Fresh tissue samples from both tumor and adjacent normal tissues used in this study were obtained from The Medical Ethics Committee of the Affiliated Qinyuan Hospital of Guangzhou Medical University (Qinyuan People’s Hospital). The use of tumor samples was approved by the hospital’s Institutional Review Board (IRB-2024-068), and informed consent was obtained from all patients. The tissues were collected in a sterile operating room at Qinyuan People’s Hospital and immediately placed in 50 ml conical tubes containing sterile DMEM culture medium. Subsequent procedures were carried out using high-pressure sterilized dissection tools within a laminar flow hood to minimize contamination.

F. nucleatum strains were isolated and maintained by our research group. The cell lines MD Anderson-Metastatic Breast-231 (MDA-MB-231), Michigan Cancer Foundation-7 (MCF-7), and human umbilical vein endothelial cells (HUVECs) were also previously preserved by our research group.

Isolation, Cultivation, and Identification of Intratumoral Bacteria

Fresh clinical tissue samples were placed into 6 cm cell culture dishes containing DMEM culture medium. The tissues were minced using high-pressure sterilized surgical scissors and then transferred to 1.5 mL centrifuge tubes, following which they were thoroughly homogenized using a tissue disruptor. Next, a 100 μL aliquot of the tissue homogenate was inoculated onto blood agar plates, Sabouraud agar plates, and chocolate agar plates using sterile swabs to ensure even distribution. Additionally, 100 μL of the homogenate was added to the nutrient broth. Each of the four types of media was prepared in duplicate and incubated under both aerobic and anaerobic conditions at 37°C for up to 15 days. Visible colonies were selected for identification using mass spectrometry, and the strains were preserved at -80°C.

Frozen F. nucleatum stocks were thawed and cultured on blood agar under anaerobic conditions at 37°C for 48 hours. Afterward, single colonies were suspended in pure DMEM to prepare a bacterial suspension, which was subsequently counted and used for infection. The BC cell lines, namely MDA-MB-231 and MCF-7, were digested into single-cell suspensions the day before infection, seeded in cell culture plates, and grown to 70-80% confluence. F. nucleatum was added at a multiplicity of infection (MOI) of 50:1, gently mixed, and incubated for 48 hours in a cell culture incubator.

Approximately 1 × 10⁴ BC cells were seeded in 96-well plates and allowed to adhere. The cells were then infected with F. nucleatum at an MOI of 50:1. At 0, 24, 48, and 72 hours of culture, the old media were replaced with fresh DMEM containing 10% FBS after washing with PBS. CCK-8 reagent (DOJINDO) was added, and the cells were incubated for approximately 1 hour. Absorbance was measured at 450 nm using a microplate reader to generate growth curves.

Upon the BC cell density reaching approximately 60% in 6-well plates, the cells were infected with F. nucleatum at an MOI of 50:1 for 48 hours. After infection, the cells were detached, centrifuged, resuspended in serum-free DMEM, and adjusted to a concentration of 2.5 × 10⁵ cells/mL. Thereafter, 200 μL of the cell suspension was added to the upper chamber of the Transwell inserts, while 800 μL of complete DMEM supplemented with 20% FBS was added to the lower chamber. The plates were incubated at 37°C in 5% CO2 for 24–36 hours, then fixed with 4% paraformaldehyde and stained with crystal violet. Cells on the upper surface of the membrane were gently wiped off, and cell migration was observed and photographed under a microscope.

BC cells were seeded in 6-well plates and grown to confluence. Horizontal lines were marked on the back of the plates at 0.5 cm intervals, with three lines per well. After infection with F. nucleatum at an MOI of 50:1 for 48 hours, scratches were made perpendicular to the horizontal lines using a 200 µL pipette tip. Floating cells were discarded using PBS, and the medium was replaced with 2 mL of serum-free DMEM. The plates were photographed under an inverted microscope at 0 hours and after 24 hours of incubation. Scratch areas were analyzed using Image J software.

Total RNA from tissues or cells was extracted using an RNA extraction kit, followed by reverse transcription into cDNA using Evo M-MLV reverse transcription premix (Aker Bio). The expression levels of mRNA (Aker Bio) and miRNAs (Ribobio) were then analyzed. Specific primers used for the analysis are listed in the Supplementary Materials ( Supplementary Table 1 ).

