Increased gut microbiota diversity in women with uterine fibroids: Insights from a pilot study on gut and reproductive tract microbiota.

Uterine fibroids (UFs) are the most common benign tumors in women of reproductive age. They may affect even over 70% of women worldwide [ 1 ]. The risk factors for UFs include age, obesity, positive familial history, low levels of vitamin D, and endogenous and exogenous hormonal factors. UFs are heterogeneous as regards their number, composition and size [ 2 ]. They may require surgical treatment and are a major source of gynecological and reproductive dysfunction [ 1 ]. UFs constitute a significant health impact for the affected women and a financial burden for healthcare systems. The human gut hosts approximately 100 trillion bacteria that play a key role in digestion and nutrient absorption [ 3 – 5 ]. Numerous intestinal microbes influence the physiological functions of the host and affect the synthesis and secretion of hormones, trace elements, growth factors and immune system functions [ 6 ]. The intestinal flora may be modified by hormonal interactions both in vitro and in vivo affecting the biological balance of the body [ 7 ]. UFs, which are hormone-dependent tumors, are primarily influenced by estrogen and progesterone [ 8 ]. Several studies indicate that the gut microbiota are involved in estrogen metabolism through enterohepatic circulation [ 9 ]. Disruptions in the composition of the gut microbiome may affect hormone levels, potentially influencing the development and growth of UFs. Changes in the gut microbiome can alter the reabsorption and excretion of estrogen, impacting estrogen-dependent conditions such as UFs [ 10 ]. The results suggested that the systemic distribution of gut bacteria extended to the patients’ fibroids, likely following dysbiosis or impaired intestinal barrier [ 10 ]. The use of next-generation sequencing (NGS) of hypervariable fragments of the 16S rRNA bacterial gene revealed the presence of numerous microorganisms within the uterine cavity, which were usually found in the vagina and in the colon [ 11 , 12 ]. However, compared to the microbiomes of the uterine cervix and vagina, the level of the bacterial colonization of the uterine cavity is relatively low [ 13 ]. Previous research allowed the determination of the so-called “healthy” microbiome of the vaginal and uterus in women of reproductive age, with the dominant Lactobacillus (LD, Lactobacillus -dominant) species, with the abundance of Lactobacillus spp . above 90% and without the dominance of Lactobacillus (NLD, non- Lactobacillus dominant), with the abundance of Lactobacillus spp . below 90% [ 14 ]. Hormonal changes affect the vaginal and endometrial microbiota. Exogenous progestogens significantly alter the microbiota of the endometrium, including the reduction of the diversity of Lactobacillus spp. [ 15 ]. To date, there is a scarcity of studies investigating the profiles of gut and uterine microbiota in women with UFs. Moreover, to the best of our knowledge, no research has comprehensively analyzed the microbial composition of endometrial tissues in this patient population. Existing literature has largely overlooked the potential relationship between site-specific microbial dysbiosis and the pathophysiology of UFs, particularly in relation to bacterial metabolites. Addressing this gap, our study offers a novel and integrated approach by analyzing both local (uterine cavity and cervix) and systemic (gut) microbiota, alongside microbial metabolites. Therefore, the aim of this study was to comprehensively characterize the microbiota composition at three anatomical sites – the uterine cavity, cervix, and stool – as well as to quantify selected stool bacterial metabolites in women with UFs compared to fibroid-free controls, in order to identify potential microbiota differences associated with the presence of uterine fibroids.

