Based on the World Health Organization (WHO) data, breast cancer (BC) is the most frequently diagnosed cancer in the world. In 2020, 2.3 million people were diagnosed with BC and 685,000 cases died from this disorder (Sung et al., 2021). It is estimated that by 2040, the number of newly diagnosed BC worldwide will increase by 40% and the number of deaths will escalate by 50% (Arnold et al., 2022). In Poland, the incidence and mortality of BC are a significant health problem. In 2020 represented 23% of all cancer cases, and was the second cause of cancer-related deaths among Polish women (Wojciechowska et al., 2023).

The human body is inhabited by a microbiota consisting mainly of bacteria, fungi, archaea, and viruses (Proctor et al., 2019). The composition and role of the intestinal microbiota is crucial in maintaining the homeostasis of the host organism. It is involved in a variety of physiological processes, including immune system modulation, metabolic homeostasis, and defense against pathogens, with its influence extending to distant organs and potentially impacting cancer development (Álvarez-Mercado et al., 2019, 2023; Vivarelli et al., 2019). In the context of BC, the microbiota may contribute to pathogenesis by regulating steroid hormone levels, energy metabolism, the synthesis of bioactive metabolites (e.g., genotoxins, lipopolysaccharides, and vitamins), and immune response modulation. It has been demonstrated that the bacterial enzyme β-glucuronidase (GUS) participates in estrogen reactivation, which may contribute to BC progression (Al-Ansari et al., 2021). Short-chain fatty acids (SCFAs), including acetate, propionate, and butyrate, may support cellular stability and modulate immune responses, potentially exerting protective effects. In contrast, secondary bile acids exhibit both carcinogenic and anti-carcinogenic properties, depending on the biological context and microbiota composition (Miko et al., 2019). The complex interactions between the microbiota and the human organism suggest that its disturbances, known as dysbiosis, may not only contribute to the development of BC but also influence its progression and treatment efficacy (Viswanathan et al., 2023). Research findings demonstrate that tumor tissues exhibit reduced microbial diversity compared to healthy tissues, along with an increased abundance of bacteria from the Pseudomonadaceae and Enterobacteriaceae families. In contrast, Propionibacterium and Staphylococcus are more frequently found in healthy tissues, suggesting their potential role in maintaining microenvironmental homeostasis (Tzeng et al., 2021).

For BC, several studies evaluated the presence of various bacteria in the human gut, breast tissue, and oral microbiome (Huybrechts et al., 2020; Samkari et al., 2022; Thu et al., 2023) with inconsistent findings. While Wang et al. (2017). and Nearing et al. (2020) found no differences in the distribution of oral microbiome between BC cases and healthy subjects, Wu et al. (2022) observed a lower abundance of Porphyromonas and Fusobacterium genus in BC cases.

The urinary microbiota has been less studied; however, recent studies showed its association with oncological disease (Wolfe and Brubaker, 2015; Perez-Carrasco et al., 2021), including bladder, colon, and prostate cancer. For BC, Wang et al. demonstrated increased amounts of Gram-positive bacteria, including Corynebacterium, Staphylococcus, Actinomyces and Propionibacteriaceae in the urine microbiota of BC cases (Wang et al., 2017).

Moreover, although fungi constitute a minor part of the microbiota community, their association with carcinogenesis has been established. Therefore, their composition and abundance could be important in BC cases. For example, Candida albicans can support cancer progression (Ramirez-Garcia et al., 2013).

The current research investigated the composition of urinary and oral microbiota in Polish BC patients relative to healthy subjects, marking Poland’s first study of its kind. The findings may help identify potential biomarkers and offer new prevention or treatment options for BC through microbiota intervention.

A total of 48 participants were included in this study: 24 women with BC and 24 HC. The characteristics of the study groups are shown in Table 1 and were reported during the interview. BC cases were similar to healthy controls with regard to all presented parameters (p > 0.05).

In the NGS analysis for urine and oral metaprofiling, the obtained parameters were as follows: the flow cell cluster density 1,006 ± 32, the Phred base-calling score GQX ≥ 30, an average of 74.87%, and the passing filter (PF) reads 22,703,315 constituting 92.25% of the total reads 24,597,294. In the bioinformatics analysis, the number of input reads per sample was on average 410,976 [SD 68509.5 (median 402,000, min. 278,462, max. 698,298)], and the number of reads after filtering on length, ambiguity, and expected errors was at the level of 24639.6 (SD 6637, median 24,318, min. 12,764, max. 47,286). Refraction curves of amplicon sequence variant (ASV) richness at read depth are presented in Figure 1A. In our study, both urine and oral rinse samples of BC patients and controls demonstrated similar microbial taxonomic diversity based on the deepest available analysis of all regions V1-V9 of the 16S rRNA gene (Figures 1B,C). Analysis of the beta diversity determined by the Bray–Curtis dissimilarity revealed that the bacterial microbiota of urine samples create a separate cluster from oral microbiota samples (Figure 1D).

