Analysis of vaginal and endometrial microbiota communities in infertile women with a history of repeated implantation failure

Abstract Purpose To identify specific bacterial communities in vaginal and endometrial microbiotas as biomarkers of implantation failure by comprehensively analyzing their microbiotas using next‐generation sequencing. Methods We investigated α‐ and β‐diversities of vaginal and endometrial microbiotas using 16S rRNA gene sequencing and compared their profiles between 145 women with repeated implantation failure (RIF) and 21 controls who lacked the factors responsible for implantation failure with a high probability of being healthy and fertile to identify specific bacteria that induce implantation failure. Results The endometrial microbiotas had higher α‐diversities than did the vaginal microbiotas (P < .001). The microbiota profiles showed that vaginal and endometrial samples in RIF patients had significantly higher levels of 5 and 14 bacterial genera, respectively, than those in controls. Vaginal Lactobacillus rates in RIF patients were significantly lower at 76.4 ± 38.9% compared with those of the controls at 91.8 ± 22.7% (P = .018), but endometrial Lactobacillus rates did not significantly differ between the RIF patients and controls (56.2 ± 36.4% and 58.8 ± 37.0%, respectively, P = .79). Conclusions Impaired microbiota communities containing specific bacteria in both the endometrium and vagina were associated with implantation failure. The vaginal Lactobacillus rates, but not the endometrial, may be a biomarker for RIF.


| INTRODUC TI ON
In vitro fertilization (IVF) technology and quality have rapidly advanced. Recent reports of preimplantation genetic testing for aneuploidy reported >60% clinical pregnancy rates after embryo transfer (ET) cycles. 1,2 However, pregnancy requires competent embryos and a receptive endometrium; therefore, repeated implantation failure (RIF) with euploid embryos is difficult to treat. 3 Reproductive-related microbiota communities in women can affect reproductive and obstetric outcomes. 4 Bacterial vaginosis (BV) is associated with obstetric complications, including pregnancy loss and preterm birth. [5][6][7][8][9][10] Analyzing microbiota profiles of the amniotic fluid may help predict perinatal outcomes. 11 Moreno et al 12 revealed vaginal and endometrial bacterial communities from vaginal aspirates and endometrial fluid from fertile and infertile women who underwent IVF. Data from the endometrial samples showed that bacterial communities from women experiencing implantation failure or pregnancy loss after ET contained more Gardnerella and Streptococcus and fewer Lactobacillus than did those from women who had successful livebirths. Furthermore, infertile patients with >90% Lactobacillus in their endometrial microbiota (EM) had significantly good pregnancy prognoses after IVF than did those with <90%. Therefore, endometrial microbiome analyses are used to determine individual EM profiles in infertile women. 13,14 The importance of the abundance of Lactobacillus in the endometrium is currently being debated. If lower Lactobacillus rates in the EM are associated with lower implantation rates, patients with repeated implantation failure would be expected to have abnormal incidence rates of Lactobacillus and high rates of pathogenic bacteria. In addition, we expected similar results for the vaginal microbiota (VM), because the vagina prevents the invasion of pathogens into the uterus.
We compared the EM and VM communities between patients with RIF and healthy women at the genus level using next-generation sequencing and analyzed the abundance of Lactobacillus and the presence of specific bacteria responsible for RIF.

| Patients
We diagnosed patients with RIF if they failed to achieve clinical pregnancy after at least three ET cycles, using the Gardner scoring system 15  Sixty-six patients with RIF were excluded for various reasons, including the presence of obvious risk factors for RIF. Patients with uterine cavity ultrasounds that revealed possible causes of infertility received a diagnostic hysteroscopy to rule out intrauterine disorders (eg, endometrial polyps, submucosal myomas, and intrauterine adhesions). Twenty-nine women were excluded after ultrasound and hysteroscopy examinations. We also excluded 30 women with other possible risk factors for reproductive failure, including 13 with thrombophilia (eg, antiphospholipid syndrome), 15 with endocrinologic abnormalities or collagen disease, and 2 with parental chromosomal imbalances or translocations. Seven women who had received antibiotics within 1 month of sampling were excluded because antibiotics can affect microbiota communities.
Finally, 145 women were included. Forty did not provide vaginal specimens; thus, we obtained 145 endometrial samples and 105 vaginal aspirates. We also recruited 21 healthy women with etiologies of infertility because their husbands had azoospermia as the control group; these women had regular menstrual cycles without causes of infertility such as tubal factors, ovulation disorder, endometriosis, endocrinologic abnormalities, or immunological abnormalities. Figure 1 shows the participant selection methods.

| Vaginal and endometrial sampling
Both vaginal and endometrial samples were taken 5-7 days after ovulation or the beginning of the high-temperature phase in the basal body temperature. All the specimens were collected in a hormonefree cycle, except in the four patients with irregular menstruation.
There were four women with irregular menstrual cycles in RIF group, and those samples were taken during the hormone replacement cycle. From days 1-3 of the menstrual cycle, 2-8 mg of oral estradiol valerate (Progynova ® , Bayer Health Care, Schering, Germany) was administered. From day 13, oral chlormadinone acetate (8 mg; Lutoral, Shionogi, Osaka, Japan) was administered for 13 days.
Samples were obtained 5-7 days after initiating oral progesterone intake.
Vaginal discharge was first collected in the posterior fornix of the vagina using two sterilized swabs, after placing a sterilized vaginal speculum. One swab was submitted for Nugent scoring, 16  Participants for whom pipette insertion was difficult due to strong uterine flexion or other reasons were excluded from this study.  Sequenced reads were merged using EA-Utils fastq-join 18 to obtain a 291-bp median merged sequence length. Quality control of the merged sequence was performed using USEARCH v10.0.240 19 to remove PhiX reads, truncate primer-binding sequences, and discard sequences of <100 bp and with a sequence quality <Q20. Quantitative Insights Into Microbial Ecology (QIIME) 1.9.1 was used with the default parameters for quality filtering, chimera checking, sequence clustering into operational taxonomic units (OTUs), and taxonomic assignment. 20 Sequences were clustered into OTUs using an open-reference OTUpicking strategy using the UCLUST method based on 97% sequence identity. Taxonomy was assigned to each OTU using RDP Classifier 21 with a 0.50 confidence threshold against the Greengenes database, version 13_8. 22 Taxonomy was determined at the genus level.