We recruited 55 patients for the study. Finally, material collected from 51 patients, including 29 from the group with uterine fibroids (UFs) and 22 from the control group, met the qualitative and quantitative criteria for inclusion in further analysis. During recruitment, a total of 28 and 19 cervical swabs, 21 and 16 samples of pathologically altered endometrial tissue, and 25 and 17 samples of normal endometrial tissue were collected from UF patients and the control groups, respectively, as well as 8 stool samples from each group. The clinical characteristics of patients participating in the study are presented in Table 1 . The mean age of participants was 44.9 ± 11.01 years in the UF group and 45.2 ± 11.37 years in the control group. Median age was 44.1 and 44.0 years, respectively. The majority of patients in both groups were premenopausal; specifically, 26 out of 29 UF patients and 20 out of 22 control patients were premenopausal. Three UF patients and two control patients were perimenopausal. None of the participants were postmenopausal. Use of hormonal contraception within the last 12 months was reported by 5 patients in the UF group and 4 in the control group. Polycystic ovary syndrome (PCOS) or endometriosis were noted in 2 patients with UFs; neither diagnosis occurred in the control group. Seven patients in each group had never been pregnant. Primiparity was reported by 13 patients in the UF group and 11 in the control group. At least one miscarriage was reported by 6 UF patients and 5 control patients. Various forms of infertility treatment had been conducted in 6 UF patients and 2 from the control group. A history of inflammation (e.g., recurrent vaginal infections, adnexitis, endometritis) was reported by 7 patients with UFs and 4 in the control group. Treatment with antibiotics up to 6 months prior to the procedure was reported by 6 patients in the UF group and 4 in the control group. The antibiotics listed by the patients included: rifaximin, azithromycin, amoxicillin with clavulanic acid, cefuroxime, metronidazole, and tetracycline. Vitamin D supplementation was noted in 17 UF patients and 9 control patients. A total of 21 patients with UFs and 13 in the control group had a history of procedures or surgeries, including abdominal laparoscopic surgeries, laparotomies due to gynecological or surgical indications, gynecological procedures such as hysteroscopies or curettages, as well as soft tissue operations. To reduce potential confounding, control participants were recruited to closely match UF patients in terms of age and reproductive history, and were confirmed to be free from fibroids or other uterine pathologies via clinical and ultrasonographic examination. UF, uterine fibroid; SD, standard deviation. There were, on average, 88954, 90347, 67819 and 45494 sequencing reads generated for cervical swabs, normal endometrium tissues, pathologically altered endometrium tissues and stool samples, respectively. From 2666 (normal endometrium tissue) to 641 (stool) bacterial species were present, with 383 (normal endometrium tissue) to 102 (cervical swab) species being present in more than 0.01% of reads ( Fig 1 ). The 16S rRNA bacterial gene sequencing of DNA isolated from stool samples showed significantly higher bacterial α-diversity, as measured by the Shannon index ( p  = 0.043), in the group of women with UFs compared to the control group ( Fig 2A ). However, we found no significant differences between the groups both in terms of the richness of bacterial species measured with the Chao index ( Fig 2B ) and their β-diversity ( Fig 2C ). We identified 15 species differentiating the group of women with UFs from the control group at the level of p -value < 0.05 significance. Eight and seven of them were more abundant ( Prevotellaceae NK3B31 group ) and less abundant ( Bifidobacterium ), respectively, in women with UFs ( S1 Table ). However, for none of the identified species statistical significance reached the level of the p adj -value < 0.05. When analyzing samples of cervical swabs, no significant difference between the groups at both α- and β-diversity levels was shown ( Fig 3 ). The identified possibly differentiating bacteria (significance at p  < 0.05, without adjustment), were less abundant in swab samples obtained from women with UFs than control group ( S2 Table ). However, searching for the Lactobacillus species colonizing the cervical canal, we found that Lactobacillus iners was significantly ( p adj  = 0.05; p  = 0.0035) overrepresented in the cervix of women with UFs. Subsequently, we compared the microbiota inhabiting the endometrial tissue. This was the first time that the microbiome composition of pathologically altered endometrial tissues was compared between a group of women with UFs and the control group. Similarly to cervical microbiome, no significant difference between the groups was shown at both α- and β-diversity levels ( Fig 4 ), However, 62 bacterial species were identified differentiating pathologically altered tissue samples of women with UFs and the control group, at the level of p  < 0.05 significance. Of these, 22 ( Atopobium, Ezakiella ) and 40 ( Ilumatobacter, Rhodocyclaceae C39 ) species were more and less represented in the group of women with UFs, respectively ( S3 Table ). Among Lactobacillus species, Lactobacillus curvatus , was less abundant in endometrial tissues from patients with UFs, but at the significance level without adjustment. Similarly to the comparison of the pathologically altered endometrial tissue, no significant intergroup differences were noted as regards the lesion-free endometrial tissue both at the level of the α- and β-diversity ( Fig 5 ). Also, at the level of p  < 0.05 statistical significance, we identified 28 species which differentiated the lesion-free tissue samples from women with UFs and control group. Half of them were more abundant ( Atopobiaceae unclassified ) and half were less abundant ( Acidovorax ) in the group of women with UFs ( S4 Table ). In this tissue, we found no Lactobacillus species significantly differentiating the UF patients and control groups. Next, we compared the relative concentrations of stool metabolites between UF patient and control groups. However, no statistically significant differences were obtain between the compared groups, both for SCFAs ( Fig 6 ) and AAs ( Fig 7 ). Nevertheless, a distinct trend of either increasing or decreasing concentrations of the studied metabolites was observed.

