Microbial community profiles in breast cancer and normal adjacent tissues: associations with clinicopathological characteristics

Introduction

According to the Global Cancer Statistics released by the International Agency for Research on Cancer (IARC), in 2022, there were approximately 2.3 million new cases of breast cancer (BC) worldwide, and BC accounted for 11.6% of all newly diagnosed cancers. BC was the second most common cancer globally, following lung cancer (12.4%). More than 660,000 individuals died from BC, and BC accounted for 6.9% of all cancer-related deaths worldwide, and was the 4^th leading cause of cancer mortality globally (1). Early screening technologies for BC, such as mammography, ultrasound, and MRI, have advanced and are widely used in clinical practice; however, they still face notable limitations in terms of their sensitivity, specificity, and applicability, especially in certain populations like young women and those with dense breast tissue. Further, while advancements in treatment strategies, such as surgery, radiotherapy, chemotherapy, endocrine therapy, and targeted therapy, have significantly improved overall survival, clinical challenges, including tumor heterogeneity, therapeutic resistance, distant metastasis, and recurrence, persist. Therefore, effective biomarkers urgently need to be identified to facilitate the early diagnosis and personalized treatment of BC, enabling the identification of high-risk individuals and the optimization of clinical intervention strategies.

In recent years, precision-targeted therapies leveraging tumor-associated bacteria have emerged as a promising and innovative strategy for cancer treatment, showing great potential in the prevention and treatment of BC (2). Traditionally, the breast was considered a “sterile” organ; however, this notion has been increasingly challenged by recent studies. The breast microenvironment, rich in fat, blood vessels, lymphatics, ducts, and lobular structures, provides favorable conditions for microbial colonization and survival (3). Multiple hypotheses have been proposed regarding the origin of breast microbiota, including translocation from the skin, breast milk transmission (the gut-mammary axis), gut migration, oral transmission, sexual contact, and environmental exposure (4-6).

Currently, 16S ribosomal RNA (rRNA) gene sequencing is the primary approach for characterizing microbial communities. Among the variable regions, V3–V4 is the most commonly targeted in BC research. A previous study has shown that region V3 provides the highest species-level resolution, while regions V8 and V9 yield less taxonomic information (7). In this study, five variable regions (V2, V3, V5, V6, and V8) were selected for multiplex polymerase chain reaction (PCR) amplification and sequencing, covering approximately 68% of the full-length 16S rRNA gene. This approach has high sensitivity and broad applicability, particularly in low-biomass tissue samples (8). Accumulating evidence indicates that microbial profiles differ between healthy breast tissue and BC tissue (6,9,10). Normal adjacent tissue (NAT), which refers to non-cancerous breast tissue collected from the same patient in proximity to the tumor, may better reflect the interactions between the tumor microenvironment and microbiota than healthy tissue. A comprehensive investigation of the microbial characteristics associated with breast malignancies may provide valuable insights for the development of large-scale diagnostic and prognostic biomarkers for BC (11).

Among the clinicopathological characteristics of BC, the proliferation index (Ki-67) is an important marker for assessing the proliferative activity of BC and is closely associated with tumor size, grade, and hormone receptor (HR) status. High Ki-67 (H-Ki-67) expression generally indicates a poor prognosis and is correlated with shorter disease-free survival and overall survival, particularly in luminal A subtype patients. Moreover, elevated Ki-67 levels are more frequently observed in highly aggressive subtypes such as triple-negative BC (TNBC), with median expression levels reaching up to 50% (12,13).

In this study, we analyzed the differences in microbial community structure between BC and NAT using 16S rRNA gene sequencing and predicted their potential functions. Data on the clinical and pathological baseline characteristics of the patients were collected to investigate microbial variations across different stratifications, including molecular subtypes, clinical stages, and histological grades. The aim of the study was to identify microbiota-related biomarkers closely associated with the occurrence and progression of BC. This study revealed differences in the microbiota and related functional pathways between groups with different levels of Ki-67 expression. Our findings may have important prognostic implications and provide a novel theoretical foundation and research direction for personalized treatment and microbiota-targeted interventions in BC. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1570/rc).

Methods

Subjects and sample collection

Tissue specimens were obtained from adult female BC patients undergoing modified radical mastectomy. During surgery, paired fresh tissue samples (approximately soybean-sized), consisting of BC and NAT from the same patient, were aseptically collected, placed into sterile Eppendorf (EP) tubes, and immediately snap-frozen at −80 ℃ in an ultra-low temperature freezer. Clinical and pathological data were collected for all participants, including age, menopausal status, body mass index, molecular subtype, clinical stage, histological grade, and Ki-67 expression level.