Based on a total number of identified ASVs, the urine samples presented a similar richness of microbiota compared to oral rinses, 12,086 vs. 12,035, respectively, of which 492 ASVs are found in both types of samples, giving a total richness of 24,613 ASVs. To narrow down the number of ASVs and to exclude sporadically occurring microorganisms, only ASVs that were observed in at least 20% of the samples within each group were counted (Figure 2).

After applying this criterion, urine samples presented 5 unique ASVs and 9 oral rinses, and none were joint. For urine, BC cases showed more unique ASVs than controls (24 vs. 5), with 2 ASVs present in both groups. In oral rinses, BC patients demonstrated also a richer composition of unique ASVs (11) compared to healthy controls (1) (Figure 2).

NGS raw data are available in the SRA database (accession number PRJNA1180501).

In our study, metagenomic profiling based on next-generation sequencing allowed us to classify bacteria and fungi inhabiting the urine and oral cavity of BC patients and healthy participants.

Despite contamination of urine samples by bacteria originating from the uroepithelium, urethra, or genital tract, analyzing urine microbiota can provide valuable health insights into its role in disease development (Gottschick et al., 2017; Perez-Carrasco et al., 2021). Previous research on the urine microbiota indicated its lower richness compared to the number of bacteria inhabiting other niches of the human body (Pearce et al., 2014), a finding confirmed in our study, where urine samples had fewer ASVs than oral rinse samples. This is probably the result of greater external environmental pressure in the case of oral wash samples compared to urine samples.

We observed a slight increase in abundance (not significant statistically) of Proteobacteria phylum in the group of BC patients (34.5%) compared to healthy controls (27.7%) (Figure 4A). This phylum includes a wide variety of potentially pathogenic genera Escherichia, Klebsiella, and Shigella, whereas Firmicutes were presented decreased in BC cases to healthy controls, with an abundance of 25.8 and 35.3%, respectively. In general, the Proteobacteria increase with Firmicutes depletion is a microbial signature of dysbiosis in gut microbiota (Kumar and Kumar, 2022). Furthermore, in urine BC samples, we detected a significant increase in the genera Pseudomonas (p = 0.0257) and Gammaproteobacteria (p = 0.0424) compared to controls. This was also observed at the species level, although the fold change was not so high (Figures 4E,F). There are reports that some strains of bacteria, including Pseudomonas, produce toxins that can affect host cells, damaging their DNA and causing genetic changes that promote the development of cancer. On the other hand, there is evidence that cancer cells can stimulate the secretion of toxins by bacteria, such as those from the Pseudomonas genus (Choi et al., 2023). When considering the potential link between Gammaproteobacteria and BC, we came across research showing that tumor-adjacent-normal tissue was found to contain higher relative abundances of Gammaproteobacteria (unclassified) in comparison with healthy normal tissue (Giallourou et al., 2021). The exact role of this bacteria in BC development is still unclear; however, it is known that Gammaproteobacteria is associated with a specific lipid profile.

The highest fold change (5.2-fold) with a predominance in BC was presented for Corynebacterium. Although most species of Corynebacterium are symbiotic to the host, some of them may become pathogenic and lead to infections (Burkovski, 2018). Importantly, in urine samples of healthy women, we observed a higher content of probiotic bacteria from the Bifidobacteriaceae and Lactobacillaceae families compared to BC. Moreover, BC patients had an increased abundance of Micrococcales in their urine. Micrococcales belongs to the Actinomycetes class, which in general are commensal organisms and can also be an opportunistic pathogen, particularly in a host with an impaired immune system (Li et al., 2022). The presence of the Micrococcales bacteria has also been reported in a study analyzing microbiota composition in BC and healthy tissue in Slovakian and Chinese women (Hadzega et al., 2021). In that study, a higher level of Micrococcales was observed in the breast tumor tissue of Slovakian women in comparison with healthy breast tissue samples but not for the Chinese women. The observed enrichment of Micrococcales in the breast tissue among the Slovakian BC cases and in urine among Polish BC cases, two similar ethnic groups, makes this bacterium a possible candidate as a biomarker for BC in the Polish population.