Despite the predominantly negative findings, our study identified several noteworthy outcomes that underscore the potential of this research path. Specifically, Lactobacillus iners was fund to be overrepresented in cervical swab samples from women with UFs, and stool samples from patients with UFs exhibited higher Shannon-indexed α-diversity compared to the control group. These observations suggest a possible link between microbial diversity and UFs, offering a valuable starting point for further exploration. While these findings are preliminary, they highlight the importance of continuing research in this area. Investigating the role of the microbiome in UFs may provide deeper insights into their pathogenesis, open new diagnostic opportunities, and potentially lead to innovative therapeutic strategies. This promising direction deserves sustained attention and more comprehensive studies. Future research should include larger, demographically homogeneous cohorts and apply longitudinal study designs to monitor microbiota dynamics over time and in response to treatment. Incorporating multi-omics approaches such as metagenomics, metatranscriptomics, and metabolomics could help elucidate functional microbial roles and host–microbiome interactions. Additionally, studies should account for hormonal status, dietary habits, and other lifestyle factors known to influence microbiota composition. The potential of microbial markers like L. iners as indicators of reproductive tract instability or early UF risk also warrants further investigation.

The study was conducted in accordance with the guidelines of the Helsinki Declaration of 1964 and approved by the Bioethics Committee at the Center of Postgraduate Medical Education, no. 63/2022. All participants gave their informed consent to participate in the study. The study was conducted in cooperation with the 2nd Department of Obstetrics and Gynecology based in the Warsaw Institute of Women’s Health and the Department of Gastroenterology, Hepatology and Clinical Oncology, Center of Postgraduate Medical Education in Warsaw. The study group included women of Polish origin (Caucasian race) who had given written informed consent before being included in the study. The inclusion criteria were: women aged over 18 with gynecological indications for hysteroscopy. The exclusion criteria were: pregnancy and lactation, a history of bowel surgery requiring an intestinal stoma, chronic immunosuppression, oncological treatment, contraception, menopausal hormone therapy, hormonal treatment for endometrial hyperplasia up to one month before the procedure, hormonal intrauterine device (IUD), abnormal cervical cytology result, use of antibiotics up to 30 days before the procedure. The patients were recruited between June 2022 and June 2023. The study group included patients with UFs who underwent hysteroscopy. The control group included patients without UFs requiring hysteroscopy for other reasons, such as cervical and uterine polyps, abnormal bleeding from the genital tract or infertility. In patients undergoing hysteroscopy, specimens of the endometrium covering a possible pathological lesion in the uterine cavity and specimens of the normal endometrium located at a distance of at least 1 cm from any pathology of the uterine cavity were collected into sterile Eppendorf tubes. Prior to hysteroscopy, cervical swabs were collected from patients using 4N6FLOQSwabs™ (Thermo Fisher Scientific, USA). Endometrial samples were collected using sterile, single-use sampling brushes under direct visualization with the aid of a sterile speculum. Prior to insertion, the cervical os was gently cleansed with sterile saline to reduce potential contamination from vaginal or cervical flora. The sampling instrument was carefully inserted transcervically without contact with the vaginal wall. All procedures were performed by experienced gynecologists following strict aseptic protocols. To monitor for contamination, blank swabs (negative controls) were collected during each sampling session and processed alongside clinical samples. Moreover, the patients were asked to provide a stool sample in sterile containers. The collected material was secured and frozen at -20 o C. Genomic bacterial