According to the 2024 Chinese Society of Clinical Oncology guidelines (14) for the diagnosis and treatment of BC, the patients were classified into four molecular subtypes: luminal A: estrogen receptor (ER) positive, progesterone receptor (PR) positive with high expression and Ki-67 <14%, and human epidermal growth factor receptor 2 (HER2) negative; luminal B: ER positive, PR negative or low expression, with H-Ki-67 expression, and HER2 negative; HER2 positive: HER2 positive regardless of ER/PR status; and TNBC: ER negative, PR negative, and HER2 negative. Clinical staging was based on the 8^th edition of the tumor-node-metastasis (TNM) classification system released by the American Joint Committee on Cancer in 2017 (15), which classifies BC into stages 0, I, II, III, and IV based on tumor size (T), regional lymph node (N) involvement, and distant metastasis (M). Histological grading was conducted using the Nottingham Histological Grade system, which evaluates the following three parameters: glandular (tubule) formation, nuclear pleomorphism, and mitotic count. Each parameter is scored from 1 to 3, yielding a total score between 3 and 9, and breast tumors were subsequently categorized into three grades: grade 1 (well differentiated), grade 2 (moderately differentiated), and grade 3 (poorly differentiated). Although the threshold for H-Ki-67 expression varies among clinical guidelines and studies, the 2021 “Breast Cancer Ki-67 International Working Group Assessment Guidelines” recommend using a standardized “typing machine” visual evaluation method, noting that Ki-67 expression levels ≤5% or ≥30% can serve as references for clinical treatment decisions and prognosis prediction (16). Given that 30% as the cutoff allows for a relatively balanced sample distribution, this study defined high proliferation and an elevated risk of recurrence as a Ki-67 index of ≥30% (17,18).

The following exclusion criteria were applied in this study: pregnancy or lactation; use of antibiotics within the previous 6 months; diagnosis of other malignancies; presence of systemic infectious diseases; and/or incomplete or missing clinical data.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Ethical approval was obtained from the Ethics Committee of The Second Hospital of Shanxi Medical University (approval No. [2024] YX-460). Written informed consent was obtained from all participants prior to inclusion in the study.

Microbial DNA extraction and PCR amplification

Genomic DNA was extracted from frozen tissue samples (40–70 mg) using the cetyltrimethylammonium bromide method, and DNA concentration and quality were assessed via 1% agarose gel electrophoresis. Primers targeting specific regions of the bacterial 16S rRNA gene were designed, and five primer pairs were diluted to 1 µM with double-distilled water (ddH₂O) for the PCR amplification of five variable regions of the 16S rRNA gene (i.e., V2, V3, V5, V6, and V8). Two rounds of PCR amplification were performed using different reaction systems comprising DNA templates, primers, deoxynucleotide triphosphates (dNTPs), polymerase, and reaction buffer. PCR conditions, such as the annealing temperature and cycle number, were optimized to ensure efficient amplification (8). The primer sequences for the two rounds of PCR amplification are shown in Table 1.

Library pooling and sequencing

The PCR products were purified using AMPure XT beads (Beckman Coulter Genomics, Danvers, MA, USA) and quantified using Qubit (Invitrogen, Carlsbad, CA, USA). The purified PCR products were evaluated using the Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA) and the Illumina library quantification kit (Kapa Biosciences, Woburn, MA, USA). The qualified library concentration had to be above 0.3 ng/µL. After the gradient dilution of the qualified sequencing libraries (with non-repetitive index sequences), they were pooled in appropriate proportions according to the required sequencing depth. The pooled libraries were denatured with sodium hydroxide (NaOH) to single-stranded DNA and then subjected to sequencing. Sequencing was performed on the Illumina NovaSeq 6000 platform using the PE150 paired-end sequencing mode, with the NovaSeq 6000 SP Reagent Kit (500 cycles) as the corresponding reagent.

Sequencing data processing and microbiota identification

The libraries were sequenced on the Illumina NovaSeq 6000 system using paired-end sequencing. First, the samples were demultiplexed based on barcode information, and adapter and barcode sequences were removed to ensure that the sequence data from each sample were independent and free of cross-contamination. Next, the sequencing data underwent quality control and filtering using the fqtrim (version 0.94) software. The specific parameters were set as follows: “-A -P 33 -w 10 -q 20 -l 100 -m 5 -p 1 -V -o fastq.gz.” This processing step involved quality scanning using a sliding window method, removing sequences that were too short, and filtering out sequences with an excessive number of ambiguous bases, ensuring high-quality data. Subsequently, reads from each sample were demultiplexed and filtered based on the primer sequences, and then aligned to the five targeted hypervariable regions (i.e., V2, V3, V5, V6, and V8) to ensure consistent and accurate mapping across all samples. The Short MUltiple Regions Framework (SMURF) analysis pipeline was employed to integrate and analyze the sequences from the five amplification regions, allowing for preliminary microbiota identification and relative abundance calculation (19). The database used for this project was the optimized version of Greengenes (May 2013). Contaminant bacteria introduced during sample collection, DNA extraction, or PCR amplification (based on negative control data) were removed. Finally, based on the filtered microbiota data, within-group and between-group diversity analyses and differential microbiota identification were conducted.