At this moment, it is difficult to clearly interpret the role of bacteria that significantly distinguish BC patients from healthy women, calling them radical pathobionts or commensals. As defined by the International Cancer Microbiome Consortium, there is a large and poor-defined space between a “pathogen” coexisting with the disease of the host (e.g., Helicobacter pylori) and the “commensal” which may contribute to disease under certain conditions. Therefore, the term “dynamic symbiont” was introduced and is often used, also in our study, when interpreting the bacterial genera, such as Micrococcales, Pseudomonas, and Gammaproteobacteria which were enriched in the urine of BC women (Scott et al., 2019).

Searching in our study for the microbial signature of BC in oral rinses, we have observed a significant depletion of bacterial species Abiotrophia para-adiacens, Cellulosimicrobium cellulans LMG 16121, Pseudomonas sp. (uncultured), and Pantoea agglomerans in BC cases compared to healthy controls.

To discover the potential functional role of bacteria in the host, we applied metabolomic analysis by PICRUSt2 software to predict relevant pathways of urine and oral cavity microbiota in BC patients and HC. Of the 1,352 identified pathways based on interfered Enzyme Commission (EC) profiles in urine samples and 1,323 in oral rinses, 10 pathways together from both niches differentially enriched between the BC patients and HCs were indicated (Supplementary Tables S15, S16). Based on the urine microbiota composition, we have identified three processes that mostly differentiate the two studied groups, benzoyl-CoA degradation III and hydrogen production III and VIII, which dominate in BC cases. Urinary hydrogen peroxide was postulated to be a biomarker of oxidative stress in general malignancies (Banerjee et al., 2003), whereas increased benzoyl-CoA degradation III (anaerobic) in urine samples of BC patients could be the effect of different factors. Benzoyl-CoA is the most common central intermediate of anaerobic aromatic metabolism, including those aromatic compounds derived from diet, drugs such as chemotherapy, or tumor metabolism (Koch et al., 1993). However, based on the oral rinse microbiota, we have observed that m-cresol and L-lysine degradation was significantly increased in BC cases. There are speculations that cancers alter aromatic compound metabolism leading to an increased degradation process of m-cresol, but there is a lack of evidence. Studies on full microbial metabolites are needed to elucidate these process correlation with BC development.

Fungi were remarkably less present in the tested samples of both urine and oral rinses in our study. In general, we have detected two main fungal phyla, Ascomycota and Basidiomycota, and also Incertae sedis phylum. In the case of oral rinses, we did not find any relevant differences in the composition of mycobiota between patients and healthy women. However, urine samples of BC cases showed a significant reduction in Basidiomycota abundance and enrichment in Agaricomycetes, Capnodiales, Cladosporiaceae, and Cladosporium abundance, whereas global investigations revealed a strong correlation of particular fungi presence with cancer development (Hosseini et al., 2022). Among them, Aspergillus, Candida, Coccidioides, Cunninghamella, Geotrichum, Pleistophora, Rhodotorula, Filobasidiella, Mucor, Trichophyton, Epidermophyton, Fonsecaea, Pseudallescheria, Penicillium, Ajellomyces, Alternaria, Rhizomucor, Piedraia, and Malassezia were found to be linked with breast cancer development (Banerjee et al., 2018). However, these observations concern microbial niches such as intestines or cancer tissue. In our study, we were focused on microbial composition in oral rinses and urine only.

To summarize, the current study has shown bacterial differences in the urinal and oral microbiota composition between BC patients and healthy Polish women. Although the identified differences in bacterial abundance did not maintain statistical significance after FDR adjustment, they suggest trends for further analysis using the full PWHS dataset. We acknowledge the pilot study’s small sample size limits the statistical power. The obtained results will guide us in larger investigations, in which we plan to perform metabolomic analysis parallel to metagenomics profiling in the full sample of 514 HCs and 417 BC cases. Metabolomics data would be a valuable complement to our research, particularly due to estrobolome characterization.

Our study represents the first attempt to compare groups of women that do not differ on age, menopausal status, smoking and drinking habits, and BMI. In previous studies on the diversity of urine and/or oral microbiota in patients with BC conducted by Wu et al. (2022)and Wang et al. (2017), significant differences in the selection of the study and control groups were observed on age, menopausal status, mean BMI, alcohol, tobacco, and antibiotics intake, which authors indicate could have influenced their results. In a research by Nearing et al., although has a fairly extensive description of the study population available, there is a lack of data regarding menopausal status (Nearing et al., 2020). By controlling for age, menopausal status, BMI, smoking, and alcohol consumption, we minimized their impact, allowing us to detect subtle microbiota changes. These findings may have significant implications for the health of women in Poland, suggesting potential new directions for research into BC prevention, diagnostics, and treatment.