When cell confluence exceeded 90%, the culture medium was removed, and the cells were washed twice with PBS. RIPA lysis buffer with a 1% protease inhibitor mixture was introduced (200 µL per well in a 6-well plate). After complete lysis, the cells were scraped using a cell scraper and transferred to 1.5 mL EP tubes. The lysates were sonicated on ice and centrifuged at 12,000 rpm for 10 minutes at 4°C, and the supernatants were collected. Protein concentration was determined using a BCA assay kit (Beyotime). Western blot analysis was performed using antibodies against GAPDH (CST), β-actin (CST), FOXO3 (CST), E-cadherin (CST) and Vimentin (CST), followed by detection using a chemiluminescence system. The results were photographed and saved.

When BC cells in the 12-well plate reached a confluency of 60%-80%, 100 μL of the serum-free medium was added to a sterile, enzyme-free 1.5 mL EP tube. This was followed by the addition of 2.5 μL of miR-21-3p inhibitor/inhibitor NC (Ribiotech) or FOXO3 siRNA/siRNA NC (Ribiotech), with the corresponding sequences provided in the Supplementary Materials ( Supplementary Table 2 ). The mixture was gently mixed. Meanwhile, the Simple-Pect Transfection Reagent was allowed to equilibrate to room temperature. Subsequently, 4.5 μL of the transfection reagent was added to the RNA mixture, gently mixed, and allowed to stand at room temperature for 15–20 minutes. During this incubation period, the cell culture medium in the 12-well plate was discarded and replaced with 400 μL of fresh medium containing 10% serum. After the incubation period, the RNA-transfection reagent complex was immediately added to the 12-well plate, ensuring even cell coverage. The plate was then incubated at 37°C with 5% CO2 for 6 hours, after which the complex was removed, and 1 mL of culture medium was added. The cells were further incubated at 37°C with 5% CO2. RNA or protein was extracted for subsequent experiments 48 hours post-transfection.

The Cancer Genome Atlas (TCGA) (26) BRCA expression profile data were downloaded from the UCSC XENA database (https://xena.ucsc.edu/) (27). The CancerMIRNome database (http://bioinfo.jialab-ucr.org/CancerMIRNome) (28) was utilized to examine the expression levels of miR-21-3p in BC and to evaluate its diagnostic potential. The miRDB online database (https://mirdb.org/) (29) was utilized to predict the binding sites between miR-21-3p and the FOXO3 gene. The Kaplan-Meier Plotter database (https://kmplot.com/analysis/) (30) was used to assess the impact of miR-21-3p expression on overall survival in BC patients.

All experiments were conducted in triplicate, and data analysis was carried out using GraphPad Prism 8 software. Independent sample t-tests were applied to compare data between two experimental groups, while one-way analysis of variance (ANOVA) was used for comparisons among multiple groups. Western blot grayscale values were analyzed through Image J software, and graphs were generated using GraphPad Prism 8. Statistical significance was denoted as follows: * for p < 0.05, ** for p < 0.01, and *** for p < 0.001.

To determine the presence of F. nucleatum in BC tissues, bacterial isolation and culture were conducted using fresh BC tissues and healthy breast tissues ( Figure 1 ). Colonies were identified through mass spectrometry, which confirmed the presence of F. nucleatum in BC tissues. Subsequently, an in vitro model was established using F. nucleatum isolated from clinical samples to infect MDA-MB-231 and MCF-7 cells at a multiplicity of infection of 50:1. The impact of F. nucleatum on BC cell proliferation and migration was assessed through CCK-8, Transwell, and scratch assays. As anticipated, the results indicated that while F. nucleatum infection did not affect BC cell proliferation compared to the control and heat-killed (HK) bacteria groups ( Figure 1 ), it significantly enhanced the migratory capacity of breast cancer cells (p < 0.01) ( Figures 1 , Supplementary Figure 1 ). Furthermore, the expression levels of the epithelial marker E-cadherin and the mesenchymal marker Vimentin in MDA-MB-231 and MCF-7 infection models were analyzed through Western blot analysis and RT-qPCR. The findings unveiled significant downregulation of E-cadherin and upregulation of Vimentin in the F. nucleatum infection group compared to the negative control (NC) group, while inactivation of F. nucleatum reversed these gene expression changes ( Figures 1 ).

