acetate. SmEV morphology was observed using a transmission electron microscope (Hitachi, Japan).

Nanoparticle tracking analysis (NTA)

SmEV particle size distribution and concentration were measured using a ZetaView nanoparticle tracking analysis system (Particle Metrix, Germany). System settings included a laser wavelength of 520 nm and a detection temperature maintained at 26.08 °C. Each sample was analyzed in three independent runs.

SmEVs proteomic analysis

SmEV samples were submitted to Beijing Novogene Co., Ltd. (Novogene, China) for DIA-based quantitative proteomic analysis. DIA-NN was employed to detect and correct tandem mass spectrometry interferences by selecting minimally affected fragments for each presumed elution peak and comparing their elution profiles to other fragments, enabling protein identification and quantification. Systematic functional annotation of identified proteins was performed.

Cell culture and cell transfection

The oral cancer cell line CAL-27 was maintained at the Department of Microbiology, Harbin Medical University (Harbin, China). This cell line was originally derived from a tongue lesion of a 56-year-old male patient with OSCC (ATCC^® CRL-2095™). As a typical poorly differentiated OSCC cell line, CAL-27 retains key biological characteristics of the primary tumor, including high proliferative activity and invasive capacity, and is widely used in studies of oral cancer pathogenesis and therapeutic evaluation. Cells were cultured in DMEM (Thermo Fisher, Beijing, China) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin (Beyotime, Shanghai, China) at 37 °C in a humidified incubator with 5% CO₂. CTNNB1 siRNA (Human Pre-designed siRNA Set A, MCE, HY-RS03306), containing three target-specific siRNAs and one negative control (NC), was used in this study. The siRNA sequences are listed in Table 1. CAL-27 cells were seeded into 6-well plates at 30–50% confluence 24 h prior to transfection. SiRNA (30 nM final concentration) was transfected using Lipofectamine™ RNAiMAX (Invitrogen) according to the manufacturer’s instructions. Briefly, siRNA and RNAiMAX were separately diluted in Opti-MEM, mixed at a 1:1 ratio, incubated for 5–10 min at room temperature, and then added to the cells in antibiotic-free medium. After 48 h of incubation, cells were harvested for Western blot analysis to evaluate β-catenin knockdown efficiency. Each experiment was performed with at least three independent replicates.

Table 1.

Primer sequences.

Fra1-F	TGACCTACCCTCAGTACAGCC
Fra1-R	CTTCCAGTTTGTCAGTCTCCTGTT
TCF7-F	CAAGAATCCACCACAGAGACAA
TCF7-R	CAAGCTGGGCTAGAGGAAGA
CTNNB1-F	GCGCCATTTTAAGCCTCTCG
CTNNB1-R	AAATACCCTCAGGGGAACAGG
GAPDH-F	ATCACTGCCACCCAGAAGGAC
GAPDH-R	TTTCTAGACGGCAGGTCAGG
si-CTNNB1-F	UCAUUAUAUUUACUAAAGCUU
si-CTNNB1-R	GCUUUAGUAAAUAUAAUGAGG

SmEVs labeling and uptake assay

To assess SmEV internalization by oral cancer cells, SmEV membranes were labeled using the ExoGlow-Membrane™ EV Labeling Kit (System Biosciences, USA). Reaction buffer was mixed thoroughly with the fluorescent dye, added to SmEVs (300 µg/mL), and incubated at room temperature protected from light for 30 min. Labeled SmEVs were then incubated with oral cancer cells at 37 °C for 24 h. SmEV RNA was labeled using the ExoGlow™-RNA EV Labeling Kit (System Biosciences, USA); SmEVs (300 µg/mL) were mixed with the kit buffer and dye and incubated at 37 °C protected from light for 1 h, followed by incubation with cells at 37 °C for 24 h. After 24 h, nuclei were stained with Hoechst, and uptake was visualized using a CellVoyager CV1000 confocal imaging system (Yokogawa, Yokohama, Japan).

