Dynamics of Microbiome Changes in the Endometrium and Uterine Cervix during Embryo Implantation: A Comparative Analysis

Background

Fertility and pregnancy disorders affect an increasing number of people around the world. They may be associated with anatomical anomalies of the reproductive organs, both congenital and acquired [1,2]. Hormonal balance is also important, as well as the presence of inflammatory processes that can lead, for example, to obstruction of the fallopian tubes [3]. Eating disorders, sleep disorders, and depression are also believed to have a negative impact on fertility. Problems with getting pregnant are a source of stress for many couples, which in turn can be the cause of idiopathic infertility [4,5].

Assisted fertilization techniques (ARTs), including in vitro fertilization (IVF), offer hope for parenthood. They are constantly being improved in order to maximize the chance of successful conception and live birth, reducing the number of possible complications [6,7]. Studies indicate that the IVF result is mainly influenced by the woman’s age, the number of oocytes, the length of infertility, and the level of basal (follicle-stimulating hormone) FSH [8]. However, additional factors that may affect the success of IVF are sought. One of them may be the bacterial microbiome of the female reproductive system [9].

The composition of the human microbiome is related to the state of health. Its change can lead to the expansion of certain bacteria or pathogens, which is referred to as dysbiosis [10]. Vaginal Lactobacilli and lactoferrin play a significant role in maintaining microbial homeostasis. During dysbiosis, a reduced level of Lactobacilli is observed with a simultaneous increase in the number of endogenous anaerobic bacteria. Interestingly, high levels of lactoferrin promote innate and adaptive immune responses [11]. The invasion of pathogens and progressive dysbiosis are conducive to gynecological diseases such as endometriosis, endometritis, pelvic inflammatory disease, and cancer [12,13]. Remodeling of the bacterial microbiome during pregnancy is thought to be related to a status change of the immune system to allow for immunological and metabolic adaptations leading to a successful pregnancy [14]. Lactobacilli make up the majority of vaginal microbes, affecting fertility and length of pregnancy [15]. Interestingly, increasing their number may improve implantation during IVF [13].

Therefore, this study aimed to evaluate the cervical and endometrial bacterial microbiome in 177 women treated for infertility before, during, and after embryo implantation, and the outcomes.

Material and Methods

ETHICS:

This study was performed according to the guidelines of the Declaration of Helsinki and was approved by the Institution of the Bioethical Committee operating at the Regional Medical Chamber in Krakow (No. 161/KBL/OIL/2021). Informed consent for participation in the study and publication of this article was obtained from all patients.

PATIENTS:

Table 1 presents the inclusion and exclusion criteria, which are the same as in our previous publication [16].

Out of 250 women diagnosed with infertility who were qualified for the in vitro procedure at the Gyncentrum Clinic in Poland, 177 patients (71%) were enrolled for the study, from whom cervical and endometrial swabs were taken at 3 time points. Based on the assessment of beta-human chorionic gonadotropin (beta hCG) levels 14 days after embryo implantation, 67 women were pregnant. In 65 patients (97%) the pregnancy ended in childbirth, while the remaining 2 suffered a miscarriage. The characteristics of the patients included in the study are listed in Table 2. Measurable data were presented as mean and standard deviation (X±SD) and median (Me) and quartiles 1–3 (Q1–Q3).

MATERIAL COLLECTION:

Cervical and endometrial swabs were collected at 3 time points: (1) during the initial examination and Endometrial Receptivity Analysis (ERA) test, approximately 1 month before the planned embryo implantation, (2) during implantation, (3) 10–14 days after embryo implantation during routine diagnostic tests. At the third time point, the material for molecular testing was taken only from the cervix, as collection from the endometrium would be invasive. Swab Collection and DNA Preservation Tube Transport Medium (Norgen Biotek Corp., Thorold, ON, Canada) were used for cervical and endometrial swab collection.

DNA ISOLATION:

DNA isolation was performed using the NucleoSpin Tissue kit with glass beads (NucleoSpin Bead Tubes Type B, Macherey-Nagel, Oensingen, Switzerland), according to the manufacturer’s recommendation. The obtained DNA extracts were assessed qualitatively with 1% agarose gel electrophoresis. The concentration and purity were determined with spectrophotometric measurement (Nanodrop, Thermo Fisher Scientific, Waltham, MA, USA). The concentration of the extracts was evaluated at the wavelength of 260 nm, and their purity by determining the ratio of absorbance at the wavelength of 260 nm to 280 nm (standard 1.8–2.0). This allowed the qualification of DNA extracts for the analysis of cervical and endometrial bacterial microbiome using the next-generation sequencing (NGS) technique.