DNA was isolated from the collected uterine tissues and cervical swabs using the commercial QIAamp DNA Mini Kit (Qiagen, Germany). The QIAamp Fast DNA Stool Mini Kit (Qiagen, Germany) was used to isolate DNA from fecal material [ 6 ]. Subsequently, the construction of bacterial 16S rRNA gene libraries was created using the Ion 16S™ Metagenomics Kit (Thermo Fisher Scientific, USA) and the Ion Plus Fragment Library Kit (Thermo Fisher Scientific, USA). As part of contamination control, negative extraction controls and blank swabs were included during DNA extraction and library preparation. Additionally, a commercially available mock microbial community (ZymoBIOMICS Microbial Community Standard, Zymo Research, USA) was used as a positive control to validate sequencing accuracy and detect potential technical biases. Internal control steps were also included during library construction to monitor quality and ensure reproducibility. The constructs were sequenced using Ion Torrent technology with the Personal Genome Machine (Thermo Fisher Scientific, USA) using PGM™ Hi-Q™ View Sequencing Kit reagents (Thermo Fisher Scientific, USA), as described previously [ 16 ]. Then, a bioinformatic analysis was performed to assess the composition of the cervical microbiota, endometrial tissue biofilm and stool. Raw BAM files were transformed into FASTQ format using the SamToFastq utility from the Picard toolkit. Subsequent analyses were performed using Mothur version 1.43. The resulting FASTQ files were then converted into the PASTA format for further processing [ 17 ]. Sequences ranging from 200 to 300 base pairs in length were selected, ensuring an average base quality score of at least 20 within a 50-base sliding window and limiting homopolymer runs to a maximum length of 10. Chimeric sequences were detected using the vsearch algorithm with default settings, referencing an internal sequence database [ 18 ]. Chimeric sequences were excluded, and the remaining 16S rRNA sequences were taxonomically classified employing the Wang method with the SILVA bacterial 16S rRNA database as the reference [ 19 ]. The bootstrap threshold was set at 80%. Alpha-diversity analysis was performed using the Shannon and Chao indices [ 20 ]. Principal Coordinate Analysis (PCoA) was conducted utilizing the Bray-Curtis distance metric [ 21 ]. The significance of the observed clustering patterns was evaluated using the ANOSIM test. The LinDA approach, applied with default configuration, was used to evaluate taxonomic abundance differences [ 22 ]. The Mann-Whitney U test was used to assess differences in diversity indices between control samples and fibroid samples, while the paired Wilcoxon test was used to identify statistically significant differences in paired patient samples. After controlling for the false discovery rate (FDR), adjusted p -value ( p adj ) of less than 0.05 was viewed as significant. Species-level identification of Lactobacillus was achieved by mapping the reads against the Greengenes v13_8 reference database [ 23 ]. In addition, concentrations of bacterial metabolites, including short chain fatty acids (SCFAs) and amino acids (AAs) were examined using gas chromatography mass spectrometry (GC-MS). A 100 mg portion of stool was placed into a 2 ml tube containing ceramic beads (Ohaus Corporation, Parsippany, NJ, USA), specifically intended for environmental sample analysis, and then thoroughly homogenized using mechanical disruption. The study used commercial calibration standards for SCFAs (formic acid, acetic acid, propanoic acid, butyric acid, isobutyric acid, pentanoic acid, isocaproic acid, and hexanoic acid) and AAs (alanine, glycine, valine, leucine, isoleucine, proline, methionine, phenylalanine, tyrosine) (Sigma-Aldrich, USA). Derivatization of both samples and standards was performed using isobutyl chloroformate as the reagent [ 24 ]. The analysis was conducted using a GC 7890 system coupled with an Agilent 7000D Triple Quadrupole MS and G4513A autosampler, utilizing a VF-5ms capillary column (30 m × 0.25 mm × 0.50 μm). Data acquisition in full scan mode (m/z 15–650) was performed at 4.9 scans/s and processed using MassHunter software (Agilent Technologies, USA) [ 24 ].

(XLSX) (XLSX) (XLSX) (XLSX)