Cell proliferation assay

Cells were seeded into 96-well plates at 3 × 10⁴ cells/well in 100 µL medium. After overnight incubation, the medium was replaced, and cells were treated with SmEVs at the indicated concentrations (100, 200, 300, and 400 µg/mL). After co-culture for 48 h at 37 °C, 10 µL of CCK-8 solution (Beyotime, China) was added to each well and incubated at 37 °C for 1 h. Absorbance was measured at 450 nm using a microplate reader (Molecular Devices, USA).

Migration and invasion assays

For migration assays, 200 µL serum-free DMEM was added to the upper chamber of Transwell inserts with 8.0 μm pores (Corning, USA); cells in the upper chamber were treated with SmEVs (400 µg/mL). The lower chamber was filled with DMEM supplemented with 10% FBS. For invasion assays, serum-free DMEM was added to both the upper and lower chambers and incubated at 37 °C for 2 h to hydrate the membrane (as performed in this study). Cells in 200 µL serum-free DMEM were placed in the upper chamber and treated with SmEVs (400 µg/mL); the lower chamber contained DMEM with 10% FBS. After incubation at 37 °C for 48 h, migrated and invaded cells were fixed with 4% paraformaldehyde (Biosharp, China), washed with PBS, and stained with 0.1% crystal violet (Biosharp, China) for 15 min. Stained cells were observed and photographed under an inverted light microscope (Leica, Germany) at 100× magnification. For quantification, five random fields per insert were captured, and the number of cells was counted using ImageJ software (NIH, USA). All experiments were performed in at least three independent biological replicates, with duplicate technical replicates in each independent experiment.

Transcriptome sequencing and analysis

Total RNA was extracted using TRIzol reagent, and samples were sent to Beijing Novogene Co., Ltd. for library preparation and sequencing. RNA quality was assessed with an Agilent 2100 Bioanalyzer. Differentially expressed genes were identified using the criteria |log₂ (fold change) | ≥ 1 and adjusted p-value (p adj) ≤ 0.05. GO analysis was used to identify characteristic biological attributes, including biological process (BP), cellular component (CC), and molecular function (MF). KEGG pathway enrichment analysis was performed to identify functional attributes²⁴. The top 30 GO terms and top 10 KEGG pathways were visualized using the OmicShare tools (www.omicshare.com). RNA-seq data for OSCC tissues and paired normal tissues were obtained from the Cancer Genome Atlas (TCGA) database.

Protein–protein docking

Protein tertiary structures were downloaded from the PDB (https://www.rcsb.org/) and SWISS-MODEL (https://swissmodel.expasy.org/) and saved in PDB format. Structures were preprocessed in PyMOL to remove water molecules and add all hydrogens. Rigid docking between proteins was performed using GRAMM (https://gramm.compbio.ku.edu/). Interface area and docking free energy were calculated using PDBePISA (https://www.ebi.ac.uk/pdbe/pisa/). Docking results were visualized and structurally analyzed using PyMOL.

Quantitative real-time PCR (qRT-PCR)

Total RNA was isolated with TRIzol and reverse transcribed into cDNA using PrimeScript™ RT Master Mix (Takara, Japan). qRT-PCR was performed using SYBR Premix Ex Taq™ (Takara, Japan) with cycling conditions: 95 °C for 10 s, 60 °C for 20 s, and 72 °C for 20 s for 45 cycles. Relative mRNA expression was calculated by the 2^−ΔΔCt method with GAPDH as the internal control. Primer sequences and siRNA sequences are listed in Table 1.

Western blotting

Total cellular protein was extracted using RIPA lysis buffer (Biyuntian, China), and protein concentrations were determined by BCA assay (Beyotime, China). Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were blocked with 5% nonfat milk at room temperature for 2 h and incubated with primary antibodies (diluted in TBST) overnight at 4 °C. Primary antibodies included anti-β-catenin (Abcam, ab305261, 1:1000), TCF7 (Cell Signaling Technology, #2203T, 1:1000), Fra-1 (Invitrogen, PA5-76185, 1:1000), and β-tubulin (ABclonal, AC008, 1:2000). After washing, membranes were incubated with HRP-conjugated secondary antibodies (A0208, A0216; Biyuntian, 1:2000) at room temperature for 1 h. Signals were developed using ECL reagents (Biosharp, China), and band intensities were quantified with ImageJ software.