Statistical analysis

Based on the obtained species-level abundance table, alpha and beta diversity analyses were performed for the two groups of microorganisms. Alpha diversity was assessed using the following five indices: Chao1, Observed Species, Goods_coverage, Shannon, and Simpson, which collectively evaluate the richness and evenness of microbial communities. Beta diversity was evaluated through principal component analysis (PCA) and analysis of similarities (ANOSIM) to assess differences in microbial composition between groups. The PCA was conducted in R (version 3.4.4, R Foundation for Statistical Computing, Vienna, Austria) using the “vegan” package (version 2.5-4), and the plots were generated using the “ggplot2” package (version 3.2.0). For groups with biological replicates ≥4, a 95% confidence ellipse was added to illustrate the sample distribution and intra-group clustering trends. ANOSIM, a non-parametric statistical test used to assess the significance of differences in the similarity between and within groups, was performed using Quantitative Insights Into Microbial Ecology (QIIME) software (version 1.8.0) based on Bray-Curtis distance matrices (20). The species abundance in each sample was statistically analyzed based on the species abundance table obtained from the SMURF. The Wilcoxon signed-rank test was used to compare the species differences between the two groups, with a significance threshold of P<0.05. Differential microbiota identification was conducted by a linear discriminant analysis effect size (LEfSe) analysis, based on the original script published by the Segata team (GitHub commit version 094f447691f0). Features with linear discriminant analysis (LDA) scores (log₁₀) >3.0 were considered significant biomarkers indicative of biological relevance. The functional prediction analysis was conducted using the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States 2 (PICRUSt2) software (version 2.2.0b) (21), and the gene function annotation results were obtained from the Kyoto Encyclopedia of Genes and Genomes (KEGG) and KEGG Orthology (KO) databases. Differences in the predicted microbial functions between the groups were analyzed using Welch’s t-test in the Statistical Analysis of Metagenomic Profiles (STAMP), with a significance threshold of P<0.05 and a 95% confidence interval.

Results

Patient baseline characteristics and sequencing data information

A total of 31 adult female patients diagnosed with BC, who were treated at The Second Hospital of Shanxi Medical University between November 2024 and February 2025, were enrolled in the study. The baseline characteristics of the patients are summarized in Table 2.

16S rRNA gene sequencing was performed to analyze both the BC and NAT groups. A total of 1.0635024×10⁷ raw reads and 1.0030571×10⁷ clean reads were obtained from 62 samples, with an average Q30 percentage of 95.23%.

Microbial features of the BC and NAT groups

To assess the overall diversity of the microbiomes between the two groups, their alpha and beta diversity were analyzed. The alpha diversity rarefaction curve tended to plateau, indicating sufficient sequencing depth, and the Goods_coverage values in both groups were close to 1.0 (Figure 1A). Although the Chao1, Simpson, and Shannon indices suggested slightly higher alpha diversity in the BC group than the NAT group, the differences were not statistically significant (Figure 1B). The PCA revealed no distinct separation in beta diversity between the BC and NAT groups (Figure 1C, P<0.05).

Figure 1 Analysis of microbial diversity and community composition in the BC tissue and NAT groups. (A) Rarefaction curve based on Goods_coverage index. (B) Chao1, Shannon, and Simpson indices reflecting the alpha diversity of the two groups. (C) PCA plot illustrating the beta diversity distribution between the BC and NAT samples. (D) Stacked bar plot showing the relative abundance of dominant phyla. (E) Heatmap illustrating the composition of dominant microbial families. (F) Heatmap showing the composition of dominant microbial genera. BC, breast cancer; NAT, normal adjacent tissue; PCA, principal component analysis.

Across all samples, a total of 14 phyla (Figure 1D), 31 classes, 58 orders, 126 families, 328 genera, and 998 species were identified in the BC and NAT groups. At the phylum level, the identified microbial taxa comprised four dominant phyla (i.e., Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes), along with several low-abundance phyla, among which Proteobacteria was the most predominant phyla in both groups. Notably, Synergistetes, Tenericutes, and Verrucomicrobia were unique to the BC group. The relative abundance composition of each phylum is detailed in Table 3, and no statistically significant differences in microbial composition were observed between the two groups at the phylum level (P>0.05). At the family level, Pseudomonadaceae was the most abundant family in both the BC and NAT groups (Figure 1E). This family includes several strains associated with antibiotic resistance, biofilm formation, and pathogenicity, suggesting a potential role in the tumor microenvironment or host immune regulation. At the genus level, Pseudomonas, Acinetobacter, and Massilia were the dominant genera in both groups (Figure 1F). Further, at the family level, Rhodocyclaceae, Rhodospirillaceae, Caulobacteraceae, and Xanthomonadaceae showed significantly higher relative abundance in the BC group (Figure 2A, P<0.05). Compared with the NAT group, nine genera exhibited significantly higher relative abundances in the BC group (Figure 2B, P<0.05).

Figure 2 Differential analysis of microbial communities between the BC and NAT groups. (A) Boxplots of significantly different bacterial taxa at the family level between the two groups (P<0.05). (B) Boxplots of significantly different bacterial taxa at the genus level (P<0.05). (C) Bar chart of LEfSe showing differential microbial taxa between the two groups, with LDA scores >3 indicating a statistically significant difference. (D) Differential metabolic pathways between groups based on KEGG functional annotation (P<0.05). BC, breast cancer; KEGG, Kyoto Encyclopedia of Genes and Genomes; LDA, linear discriminant analysis; LEfSe, linear discriminant analysis effect size; NAT, normal adjacent tissue.

Additionally, LEfSe analysis was performed to identify the microbial taxa that differed significantly between the BC and NAT groups, and to screen for potential cancer-associated microbial biomarkers. The results showed that Flavobacteriales, Comamonas, and Delftia were predominantly present in the BC group (LDA >3.9), while Methylobacterium was most significantly enriched in the NAT group (Figure 2C, LDA >4). The KEGG level 3 pathway analysis (Figure 2D) revealed that the ascorbate and aldarate metabolism, lysosome, and other glycan degradation pathways were more abundant in the BC group than the NAT group (P<0.05). Additionally, 14 KOs were detected with significant differences between the two groups (Figure S1A, P≤0.01).