In previous studies conducted by our research group, miRNA sequencing was performed on tissue samples collected from five BC patients and five healthy breast tissues (31). An analysis of the sequencing data and existing literature initially yielded six miRNAs ( Supplementary Table 3 ). RT-qPCR validation demonstrated that in MDA-MB-231 cells, miR-21-3p, miR-182-5p, and miR-425-5p levels were significantly higher in the F. nucleatum infection group compared to NC and HK groups. In MCF-7 cells, only the expression level of miR-21-3p was increased ( Figures 2 ). As a result, miR-21-3p was selected for further experiments.

Initially, online database analysis was used to evaluate the expression of miR-21-3p in BC tissues and its impact on overall survival in BC patients. CancerMIRNome database analysis uncovered that the expression level of miR-21-3p was significantly higher in BC tissues compared to healthy tissues (p < 0.05) ( Figure 2 ). At the same time, Kaplan-Meier analysis illustrated that high miR-21-3p expression was significantly associated with shorter survival in BC patients ( Figure 2 ). Moreover, ROC analysis indicated an AUC value of 0.95 for miR-21-3p, suggesting its high diagnostic value in BC ( Figure 2 ). These results collectively highlight the critical role of miR-21-3p in BC development.

To explore the contribution of miR-21-3p in enhancing F. nucleatum-mediated BC cell metastasis, miR-21-3p knockdown BC cell lines were established using siRNAs. Transwell and scratch assays were utilized to investigate the effect of miR-21-3p knockdown on F. nucleatum-induced BC cell migration. Interestingly, the results demonstrated that cells transfected with the miR-21-3p inhibitor exhibited significantly reduced migratory ability compared to those transfected with the inhibitor control in the F. nucleatum infection group ( Figures 2 ). Similarly, scratch assay results indicated a decreased migratory rate in cells transfected with the miR-21-3p inhibitor compared to the inhibitor control group in the F. nucleatum infection group ( Supplementary Figure 2 ). Taken together, these results suggest that miR-21-3p inhibition attenuates F. nucleatum-induced BC cell migration. Additionally, Western blot analysis and RT-qPCR demonstrated that miR-21-3p suppression significantly impeded F. nucleatum-induced EMT in breast cancer cells ( Figures 2 ).

To explore potential downstream targets of miR-21-3p, predictions were performed using the miRDB online database, which identified FOXO3 as a target gene, whilst the presence of binding sites was corroborated through TargetScan database analysis ( Supplementary Figure 3 ). Of note, previous studies have established the tumor-suppressive role of FOXO3 in various cancers. In this study, miR-21-3p expression was markedly upregulated in both BC tissues and F. nucleatum-infected BC cell models. Consequently, FOXO3 was hypothesized to be a downstream target of miR-21-3p. Validation through RT-qPCR and Western blot analysis revealed no significant change in FOXO3 mRNA levels. However, the protein expression of FOXO3 was significantly higher in the miR-21-3p knockdown groups compared to controls ( Figures 3 ). Furthermore, F. nucleatum infection lowered FOXO3 protein levels in BC cells without affecting mRNA expression ( Figures 3 ). Taken together, these findings signal that FOXO3 functions as a downstream target of miR-21-3p.

To validate the involvement of FOXO3 as a miR-21-3p downstream target in F. nucleatum-induced BC cell migration, further experiments were conducted. TCGA database analysis indicated significantly lower FOXO3 expression levels in breast cancer tissues compared to adjacent non-cancerous tissues ( Figure 4 ). Consistently, the results of UALCAN database analysis corroborated reduced FOXO3 protein levels in breast cancer tissues ( Figure 4 ). More importantly, low FOXO3 levels were significantly associated with poorer overall survival in BC patients ( Figure 4 ). Subsequent experiments assessing the impact of miR-21-3p knockdown on FOXO3 protein expression in F. nucleatum-infected BC cells revealed that miR-21-3p inhibition attenuated the suppressive effect of F. nucleatum on FOXO3 ( Figure 4 ), implying that miR-21-3p may promote BC cell migration through FOXO3 inhibition. To validate these observations, a FOXO3 knockdown cell model was generated. Transwell assays demonstrated that FOXO3 knockdown significantly promoted BC cell migration ( Figures 4 ), in line with the results of the scratch assays ( Supplementary Figure 4 ). Likewise, the results of Western blot analysis and RT-qPCR delineated that FOXO3 knockdown further elevated EMT-related gene expression at both the protein and mRNA levels ( Figure 4 ).