Animal experiments

Female BALB/c nude mice (4–5 weeks old) were purchased from Beijing Weitong Lihua Experimental Animal Technology Co., Ltd. (Beijing, China). Mice were housed in specific-pathogen-free conditions with sterilized bedding, food, and water. Nude mice were randomized into two groups (n = 5 per group), and ear tags were used for identification. Groups were: (a) CAL-27 cells alone (1 × 10⁶ cells in 0.15 mL PBS); (b) CAL-27 cells (1 × 10⁶ cells in 0.15 mL) mixed with SmEVs (400 µg/mL). The cell suspension (with or without SmEVs) was slowly injected into the posterior region of the right axilla of each mouse. Tumor growth was monitored every two days, and tumor size was measured with calipers. Tumor volume was calculated as 0.5 × (longest diameter, mm) × (shortest diameter, mm)². The experiment was terminated when subcutaneous tumor diameters approached but did not exceed 15 mm, in accordance with institutional animal welfare guidelines. At the endpoint, mice were deeply anesthetized with 5% isoflurane and euthanized by cervical dislocation. Death was confirmed by the absence of reflexes and cessation of breathing prior to tumor collection. Subsequently, tumors were harvested, and their weight and dimensions were recorded. Finally, excised xenograft tumor tissues were fixed, embedded, and sectioned for subsequent analysis. The animal experiments were conducted in accordance with the guidelines of the Institutional Animal Ethics Committee of Harbin Medical University. All animal experiment protocols were reviewed and approved by the Harbin Medical University Institutional Animal Ethics Committee (HMUIRB2025026), and are in compliance with ARRIVE guidelines.

Immunohistochemistry (IHC)

Tumor tissues were paraffin-embedded and sectioned. After deparaffinization and rehydration, sections were incubated with 3% hydrogen peroxide at room temperature for 30 min to quench endogenous peroxidase activity. Non-specific binding sites were blocked, and sections were incubated overnight at 4 °C with primary antibodies (β-catenin, 1:500; TCF7, 1:200; Fra-1, 1:200). After washing, sections were incubated with Bio-anti-rabbit IgG at 37 °C for 30 min. Immunoreactive signals were visualized with 3,3′-diaminobenzidine (DAB), counterstained with hematoxylin, and examined by light microscopy. For each tumor sample, five random fields were photographed at 100× magnification. The mean optical density of positively stained areas was quantified using Image J software (NIH, USA). All images were captured and analyzed under identical exposure and threshold settings. Representative images were selected based on the mean value closest to the group average. A total of 10 images per group (5 mice × 2 fields/mouse) were analyzed, and the average value for each mouse was calculated for statistical comparison. Representative images in Fig. 4F-K were selected based on their mean values being closest to the group average.

Fig. 4.

β-catenin is required for SmEV-induced oncogenic effects in CAL-27 cells. (A) Representative western blots showing β-catenin knockdown efficiency in CAL-27 cells transfected with CTNNB1 siRNA or negative control (NC) siRNA. (B) Densitometric quantification of β-catenin protein levels normalized to GAPDH (n = 3). (C) Cell viability assay of CAL-27 cells under indicated conditions (n = 3). (D,E) Transwell migration assays of CAL-27 cells treated with or without SmEVs following β-catenin knockdown (n = 3). Scale bar, 100 μm; original magnification, 100×. (F) Representative western blots for β-catenin, TCF7 and Fra-1 (n = 3). (G–I) Densitometric quantification of protein bands. All values are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

Statistical analysis

All experiments were independently repeated at least three times. Data were analyzed and plotted using GraphPad Prism 8.0. Results are presented as mean ± standard deviation (mean ± SD). For comparisons between two groups (e.g., SmEV-treated vs. control), statistical significance was assessed using an independent-samples t-test. For multiple group comparisons (e.g., different SmEV concentrations in CCK-8 assays), one-way ANOVA followed by Dunnett’s or Tukey’s post hoc test was used. Statistical significance was set at P < 0.05.