Microbial differences in the BC molecular subtypes, clinical stages, and histological grades

Distinct microbiome profiles were observed among different subtypes of BC. The alpha diversity analysis (Chao1 and Observed_species) revealed higher microbial diversity in the ER-positive subtypes, of which, the luminal A subtype exhibited significantly greater diversity than the HER2-positive subtype (Figure 3A,3B, P<0.05). LEfSe analysis further uncovered 9microbial differences across the BC subtypes: Lactobacillales and Ralstonia were dominant in the TNBC group, while Burkholderiales was the most enriched order in the HER2 group (Figure 3C). The KEGG level 3 pathway analysis identified 15 metabolic pathways with significant differences among the four subtypes (Figure 3D, P<0.05), of which, the TNBC subtype showed a notable upregulation in pathways related to transporters. Further, 20 KOs were found to differ significantly across the four groups (Figure S1B, P<0.01).

Figure 3 Microbial characteristics associated with different molecular subtypes and clinical stages of BC. Due to the limited number of S 0 samples, S 0 and S I were combined as S 0/I for the analysis; the other stages are represented as S II and S III. (A,B) Chao1 and Observed_species indices reflect microbial alpha diversity across different molecular subtypes. (C) Bar chart of LEfSe showing taxa with statistically significant differences among molecular subtypes. (D) The KEGG level 3 functional pathways that differed significantly among the molecular subtypes. (E) Shannon index comparing microbial alpha diversity among different clinical stages and NAT samples. (F) LEfSe bar chart indicating stage-related differential taxa. (G) The KEGG level 3 pathways that differed significantly across the clinical stages. LDA scores >3 and P<0.05 indicate statistically significant intergroup differences. *, P<0.05; ns, not significant (P>0.05). BC, breast cancer; HER2, human epidermal growth factor receptor 2; KEGG, Kyoto Encyclopedia of Genes and Genomes; LDA, linear discriminant analysis; LEfSe, linear discriminant analysis effect size; NAT, normal adjacent tissue; S, stage; TNBC, triple-negative breast cancer.

We further explored the association between microbial communities and BC staging. The Shannon index revealed significant differences in microbial alpha diversity between the stage III group and both the stage 0/I and NAT groups (Figure 3E, P<0.05). LEfSe analysis identified representative differential microbes for each stage (Figure 3F, LDA >4): stage 0/I was mainly enriched with Acinetobacter_ursingii and Pseudomonas_Unknown_species792; stage II was mainly enriched with Bacillales and Bacillaceae; while stage III exhibited the richest diversity, with Pseudomonas_jessenii being the most representative differential microbe. The KEGG level 3 pathway analysis identified 10 functional pathways with statistically significant differences across the three groups (Figure 3G, P<0.05). These pathways primarily involved pyruvate metabolism, pores, ion channels, and peroxisomes. Additionally, 50 KOs showed significant differences between the groups (Figure S1C, P<0.01).

In addition, we analyzed the differences in microbial communities across different histological grades of BC. The alpha diversity analysis revealed significant differences in species richness (based on the Chao1 and Goods_coverage indices) between the grade I/II group and both the NAT and grade III groups (Figure 4A,4B, P<0.05). The PCA further revealed a significant difference in beta diversity between the grade I/II and grade III groups (Figure 4C, P=0.001). LEfSe analysis indicated that Acinetobacter was the dominant genus in the grade I/II tissues, while Corynebacterium was the dominant genus in the grade III tissues (Figure 4D). The KEGG level 3 functional pathway analysis revealed seven metabolic pathways that differed significantly between the two groups (Figure 4E, P≤0.01). Additionally, a total of 28 KOs were found to differ significantly between the groups (Figure S1D, P<0.01).

Figure 4 Microbial characteristics associated with different histological grades of BC. Due to the limited number of G I samples, G I and G II were combined as G I/II for the analysis. (A,B) Chao1 and Goods_coverage indices reflect microbial alpha diversity among different histological grades and NAT samples. (C) PCA plot illustrating the microbial beta diversity distribution between different histological grade groups. (D) Bar chart of LEfSe showing differential microbial taxa between different histological grade groups, with LDA scores >3 indicating statistically significant differences. (E) The KEGG pathways that differed significantly between the different histological grade groups (P≤0.01). *, P<0.05; **, P<0.01; ns, not significant (P>0.05). BC, breast cancer; G, grade; KEGG, Kyoto Encyclopedia of Genes and Genomes; LDA, linear discriminant analysis; LEfSe, linear discriminant analysis effect size; NAT, normal adjacent tissue; PCA, principal component analysis.

In the H-Ki-67 group, Brevundimonas was the predominant genus, and glycolysis/gluconeogenesis was a significantly upregulated metabolic pathway

Notably, this study explored the microbial differences between different levels of tumor Ki-67 expression. An analysis of the Chao1, Goods_coverage, and Observed_species indices revealed significant differences in alpha diversity between the low Ki-67 (L-Ki-67) group and both the H-Ki-67 and NAT groups (Figure 5A-5C, P<0.05). The genera enriched in the L-Ki-67 expression group with an LDA score >4 included Acinetobacter_ursingii and Delftia_acidovorans, while Brevundimonas was the dominant genus in the H-Ki-67 group (Figure 5D). The KEGG level 3 pathway analysis identified 14 significantly different pathways between the L-Ki-67 and H-Ki-67 groups (P<0.05), with glycolysis/gluconeogenesis being a key upregulated metabolic pathway in the H-Ki-67 group. In addition, pathways such as bacterial toxins and isoflavonoid biosynthesis were also found to be upregulated in the H-Ki-67 group (Figure 5E). Further analysis identified 12 KOs with significant differences between the two groups (Figure S1E, P≤0.01).