NEXT-GENERATION SEQUENCING:

The libraries were prepared according to the Ilumina 16S Metagenomic Sequencing Library Preparation (16S Sequencing) protocol, and their indexing was performed using the Nexter XT Kit (Illumina, San Diego, USA), as recommended by the manufacturer. The fragments were then purified using a MagSi-NGS apparatus (Magtivio, HK Nuth, The Netherlands) and analyzed using the Fragment Analyzer (Agilent, Santa Clara, CA, USA). The QuantiFluor ONE dsDNA Kit (Promega Corporation, Madison, WI, USA) was used to measure the concentration of libraries on a Quantus™ fluorometer (Promega Corporation, Madison, WI, USA) and normalized to 4 nM. Sequencing was performed on the Illumina Miseq 2×300 bp platform, and the analysis of the obtained sequences was based on the resources of the EzBioCloud platform (EzBiome Inc., Gaithersburg, MD, USA). The Usearch algorithm was used to taxonomically identify the bacteria down to the species level. Results were validated using the ZymoBIOMICS Microbial Community Standard Microbial Controls (Zymo Research, Irvine, CA, USA).

STATISTICAL ANALYSIS:

Statistical analysis was performed using Statistica 13.3 PL software (StatSoft, Cracow, Poland) and R 3.5.1 statistical software. The level of significance was set at α=0.05. In the case of measurable data, after the Shapiro-Wilk distribution normality test, further statistical analysis was performed using the Kruskal-Wallis and Dunn’s post hoc tests or Friedman’s and post hoc tests. For non-measurable data, chi-square (χ²) analysis was performed.

Results

BACTERIAL MICROBIOME COMPOSITION OF THE CERVIX AND ENDOMETRIUM BEFORE, DURING, AND AFTER EMBRYO IMPLANTATION:

A total of 105 strains of bacteria were distinguished during the observation (Table 3). The statistical analysis showed only 1 significant change in the number of patients in whom Escherichia coli was identified in the microbiome of the cervix and endometrium (P=0.03). This screening allowed for the selection of 10 strains of bacteria for further analysis, which were present in at least 2 samples in at least 6 patients: Bifidobacterium longum, Escherichia coli, Gardnerella vaginalis, Lactobacillus gasseri, Lactobacillus helveticus, Lactobacillus iners, Lactobacillus paracasei, Lactobacillus reuteri, Lactobacillus jensenii, and Staphylococcus aureus.

In addition, there were no significant changes between the selected bacterial strains and the place of residence (Table 4), BMI (Table 5), and the type of treatment (Table 6).

CORRELATION ANALYSIS BETWEEN THE CONTENT OF SELECTED BACTERIAL STRAINS BEFORE, DURING, AND AFTER EMBRYO IMPLANTATION:

Next, we assessed whether there was a statistically significant relationship between the selected strains of bacteria (Table 7).

In swabs obtained during the first collection, we showed significant negative relationships between the percentage of Gardnerella vaginalis and Lactobacillus helveticus (r=−0.86); Lactobacillus gasseri and Lactobacillus helveticus (r=−0.73); Lactobacillus jensenii, and Lactobacillus helveticus (r=−0.84). In swabs obtained during the second collection, there was only 1 statistically significant relationship, between Lactobacillus helveticus and Lactobacillus jensenii (r=−0.84). On the other hand, in swabs obtained during the third collection, significant relationships were found between: Escherichia coli and Lactobacillus helveticus (r=−0.90); Lactobacillus gasseri and Escherichia coli (r=−0.78); Lactobacillus helveticus and Lactobacillus iners (r=−0.76); and Lactobacillus helveticus and Lactobacillus jensenii (r=−0.72).

We additionally assessed the correlations between the discussed bacterial strains and the patients’ anthropometric data and the period of infertility (Table 8).

In the first swab collection, significant relationships were observed between age and the percentage of Lactobacillus paracasei. Moreover, the content of Lactobacillus jensenii was correlated with age in both the first and second collections. The percentage of Lactobacillus iners was correlated with the patient’s weight during the first collection.

BACTERIAL MICROBIOME DIVISION BEFORE, DURING, AND AFTER EMBRYO IMPLANTATION:

In the next step, the patients were divided according to the type of flora: normal, mild/moderate dysbiosis, and dysbiosis (Figure 1).

A similar distribution of the bacterial microbiome was observed in all 3 swab collections. The largest percentage were patients with normal microbiome, while the highest dysbiosis was noted for the second collection. The metric profile of the patients was also analyzed in relation to the nature of the microbiome (Table 9).

The analysis showed that the BMI significantly affected the type of the patient’s bacterial microbiome in the second swab collection. Then, we checked whether there were any relationships between the clinical profile of the patients and the type of microbiome (Table 10).

An association was observed between the use of metronidazole and the nature of the microbiome at the second swab collection. In addition, the type of microbiome at the third collection depended on fertilization methods other than insemination, cryotransfer, and embryo transfer.