Results

Characterization of SmEVs

SmEVs were isolated from the supernatant of S. mutans ATCC 35,668 by ultracentrifugation. The morphology of purified SmEVs was examined by transmission electron microscopy. SmEVs displayed a clear lipid bilayer and appeared as irregular, heterogeneous, oval-shaped structures (Fig. 1A). Particle size distribution measured by nanoparticle tracking analysis revealed that SmEVs were predominantly ~ 200 nm in diameter (Fig. 1B). To assess uptake by oral cancer cells, SmEV membranes and RNA cargo were fluorescently labeled and incubated with CAL-27 cells. Confocal imaging demonstrated abundant labeled SmEVs within CAL-27 cells (Fig. 1C).

Fig. 1.

Characterization of SmEVs. (A) TEM image of purified SmEVs. Scale bar, 200 nm. (B) NTA-derived particle size distribution of SmEVs. (C) Fluorescence images showing uptake of membrane- and RNA-labeled SmEVs by CAL-27 cells. Scale bar, 10 μm.

SmEVs promote proliferation, migration and invasion of OSCC cells

To determine the biological effects of SmEVs on oral squamous cell carcinoma cells, CAL-27 cell viability was measured by CCK-8 after treatment with a series of SmEV concentrations. Compared with control, SmEVs promoted CAL-27 proliferation in a concentration-dependent manner (Fig. 2A). The maximal proliferative effect was observed at 400 µg/mL; hence, this concentration was used for subsequent functional assays. Transwell assays were performed to evaluate effects on migration and invasion. SmEV treatment markedly increased the number of migrated and invaded CAL-27 cells relative to control (Fig. 2B–D). Together, these data indicate that SmEVs enhance proliferation, migration and invasion of OSCC cells.

Fig. 2.

SmEVs enhance CAL-27 cell viability, migration and invasion. (A) CCK-8 assay showing CAL-27 viability after treatment with indicated SmEV concentrations (n = 3). (B) Schematic of the Transwell assay. (C) Representative images of migrated and invaded CAL-27 cells following SmEV treatment. Scale bar, 100 μm (n = 4). (D) Quantification of migrated and invaded cells. All values were displayed as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

SmEVs promote OSCC via activation of the β-catenin/TCF7/Fra-1 axis

To explore the oncogenic mechanisms induced by SmEVs, transcriptomic profiling (RNA-seq) was performed on CAL-27 cells following SmEV co-culture. A total of 165 differentially expressed genes (DEGs) were identified upon SmEV treatment, including 36 upregulated and 129 downregulated genes (Fig. 3A). Gene Ontology (GO) enrichment analysis indicated that SmEV exposure activated host responses to bacteria and modulated immune-related pathways. Enrichment of terms such as neutrophil chemotaxis and leukocyte chemotaxis suggested that SmEVs may recruit innate immune cells (e.g., neutrophils), potentially contributing to immunosuppression and facilitating tumor dissemination. However, as these findings are derived from in vitro CAL-27 monoculture and subsequent in vivo studies used nude mice (which lack T cells), interpretations regarding adaptive immunity should be made with caution. Additionally, enrichment of cellular component terms related to secretory granules and vesicles (e.g., secretory granule lumen, cytoplasmic vesicle lumen), although involving a smaller number of genes, pointed to possible roles for vesicle-mediated intercellular communication in SmEV-induced tumor cell migration and invasion (Fig. 3B). KEGG pathway analysis of DEGs highlighted multiple cancer-related pathways; among the top ten, the Wnt signaling pathway emerged as a prime candidate underlying the SmEV-induced phenotype in CAL-27 cells (Fig. 3C). The Wnt/β-catenin pathway is a well-recognized driver of oncogenesis, particularly in OSCC, and a potential therapeutic target²⁵. Moreover, accumulating evidence suggests that this pathway also plays a key role in regulating tumor immune responses and facilitating immune evasion²⁶. Using TCGA data, expression levels of key pathway components were compared between normal and tumor tissues: CTNNB1 (β-catenin), TCF7 and FOSL1 (Fra-1) were significantly upregulated in OSCC relative to normal controls (Fig. 3D–F). Consistently, after SmEV co-culture, both mRNA and protein levels of CTNNB1, TCF7 and FOSL1/Fra-1 in CAL-27 cells were markedly elevated compared with untreated controls (Fig. 3G–M). To determine whether Wnt/β-catenin signaling is functionally required for SmEV-induced effects, we performed loss-of-function experiments by transfecting CAL-27 cells with specific siRNA targeting CTNNB1 (β-catenin). Western blot analysis confirmed that β-catenin protein levels were markedly reduced in siRNA-transfected cells compared to the negative control, demonstrating high knockdown efficiency (Fig. 4A,B). Functional assays revealed that β-catenin knockdown significantly attenuated the pro-proliferative and pro-migratory effects of SmEVs on CAL-27 cells (Fig. 4C–E). Furthermore, we evaluated the dependency of Wnt pathway activation on β-catenin. Silencing β-catenin not only reduced the basal protein levels of TCF7 and Fra-1, but also abrogated their upregulation in response to SmEV stimulation (Fig. 4F–I). These results demonstrate that the SmEV-induced activation of the TCF7/Fra-1 signaling axis is β-catenin-dependent, and that β-catenin is a critical mediator of the oncogenic phenotypes induced by SmEVs in OSCC cells.