Figure 5 Microbial characteristics associated with different proliferation indices (Ki-67) in BC. Patients with Ki-67 ≥30% were classified as the H-Ki-67 group, while those with Ki-67 <30% were classified as the L-Ki-67 group. (A-C) Chao1, Goods_coverage, and Observed_species indices reflect microbial alpha diversity among different Ki-67 and NAT groups. (D) Bar chart of LEfSe showing differential microbial taxa between different Ki-67 groups, with LDA scores >3 indicating statistically significant differences. (E) The KEGG pathways that differed significantly between the Ki-67 groups (P<0.05). *, P<0.05; ***, P<0.001; ns, not significant (P>0.05). BC, breast cancer; H-Ki-67, high Ki-67; KEGG, Kyoto Encyclopedia of Genes and Genomes; L-Ki-67, low Ki-67; LDA, linear discriminant analysis; LEfSe, linear discriminant analysis effect size; NAT, normal adjacent tissue.

Discussion

With the advancement of high-throughput sequencing technologies, 16S rRNA sequencing has become an essential tool for studying the microbiota associated with BC. This study used this technology to analyze the microbial diversity and composition differences between BC and NAT, and to make functional predictions. It also explored variations in microbial community structure and function across different clinical and pathological stratifications, as well as under varying levels of Ki-67 expression. The study findings identified potential biomarkers for BC prognosis and extended our understanding of its initiation and progression mechanisms, thus offering new avenues for personalized cancer treatment and prevention.

This study found that the microbial diversity was slightly higher in the BC tissue than in the NAT, although the difference was not statistically significant. This finding is consistent with previous studies; thus, the overall microbial composition of BC and NAT appears to be relatively similar (7,22,23). Our results also confirmed the similarity between the two groups. We further identified microbial communities that differed between BC and NAT via LEfSe analysis and screened potential microbial biomarkers, thereby enriching the foundational research on BC-related microbiota. The functional prediction results based on PICRUSt2 showed that the BC group exhibited significant enhancement in the pathways of ascorbate and aldarate metabolism, lysosome, and other glycan degradation. Multiple studies have reported that the occurrence of BC is closely related to metabolic abnormalities and lysosomal function. Notably, Wei et al. found that exposure to environmental pollutant polybrominated diphenyl ether-47 promoted breast tumor growth in a dose-dependent manner. The mechanism may be related to the upregulation of ascorbate and aldarate metabolism and glutathione metabolism pathways, making cells more susceptible to oxidative stress (24). Additionally, ascorbic acid can act as a pro-oxidant under specific conditions, exhibiting potential anti-tumor activity. Choi et al. further pointed out that L-ascorbic acid induced apoptosis in BC cells by increasing nuclear p62 levels through the ER stress-mediated IRE-JNK-CHOP signaling pathway (25). Nehme et al. showed that sequential treatment with CDK4/6 inhibitors and lysosome-promoting drugs effectively reduced the growth of HR positive and TNBC cell subpopulations in vivo (26). Yao et al. proposed that asparagine endopeptidase regulates lysosomal PI3K activity and participates in the autophagic process, making it one of the key targets for metabolic adaptation in cancer cells (27). The present study revealed distinct functional pathway differences between BC and NAT from the perspective of microbiota, providing novel insights into the pathogenesis of BC and potential therapeutic targets.

In addition to the comparison between BC and NAT microbiota, the microbial variations associated with different clinical and pathological features of BC are also noteworthy. Banerjee et al. conducted a pan-pathogen microarray analysis and found that the ER-positive subtype had the highest diversity, while the TNBC subtype had the lowest diversity (28). Conversely, our 16S rRNA sequencing analysis revealed higher diversity in ER-positive samples and the simplest composition in HER2-positive samples. These differences might arise from the different detection methods used. Meng et al. performed sterile percutaneous biopsy sampling and observed that stage III tissue exhibited higher alpha diversity, but found no such significant difference in relation to beta diversity (29). However, we observed higher alpha diversity in grade I/II compared to grade III, accompanied by significant differences in beta diversity; however, these differences might be due to differences in the specimen collection methods.

The differential distribution of microbial communities between various BC subtypes and their NAT may be driven by multiple molecular mechanisms. Firstly, certain microbes can regulate estrogen metabolism and related signaling pathways, which may be particularly important in HR positive subtypes (30). Secondly, factors within the tumor microenvironment—including cancer-associated fibroblasts (CAFs), tumor-associated macrophages (TAMs), and extracellular matrix remodeling— can influence microbial colonization and niche formation (31,32). Thirdly, host genetic background combined with microbial metabolites shapes immune regulation and chronic inflammation, potentially leading to differences in microbial composition (33). The specific mechanisms require extensive experimental validation to further elucidate their functional roles and clinical significance.