Discussion

In our study, we determined the bacterial microbiome composition of the endometrium and uterine cervix at 3 time points: before, during, and after embryo implantation. The micro-organisms most common in patients were Bifidobacterium longum, Escherichia coli, Gardnerella vaginalis, Lactobacillus gasseri, Lactobacillus helveticus, Lactobacillus iners, Lactobacillus paracasei, Lactobacillus reuteri, Lactobacillus jensenii, and Staphylococcus aureus. Interestingly, in the case of E. coli, the percentage of patients increased in the second and third swab collection compared to the first, which was statistically significant. In addition, for the third collection, we observed a significant negative correlation between E. coli and L. helveticus and L. gasseri as the percentage of E. coli decreased between the second and third collection, while the level of Lactobacilli increased. Our results also showed that a normal bacterial microbiome was dominant in all 3 swab collections.

Among the bacteria that contaminate the IVF medium, E. coli is frequently observed. Despite this, it is possible to give birth to a healthy baby [17]. Li et al proposed zona pellucida removal as a safe method to save embryos infected during culture. They reported that this allowed for pregnancy without further intrauterine infections [18]. Ricci et al observed that the presence of E. coli was significantly associated with reduced levels of Lactobacilli. Moreover, these bacteria were more common in patients with failed IVF, but this was not statistically significant [19]. On the other hand, the presence of E. coli alongside Staphylococcus spp. may promote lower fertilization and pregnancy rates [20,21]. In addition, Moretti et al indicated that E. coli can reduce sperm motility and induce apoptosis [22].

Lactobacilli make up the largest percentage of the female reproductive system microbiome, and their reduction in favor of anaerobic bacteria leads to dysbiosis, which is associated with an increased risk of disease and poor fertilization and pregnancy outcomes. Interestingly, individual groups of Lactobacilli show different degrees of protection against infection [23,24]. L. helveticus is considered a very beneficial bacterium that can inhibit the growth of pathogens. This is possible thanks to the production of lactic acid and hydrogen peroxide, as well as stimulating the host’s immune system [25]. Johnson-Henry et al showed that L. helveticus can inhibit adhesion of E. coli to the epithelium, which may be helpful in the treatment of intestinal pathogen infection [26], and Atassi et al reached a similar conclusion for uropathogenic E. coli and vaginosis-associated bacteria G. vaginalis [27]. Interestingly, in our study there was a negative correlation between G. vaginalis and L. helveticus in swabs obtained during the first collection. Wee et al reported that G. vaginalis is more common in the cervix of infertile women [28]. Moreno et al observed that the presence of G. vaginalis in the receptive endometrium was associated with reduced rates of implantation, pregnancy, and live births [29]gut, respiratory, skin, and vaginal microbiomes. In their other study, they also confirmed that the dysbiotic endometrial microbiome containing Gardnerella, Atopobium, Bifidobacterium, Chryseobacterium, Haemophilus, Klebsiella, Neisseria, and Staphylococcus was associated with unsuccessful reproductive outcomes. Importantly, Lactobacillus was dominant in patients with live birth outcomes [30]. Haahr et al suggested that high levels of G. vaginalis negatively affect pregnancy rates in IVF patients [9]. Koedooder et al noted that a higher proportion of G. vaginalis with low levels of Lactobacillus spp. correlates with worse IVF outcomes [31]. Bernabeu et al also confirmed that Lactobacillus spp. predominate in pregnant women [32], which is consistent with our results. Carmen Diaz-Martinez et al drew similar conclusions, and they also reported significant differences in the endometrial microbiome between women with and without a history of repeated implantation failures [33]. Toson et al emphasized the need for comparability among studies on the uterine microbiome, made possible by standardization of protocols and larger groups of patients. Analyzing previous research, they proposed that a healthy endometrial microbiome allows for embryo implantation and a live birth, despite the minimal group of pathogenic bacteria [34]. In our study, we noted that the normal microbiome was dominant regardless of sampling. Importantly, in the samples after embryo implantation, apart from a large population of Lactobacillus, G. vaginalis was also present. Interestingly, in the study by Reschini et al, Lactobacillus was dominant in the endometrium in only 8% of cases. Biodiversity was also greater in pregnant women. In addition, they proposed using embryo transfer catheters to assess the endometrial microbiome [35].

The limitation of our study is the relatively small study group, which makes it difficult to analyze the results in more detail with so many species and strains of bacteria. It would be beneficial to expand the study group to be able to draw stronger conclusions. It would also be interesting to see if the cervical or endometrial bacterial microbiome is similar to the oral microbiome in order to simplify and popularize the use of microbiome changes in the context of infertility diagnosis. Another limitation is the single-center nature of our study.

Conclusions

Lactobacillus spp. were the main bacteria identified in the cervix and endometrium, present before, during, and after successful embryo transfer. E. coli and Gardnerella vaginalis reduced the protective effect of Lactobacilli before, during, and after embryo implantation.