Fig. 3.

SmEVs promote OSCC by activating the β-catenin/TCF7/Fra-1 axis. (A) Volcano plot of DEGs. (B) GO enrichment bar plot for DEGs. (C) KEGG enrichment bubble plot (top cancer-related pathways). (D–F) TCGA comparison of CTNNB1, TCF7 and FOSL1 expression in Normal (n = 44) vs. Tumor (n = 341) tissues. (G–I) qRT-PCR analysis of CTNNB1, TCF7 and FOSL1 mRNA in CAL-27 cells±SmEVs (n = 3). (J) Representative western blots for β-catenin, TCF7 and Fra-1 (n = 3). (K-M) Densitometric quantification of protein bands. All values were displayed as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

SmEVs enhance CAL-27 xenograft growth in nude mice

To validate the tumor-promoting effects of SmEVs in vivo, a CAL-27 xenograft model was established in nude mice. No significant difference was observed between the two groups before modeling (Fig. 5A). Representative images of subcutaneous tumors from control and SmEV-treated groups are shown (Fig. 5B). Tumors in the SmEV group grew significantly faster than controls (growth curves shown in Fig. 5C). At day 30, animals were sacrificed and xenografts were excised for analysis (Fig. 5D). SmEV treatment resulted in increases in tumor volume and weight compared with control (Fig. 5E). Immunohistochemical staining of tumor sections revealed markedly higher expression levels of β-catenin, TCF7 and Fra-1 in SmEV-treated xenografts relative to controls (Fig. 5F–K). These in vivo data support that SmEVs promote OSCC progression via activation of the β-catenin/TCF7/Fra-1 pathway.

Fig. 5.

Effects of SmEVs on CAL-27 xenograft growth. (A) Baseline body weight of each nude mouse in the control and treatment groups before tumor cell inoculation. (B) Representative tumor images from each group (n = 5). (C) Tumor growth curves for each group(n = 5). (D) Morphology of excised xenografts (n = 5). (E) Comparison of tumor weights(n = 5). (F–K) IHC analysis of β-catenin, TCF7 and Fra-1 expression in tumor tissues(n = 5). Scale bar, 100 μm; original magnification, 100×. All values were displayed as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