Prognostic microbial biomarkers vary across regions and ethnicities. In our study, Lactobacillales and Ralstonia were found to predominate in the TNBC group, while Burkholderiales was found to be the characteristic taxon in the HER2 group. In a microbiome analysis of BC and NAT in Ethiopian women, Burkholderia was most strongly associated with invasive TNBC and basal-like breast tumors (22). Non-Hispanic Black and non-Hispanic White BC patients display different stage-related characteristics (34). The microbial characteristics associated with different clinical stages in our study also differed from the above-mentioned study (34). These findings suggest that microbial features may be influenced by regional, ethnic, and methodological factors, warranting further exploration.

Distinct from prior research, this study is the first to compare the microbial differences among BC patients with different Ki-67 levels and found that the L-Ki-67 group exhibited higher microbial abundance. LEfSe analysis revealed that Brevundimonas was the most dominant genus in the H-Ki-67 group, which was also associated with a worse prognosis. Previously, studies reported that Brevundimonas was significantly associated with multiple BC subtypes and linked to poorer survival rates in ER-positive patients (28,35). Additionally, Chiba et al. found that elevated levels of Brevundimonas and Staphylococcus were associated with distant metastasis in BC (36). These findings are consistent with our results, suggesting that Brevundimonas may serve as a potential microbial biomarker for a poor prognosis in BC.

Glycolysis/gluconeogenesis is a significantly upregulated pathway associated with H-Ki-67. The Warburg effect promotes tumor progression and immune evasion through aerobic glycolysis. Although related targeted therapies have shown some potential, their clinical application still faces numerous challenges (37). Novel nanoplatforms, such as a new albumin-modified layered double hydroxide resveratrol dosage form (BSA@LDHs-Res), can effectively inhibit glycolytic enzyme activity and demonstrate good anti-tumor effects (38). TNBC is characterized by a hypoxic tumor microenvironment, where CAFs secrete colony-stimulating factor 3 (CSF3), upregulating phosphoglucomutase 2-like 1 (PGM2L1) and enhancing glycolysis, suggesting that the CSF3/CSF3R axis may serve as a potential therapeutic target (32). Additionally, TNBC is highly dependent on amino acid and glycolytic metabolism. The miR-152/SLC7A5/E2F1/PTBP1 axis regulates the metabolic reprogramming of TNBC, providing a direction for new target research (39). Further, gluconeogenesis-driven glycogen accumulation may also be associated with BC metastasis (40). In summary, glycolytic and gluconeogenic metabolic pathways play a critical role in the metabolic reprogramming of BC, particularly in TNBC. Targeting these metabolic pathways provides a theoretical foundation and research direction for potential therapeutic targets in BC.

Bacterial toxins and isoflavonoid biosynthesis were also upregulated in the H-Ki-67 group, suggesting their potential roles in metabolic regulation and tumor progression in BC. Previous studies have developed several anti-tumor strategies based on bacterial toxins. For example, Schrank et al. constructed an antibody-toxin conjugate targeting the CD47 pathway (CD47-LLO), which effectively inhibited both local and metastatic BC growth (41). Additionally, Halabian et al. developed a Staphylococcal enterotoxin B carrier that was shown to induce cell apoptosis and activate anti-tumor immunity (42). As phytoestrogens structurally similar to estrogens, isoflavones exhibit dual roles in BC (i.e., they may promote tumorigenesis, but as selective ER modulators, they may also have some preventive and therapeutic potential) (43). A systematic review indicated that dietary isoflavones could reduce tumor volume and weight but might also increase the tumor area; thus, the intake of isoflavones in BC patients needs to be carefully evaluated (44). Meanwhile, fluorinated isoflavonoid compounds synthesized by Liu et al. demonstrated promising anti-proliferative activity against BC cells, suggesting potential therapeutic applications (45).

This study had several limitations. First, the relatively small sample size and single-center design might constrain the generalizability of the findings. Moreover, the observed microbial diversity may be affected by participants’ racial/ethnic backgrounds and geographic locations, as variations in genetic makeup, dietary patterns, lifestyles, and environmental exposures across populations can shape microbial community composition. Second, the analysis relied solely on 16S rRNA sequencing, and the findings lack support from other omics approaches such as metabolomics. Additionally, potential confounding factors such as diet and environmental exposures were not controlled, and the functional mechanisms of the identified pathways were not experimentally validated. In the future, larger, multi-center cohort studies should be conducted that incorporate integrated multi-omics and functional analyses to yield more comprehensive and reliable results.

Conclusions

This study found that the overall microbial diversity between the BC and NAT groups was similar. However, LEfSe analysis revealed the differences in microbial communities between the two groups and further predicted their potential molecular functions. Notably, specific microbial community compositions were identified in BC tissues with different clinical and pathological characteristics. We characterized the microbial profiles associated with different levels of Ki-67. Specifically, we identified Brevundimonas as the predominant genus in the high H-Ki-67 group and showed that the glycolysis/gluconeogenesis pathway was significantly upregulated. The microbial environment of NAT may play an important role in the initiation and progression of BC. Microbial differences across subgroups may serve as potential bacterial biomarkers for prognosis, and mechanistic studies could further support the clinical translation of microbiome-based applications. Our study offers new research ideas and technical methods for exploring targeted microbial therapy strategies.