| Sequencing results and operational taxonomic unit analysis
Seven endometrial samples with insufficient sequence reads were ex- were filtered from the OTU tables ( Figure 1). Bacterial taxa in a blank control were assumed to be contaminants from various reagents; therefore, blank-characteristic taxa were subtracted to reduce background noise as in previous studies. 11,23 Fourteen bacterial taxa detected in a blank control and known to be reagent contaminants were excluded using QIIME: Acidovorax, Acinetobacter, Chryseobacterium, Pseudomonas, Rhodococcus, Sphingomonas, Stenotrophomonas, and Yersinia (Table S1). Nineteen endometrial samples were excluded from the analysis because reads assigned to background bacteria accounted for >95% of all reads, and <5% of the reads remained after filtering.

| Statistical analysis
We calculated the Shannon diversity index and Chao1 richness, which became the index of the microbiota's α-diversity, then conducted t tests. We calculated the weighted UniFrac distance for analyzing the β-diversity of the microbiotas between the samples and conducted PERMANOVA tests. The tests were analyzed using QIIME

| Endometrial and vaginal microbiota bacterial diversities
Shannon diversity and Chao1 richness indexes as α-diversity metrics were calculated to compare the patients' vaginal and endometrial bacterial compositions (Table 2, Figure 2A) Figure 2B, Figure 1C). β-diversity was analyzed to compare compositional dissimilarities between the EMs and VMs. Principal coordinate analysis (PCoA), the multivariate analysis based on weighted UniFrac distance to compare microbiome differences between groups, revealed significant associations between microbiotas (P = .001) ( Figure S1). The few subjects had higher uterine Lactobacillus rates than vaginal Lactobacillus rates. Most subjects with lower vaginal Lactobacillus rates also had lower endometrial Lactobacillus rates; thus, individuals with vaginal dysbiosis also had uterine dysbiosis (Figure 3).

| Bacterial community differences between the RIF and control groups
To identify the relationship between bacterial diversity and implantation failure, we compared the microbiota data from the endometrial and vaginal samples between the RIF and control groups (Table S4). The α-and β-diversities did not significantly differ in the EMs or VMs between the RIF and control groups ( Figure 4A,B). We further investigated differences between the bacterial genera in these groups. Twenty-five and 131 bacterial species were detected from the vaginal and endometrial samples, respectively. Figure 5 shows To identify candidate bacterial genera as risk factors for RIF, bacterial abundances in the EMs and VMs were evaluated ( Nugent score, mean ± SD 1.9 ± 2.7 0.9 ± 1.6 .09 ≥7 (Bacterial vaginosis), n (%) 16 (11.9) 0 (0) .13 Note: Nugent score is often used for the diagnosis of bacterial vaginosis. Bacterial vaginosis is diagnosed as the score 7-10.
Abbreviations: ET, embryo transfer; RIF, repeated implantation failure; SD, standard deviation. Hydrogenophaga) were significantly higher than those in the controls (

| D ISCUSS I ON
Although the uterus was once hypothesized to be sterile via the cervical mucus, 24,25 intrauterine bacterial microbiotas have since been confirmed. 12,26,27 In the endometrial cells of fertile women, progesterone secretion from the luteum body induces subnuclear vacuole production, leading to increased glycogen levels. 28 Deposition of endometrial epithelial glycogen may allow bacteria to colonize the endometrium. 28 We found that the EMs had higher α-diversity than did the VMs, as previously reported. 29 Lactic acid produced by Lactobacillus acidifies the vagina, thus inhibiting the growth of other bacterial species 30,31 ; however, the number of bacteria in the uterus is extremely small at 1/100-1/10 000 that of the vagina, and some bacterial species dominate among the highly varied vaginal bacteria, leading to low bacterial diversity in the vagina. 32,33 Therefore, the EM community is mostly independent of the VM community.
A healthy microbiota generated by a healthy lifestyle is defined as "eubiosis," and disruption of this balance inclines toward a state of "dysbiosis," in which pathogenic bacteria predominate over endogenous bacteria due to an inappropriate immune response, inflammation, or suppressed immune response. 34 The cervical mucus plug is partially impermeable to bacterial ascension from  There is a report that shows to control the menstrual cycle using the hormone has some effects on the microbiota. Therefore, the research may have been influenced by the presence of the hormone. 45 Fourth, it was difficult to prove that a sterile organ, such as the endometrium, was completely free of contamination. The amount and nature of the cervical mucus change with the phase of menstruation, and the amount of cervical mucus increases and is highly glutinous in the growth phase, making it difficult to completely remove the samples, even if they are washed or wiped with a tampon. We therefore collected the samples in the secretory phases, being careful to prevent contamination. If we detected cervical mucus on the tip of the pipette immediately after collecting the endometrial sample, even after careful washing, we cut off the tip of the pipette; however, the usefulness of this procedure has not been proven.

ACK N OWLED G EM ENTS
The authors thank the women who participated in the study and