In silico prediction of candidate bioactive factors in SmEVs

To identify SmEV components potentially responsible for the pro-tumorigenic effects, proteomic profiling of isolated SmEVs was performed. A total of 1,393 proteins were identified in SmEVs. Subcellular localization analysis indicated that 49.17% of proteins were predicted to localize to the cytoplasm and 41.11% to the cell membrane (Fig. 6A). GO enrichment of SmEV proteins implicated roles in transcription, protein translation and biosynthetic processes (Fig. 6B), and suggested potential interactions with β-catenin that could facilitate TCF7/Fra-1 transcriptional activation. From the SmEV proteome, we selected six proteins for further in silico analysis based on abundance and functional relevance: glycosyltransferase-I (gtfB), glycosyltransferase-SI (gtfC), glycosyltransferase-S (gtfD), enolase (eno), phosphoglycerate kinase (pgk) and large ribosomal subunit protein uL5 (rplE). Their abundance rankings are provided in Supplementary Table 1. Molecular docking was performed to estimate binding free energies and interaction interfaces between these candidate proteins and β-catenin. The calculated binding free energies (kcal·mol⁻¹) were: gtfC − 4.9, gtfD − 13.3, gtfB − 2.1, eno − 7.1, pgk − 12.9 and rplE − 9.3 (Fig. 6C–H). These in silico results suggest that gtfC, gtfD, gtfB, eno, pgk and rplE have the potential to bind β-catenin and thereby promote downstream pathway activation that contributes to OSCC progression.

Fig. 6.

Proteomic analysis and in silico docking of SmEV components. (A) Pie chart of predicted subcellular localization of SmEV proteins. (B) GO enrichment results for SmEV proteins. (C–H) Best docking conformations and predicted binding energies for gtfC, gtfD, gtfB, eno, pgk and rplE with β-catenin.

Discussion

To our knowledge, this study provides the first evidence that SmEVs promote OSCC progression by activating the β-catenin/TCF7/Fra-1 signaling axis. We show that SmEVs are internalized by CAL-27 cells across species boundaries, enhance CAL-27 proliferation, migration and invasion in vitro, and accelerate tumor growth in a CAL-27 xenograft model in vivo. Proteomic profiling of SmEVs combined with in silico docking predicted six SmEV-enriched proteins that may interact with β-catenin and link SmEV cargo to oncogenic signaling.

Aberrant activation of Wnt/β-catenin signaling has been widely implicated in increased incidence, malignant progression, poor prognosis, and cancer-related mortality²⁵. β-catenin is a multifunctional protein that occupies a central position in Wnt signaling and participates in diverse cellular processes²⁶. In the nucleus β-catenin forms transcriptional complexes with DNA-binding partners such as TCF/LEF family members (including TCF7) and co-activators (e.g., p300/CBP) to regulate target gene expression^27,28. Fra-1 (encoded by FOSL1) has been described as a downstream mediator of TCF7 and is required for TCF7-driven proliferation in specific cancer contexts²⁹. Extracellular vesicles from tumor or stromal cells have been reported to modulate β-catenin signaling in several cancers (for example, via PDK1/AKT/GSK3β/β-catenin in hepatocellular carcinoma or via miRNA transfer from M2 macrophage EVs to activate β-catenin in colorectal cancer)^30,31. However, the interaction between bacterial EVs and the β-catenin pathway has not been well characterized. Our data fill this gap by demonstrating concordant upregulation of β-catenin, TCF7 and Fra-1 at both mRNA and protein levels in CAL-27 cells and xenografts exposed to SmEVs, concomitant with enhanced malignant phenotypes. Critically, β-catenin knockdown abrogated SmEV-induced upregulation of TCF7 and Fra-1, confirming that this pathway is functionally required for the observed effects. While our data highlight the Wnt/β-catenin axis, we acknowledge that other pathways enriched in our KEGG analysis, such as IL-17 and TNF signaling, may also contribute to SmEV-induced effects—particularly regarding inflammatory responses—and warrant further investigation.