Acknowledgments

We would like to thank Biotree (http://www.biotree.com.cn/) for providing the 16S rRNA gene sequence and the bioinformatics analysis.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1570/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1570/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1570/prf

Funding: This work was supported by the Health Commission of Shanxi Province (No. 2024079).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1570/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The protocol was approved by the Ethics Committee of The Second Hospital of Shanxi Medical University (approval No. [2024] YX-460). All participants provided written informed consent before participation.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Chen L, Kang Z, Shen J, et al. An emerging antibacterial nanovaccine for enhanced chemotherapy by selectively eliminating tumor-colonizing bacteria. Sci Bull (Beijing) 2024;69:2565-79. [Crossref] [PubMed]
3. Hou T, Huang X, Lai J, et al. Intra-tumoral bacteria in breast cancer and intervention strategies. Adv Drug Deliv Rev 2025;217:115516. [Crossref] [PubMed]
4. Hieken TJ, Chen J, Hoskin TL, et al. The Microbiome of Aseptically Collected Human Breast Tissue in Benign and Malignant Disease. Sci Rep 2016;6:30751. [Crossref] [PubMed]
5. Chan AA, Bashir M, Rivas MN, et al. Characterization of the microbiome of nipple aspirate fluid of breast cancer survivors. Sci Rep 2016;6:28061. [Crossref] [PubMed]
6. Tzeng A, Sangwan N, Jia M, et al. Human breast microbiome correlates with prognostic features and immunological signatures in breast cancer. Genome Med 2021;13:60. [Crossref] [PubMed]
7. Costantini L, Magno S, Albanese D, et al. Characterization of human breast tissue microbiota from core needle biopsies through the analysis of multi hypervariable 16S-rRNA gene regions. Sci Rep 2018;8:16893. [Crossref] [PubMed]
8. Nejman D, Livyatan I, Fuks G, et al. The human tumor microbiome is composed of tumor type-specific intracellular bacteria. Science 2020;368:973-80. [Crossref] [PubMed]
9. German R, Marino N, Hemmerich C, et al. Exploring breast tissue microbial composition and the association with breast cancer risk factors. Breast Cancer Res 2023;25:82. [Crossref] [PubMed]
10. Urbaniak C, Gloor GB, Brackstone M, et al. The Microbiota of Breast Tissue and Its Association with Breast Cancer. Appl Environ Microbiol 2016;82:5039-48. [Crossref] [PubMed]
11. Neagoe CX, Ionică M, Neagoe OC, et al. The Influence of Microbiota on Breast Cancer: A Review. Cancers (Basel) 2024;16:3468. [Crossref] [PubMed]
12. Ragab HM, Samy N, Afify M, et al. Assessment of Ki-67 as a potential biomarker in patients with breast cancer. J Genet Eng Biotechnol 2018;16:479-84. [Crossref] [PubMed]
13. Wang W, Wu J, Zhang P, et al. Prognostic and predictive value of Ki-67 in triple-negative breast cancer. Oncotarget 2016;7:31079-87. [Crossref] [PubMed]
14. Li J, Hao C, Wang K, et al. Chinese Society of Clinical Oncology (CSCO) Breast Cancer guidelines 2024. Transl Breast Cancer Res 2024;5:18. [Crossref] [PubMed]
15. American Joint Committee on Cancer. TNM classification of malignant tumours. 8th ed. New York: Springer; 2017.
16. Nielsen TO, Leung SCY, Rimm DL, et al. Assessment of Ki67 in Breast Cancer: Updated Recommendations From the International Ki67 in Breast Cancer Working Group. J Natl Cancer Inst 2021;113:808-19. [Crossref] [PubMed]
17. Davey MG, Hynes SO, Kerin MJ, et al. Ki-67 as a Prognostic Biomarker in Invasive Breast Cancer. Cancers (Basel) 2021;13:4455. [Crossref] [PubMed]
18. Zhu X, Chen L, Huang B, et al. The prognostic and predictive potential of Ki-67 in triple-negative breast cancer. Sci Rep 2020;10:225. [Crossref] [PubMed]
19. Fuks G, Elgart M, Amir A, et al. Combining 16S rRNA gene variable regions enables high-resolution microbial community profiling. Microbiome 2018;6:17. [Crossref] [PubMed]
20. Caporaso JG, Kuczynski J, Stombaugh J, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010;7:335-6. [Crossref] [PubMed]
21. Douglas GM, Maffei VJ, Zaneveld JR, et al. PICRUSt2 for prediction of metagenome functions. Nat Biotechnol 2020;38:685-8. [Crossref] [PubMed]
22. Desalegn Z, Smith A, Yohannes M, et al. Human Breast Tissue Microbiota Reveals Unique Microbial Signatures that Correlate with Prognostic Features in Adult Ethiopian Women with Breast Cancer. Cancers (Basel) 2023;15:4893. [Crossref] [PubMed]
23. Wang H, Altemus J, Niazi F, et al. Breast tissue, oral and urinary microbiomes in breast cancer. Oncotarget 2017;8:88122-38. [Crossref] [PubMed]
24. Wei J, Li X, Xiang L, et al. Metabolomics and lipidomics study unveils the impact of polybrominated diphenyl ether-47 on breast cancer mice. J Hazard Mater 2020;390:121451. [Crossref] [PubMed]
25. Choi YK, Kang JI, Han S, et al. L-Ascorbic Acid Inhibits Breast Cancer Growth by Inducing IRE/JNK/CHOP-Related Endoplasmic Reticulum Stress-Mediated p62/SQSTM1 Accumulation in the Nucleus. Nutrients 2020;12:1351. [Crossref] [PubMed]
26. Nehme J, Maassen S, Bravaccini S, et al. Pharmacological CDK4/6 inhibition promotes vulnerability to lysosomotropic agents in breast cancer. EMBO J 2025;44:1921-42. [Crossref] [PubMed]
27. Yao L, Zi G, He M, et al. Asparagine endopeptidase regulates lysosome homeostasis via modulating endomembrane phosphoinositide composition. Cell Death Dis 2025;15:883. [Crossref] [PubMed]
28. Banerjee S, Wei Z, Tian T, et al. Prognostic correlations with the microbiome of breast cancer subtypes. Cell Death Dis 2021;12:831. [Crossref] [PubMed]
29. Meng S, Chen B, Yang J, et al. Study of Microbiomes in Aseptically Collected Samples of Human Breast Tissue Using Needle Biopsy and the Potential Role of in situ Tissue Microbiomes for Promoting Malignancy. Front Oncol 2018;8:318. [Crossref] [PubMed]
30. Ren X, Zheng L, Huang L, et al. The role of the gut microbiota in shaping the tumor microenvironment and immunotherapy of breast cancer. Front Microbiol 2025;16:1591745. [Crossref] [PubMed]
31. Mao X, Xu J, Wang W, et al. Crosstalk between cancer-associated fibroblasts and immune cells in the tumor microenvironment: new findings and future perspectives. Mol Cancer 2021;20:131. [Crossref] [PubMed]
32. Qin W, Chen B, Li X, et al. Cancer-associated fibroblasts secrete CSF3 to promote TNBC progression via enhancing PGM2L1-dependent glycolysis reprogramming. Cell Death Dis 2025;16:249. [Crossref] [PubMed]
33. Laborda-Illanes A, Sanchez-Alcoholado L, Dominguez-Recio ME, et al. Breast and Gut Microbiota Action Mechanisms in Breast Cancer Pathogenesis and Treatment. Cancers (Basel) 2020;12:2465. [Crossref] [PubMed]
34. Smith A, Pierre JF, Makowski L, et al. Distinct microbial communities that differ by race, stage, or breast-tumor subtype in breast tissues of non-Hispanic Black and non-Hispanic White women. Sci Rep 2019;9:11940. [Crossref] [PubMed]
35. Banerjee S, Tian T, Wei Z, et al. Distinct Microbial Signatures Associated With Different Breast Cancer Types. Front Microbiol 2018;9:951. [Crossref] [PubMed]
36. Chiba A, Bawaneh A, Velazquez C, et al. Neoadjuvant Chemotherapy Shifts Breast Tumor Microbiota Populations to Regulate Drug Responsiveness and the Development of Metastasis. Mol Cancer Res 2020;18:130-9. [Crossref] [PubMed]
37. Barba I, Carrillo-Bosch L, Seoane J. Targeting the Warburg Effect in Cancer: Where Do We Stand? Int J Mol Sci 2024;25:3142. [Crossref] [PubMed]
38. Geng C, Yan L, Li Y, et al. Layered Double Hydroxide Nanoparticles Loaded with Resveratrol Inhibit Glycolysis and Show Efficacy in the Treatment of Breast Cancer. Int J Nanomedicine 2025;20:3423-44. [Crossref] [PubMed]
39. Jiang C, Qian Y, Bai X, et al. SLC7A5/E2F1/PTBP1/PKM2 axis mediates progression and therapy effect of triple-negative breast cancer through the crosstalk of amino acid metabolism and glycolysis pathway. Cancer Lett 2025;617:217612. [Crossref] [PubMed]
40. Hicks E, Layosa MA, Andolino C, et al. Gluconeogenesis and glycogenolysis required in metastatic breast cancer cells. Front Oncol 2024;14:1476459. [Crossref] [PubMed]
41. Schrank BR, Wang Y, Wu A, et al. An antibody-toxin conjugate targeting CD47 linked to the bacterial toxin listeriolysin O for cancer immunotherapy. Nat Cancer 2025;6:511-27. [Crossref] [PubMed]
42. Halabian R, Jahangiri A, Sedighian H, et al. Staphylococcal enterotoxin B as DNA vaccine against breast cancer in a murine model. Int Microbiol 2023;26:939-49. [Crossref] [PubMed]
43. Ferriere F, Aasi N, Flouriot G, et al. Exploring the Complex Mechanisms of Isoflavones: From Cell Bioavailability, to Cell Dynamics and Breast Cancer. Phytother Res 2025;39:957-79. [Crossref] [PubMed]
44. Chakravarti B, Rajput S, Srivastava A, et al. A Systematic Review and Meta-Analysis of the Effects of Dietary Isoflavones on Female Hormone-Dependent Cancers for Benefit-Risk Evaluation. Phytother Res 2024;38:6062-81. [Crossref] [PubMed]
45. Liu YQ, Fu LL, Hong LH, et al. N-Heterocyclic Carbene Organocatalysis Enabled Modular Synthesis of Fluorinated Isoflavonoids to Suppress Proliferation and Migration in Breast Cancer Cells. Adv Sci (Weinh) 2025;12:e2413851. [Crossref] [PubMed]