Mechanistically, bacterial EVs carry diverse bioactive cargos (proteins, nucleic acids, enzymes, toxins, lipids) capable of crossing species barriers and modulating host cell physiology. Prior studies have implicated oral pathogen EVs in systemic disease modulation and in oral cancer–related phenotypes^32–34. For example, Porphyromonas gingivalis and Fusobacterium nucleatum EVs have been linked to inflammatory and oncogenic processes³⁵. Nevertheless, deconvolving which EV component(s) mediate specific phenotypes is challenging because EVs are complex mixtures. RNA species (including miRNAs and other small RNAs) and membrane lipids have been reported to exert important regulatory functions in EV-mediated intercellular communication, and proteomic surveys of EVs from cancer and non-cancer sources have identified large repertoires of candidate effectors^36–39. In this study, SmEV proteomics identified 1,393 proteins; subcellular localization analysis predicted that a large fraction of these proteins localizes to the cytoplasm and membrane compartments. From this dataset, we selected six abundant proteins for in silico docking with β-catenin. These included GtfB, GtfC, and GtfD—key virulence factors of S. mutans implicated in post-radiotherapy oral microbiota shifts^40,41—as well as Eno, Pgk, and RplE, which are bacterial enzymes previously associated with cancer progression⁴². It should be noted that our docking predictions are preliminary and require experimental validation. Alternatively, SmEVs may contain pathogen-associated molecular patterns (PAMPs) that trigger host signaling cascades, such as PI3K/AKT, which has been shown to stabilize β-catenin and promote oral cancer progression^43,44. Thus, the observed effects could arise indirectly through host immune activation rather than direct bacterial protein-β-catenin interactions. Docking scores suggested feasible interactions between these proteins and β-catenin, providing a plausible molecular link by which SmEV cargo could influence β-catenin–driven transcription and thereby enhance downstream programs that promote proliferation, migration and invasion. Notably, several of these candidates have prior associations with tumor biology: ENO1 dysregulation has been reported in multiple cancers and can promote proliferation in part via β-catenin activation⁴⁵; PGK1 is implicated in glycolytic reprogramming and malignant progression⁴⁶; ribosomal proteins (e.g., RPL5) can modulate signaling pathways that affect proliferation and migration⁴⁷. Thus, our combined proteomic and docking analysis offers testable hypotheses about specific SmEV effectors.

There are several important limitations. First, although proteomics and docking identified plausible candidate effectors, these analyses are predictive and do not prove direct physical interaction or functional necessity in cells. Second, other signaling pathways or EV cargos (e.g., RNAs, lipids) may contribute to the observed phenotypes and cannot be excluded. Finally, in silico docking provides thermodynamic plausibility but requires biochemical validation (e.g., co-immunoprecipitation, pull-down assays) and functional rescue experiments to identify the key effector definitively.

Conclusions

In summary, this study demonstrates that SmEVs are internalized by human OSCC cells and enhance proliferation, migration, invasion, and tumor growth in a β-catenin-dependent manner. These phenotypes are associated with activation of the β-catenin/TCF7/ Fra-1 signaling axis. Proteomic profiling combined with molecular docking nominated six SmEV-enriched candidate effectors (gtfC, gtfD, gtfB, eno, pgk and rplE) that may interact with β-catenin and contribute to downstream oncogenic programs. Together, the data establish a molecular rationale linking bacterial EVs to Wnt/β-catenin-driven OSCC progression. Further targeted biochemical and genetic validation of the prioritized candidates and pathway perturbation experiments will help to define causality and may inform future strategies to target EV components or this signaling axis therapeutically.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

G.C. performed the main parts of the experiment; Y. J., X.L., and R.L. analyzed the experimental data and wrote the manuscript; K. L., F. B. and M. D. provided technical assistance with the experiments; T. L. and Y. F. designed and supervised the study. All authors read and approved the final manuscript.

Funding

This work was supported by the 2025 Heilongjiang Province “Double First-Class” Discipline Collaborative Innovation Achievement Project (LJGXCG2025-P10) and the Heilongjiang Provincial Health Commission Research Project (20230303110207).

Data availability

The transcriptomic sequencing data generated in this study have been deposited in the NCBI Gene Expression Omnibus under accession number GSE315030 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?&acc=GSE315030). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the iProX partner repository with the dataset identifier PXD072344.

Declarations

Competing interests

The authors declare no competing interests.