ABSTRACT
Background
Anti‐β2GPI/HLA‐DR autoantibodies may be involved in recurrent implantation failure (RIF) and recurrent pregnancy loss (RPL). We examined their association with the endometrial microbiome and chronic endometritis (CE).
Methods
In this cross‐sectional study, 141 women (54 RIF, 87 RPL) were enrolled. Serum anti‐β2GPI/HLA‐DR positivity was defined using 99th/95th percentile cut‐offs. Endometrial microbiome was assessed by 16S rRNA sequencing, focusing on reproductive‐failure–related species (Gardnerella, Prevotella, Atopobium, Dialister, Anaerococcus, Ureaplasma, Mycoplasma). Microbiome and CE were compared by antibody status in RIF and RPL.
Results
In RIF, antibody‐positive women more frequently had Lactobacillus iners (99‰ and 95‰ cut‐offs: 71.4% vs. 23.4%, p = 0.03; and 70.0% vs. 20.5%, p = 0.026) and reproductive‐failure–related bacterial species (99‰: 100% vs. 51.1%, p = 0.016; 95‰: 90.0% vs. 52.3%, p = 0.032). In multivariable analysis with 95‰ cut‐off, Lactobacillus iners (OR 13.1, p = 0.003) and reproductive‐failure–related species (OR 9.64, p = 0.029) were independently associated with antibody positivity. In RPL, Anaerococcus was more frequent in antibody‐positive women. CE frequency did not differ by antibody status in RIF or RPL.
Conclusion
Anti‐β2GPI/HLA‐DR antibody positivity was associated with endometrial dysbiosis and may serve as a biomarker of abnormal intrauterine environment in reproductive failure.
Keywords: endometrial microbiome, HLA‐DR, recurrent implantation failure, recurrent pregnancy loss, β2‐glycoprotein I
1. Introduction
Successful embryo implantation and pregnancy maintenance require a healthy intrauterine environment, and disruption of this milieu has been implicated in the pathogenesis of recurrent implantation failure (RIF) and recurrent pregnancy loss (RPL) [1, 2]. Factors such as hormonal imbalance, coagulation disorders, uterine structural abnormalities, and impaired endometrial receptivity have been considered key determinants of the intrauterine environment [1, 3, 4, 5, 6]. More recently, however, the uterine endometrial microbiome and chronic endometritis (CE) have emerged as important contributors to reproductive failure in both RIF and RPL [7, 8, 9, 10, 11, 12, 13, 14]. In RIF, CE has been associated with a breakdown of the Lactobacillus‐dominant microbiome, and an increased prevalence of anaerobic bacteria, including Prevotella, Gardnerella, and Atopobium, indicates a strong link between CE and endometrial dysbiosis [7, 8]. Moreover, antibiotic treatment for CE has been reported to improve assisted reproductive technology (ART) outcomes [9, 10].
Similarly, in RPL, the endometrial microbiome of CE‐positive patients has a reduced proportion of Lactobacillus species and an increased prevalence of anaerobic or bacterial vaginosis (BV)‐associated bacteria [11]. Shi et al. also demonstrated that endometrial dysbiosis in RPL is an independent risk factor for miscarriage and preterm birth in subsequent pregnancies [12]. Additionally, several studies have suggested that antibiotic treatment of CE improves pregnancy rates and live birth rates in RPL [13, 14].
Although endometrial dysbiosis and CE are considered potential contributors to impaired intrauterine conditions and adverse reproductive outcomes, most previous reports are cross‐sectional studies. Thus, the causal relationships among uterine endometrial dysbiosis, CE, and adverse reproductive outcomes, and their differential roles in RIF and RPL remain unclear.
Antiphospholipid antibodies (aPLs) are recognized as one of the causal factors of RPL [15], but their causal relationship with RIF has not been demonstrated; however, several studies have reported a higher prevalence of certain autoantibodies in women with RIF [16, 17]. Recently, Tanimura et al. identified anti‐β2‐glycoprotein I/human leukocyte antigen‐DR (anti‐β2GPI/HLA‐DR) antibodies, one of the so‐called neoself antibodies, in patients with antiphospholipid syndrome [18, 19]. Subsequently, women with anti‐β2GPI/HLA‐DR antibodies were shown to have an increased risk not only of RPL but also of fetal growth restriction and hypertensive disorders of pregnancy [20]. Furthermore, it was demonstrated that, among RPL patients positive for this antibody, treatment with low‐dose aspirin and/or heparin increased the live birth rate [21]. Ono et al. reported that anti‐β2GPI/HLA‐DR antibodies are present in 15.5% of infertile women undergoing ART and in 28.7% of women with RIF [22], and that antibody positivity in infertile women is associated with a trend toward lower implantation rates and expression of the β2GPI/HLA‐DR antigen complex in their endometrium [23]. Together, these findings suggest that anti‐β2GPI/HLA‐DR antibodies may influence implantation and early placentation. Because anti‐β2GPI/HLA‐DR antibodies are induced by infection or inflammation, we hypothesized that dysbiosis or CE‐related local inflammation in the endometrium may promote the expression of β2GPI/HLA‐DR complexes and thereby contribute to antibody production. Therefore, in this prospective study, we investigated the relationships among the endometrial microbiome, CE, and anti‐β2GPI/HLA‐DR antibodies in women with RIF and RPL.
2. Material and Methods
2.1. Study Design and Participants
This cross‐sectional observational study was conducted between March 2021 and August 2025 at two fertility centers: Teine Keijinkai Hospital and the University of Yamanashi Hospital. The study protocol was approved by the institutional review boards of both institutions (Teine Keijinkai Hospital: Approval #020387‐00, University of Yamanashi: Approval #2‐023209‐00), and written informed consent was obtained from all participants. Eligible participants were women diagnosed with RIF or RPL who underwent endometrial biopsy for assessment of CE by CD138 immunohistochemistry and endometrial microbiome analysis (16S r RNA sequencing) within the same menstrual cycle, together with measurement of serum anti‐β2GPI/HLA‐DR antibodies. A total of 141 women were included in the analysis, with 54 meeting the criteria for RIF and 87 for RPL. Of these, 13 fulfilled the diagnostic criteria for both conditions (Figure 1). The uterine endometrial microbiome and CE were evaluated with or without the presence of anti‐β2GPI/HLA‐DR antibodies in the overall cohort as well as separately in the RIF and RPL subgroups. Exclusion criteria included malignancy, congenital uterine anomalies, intrauterine surgery within the past 6 months, hormonal treatment within the past 3 months, and antibiotic use or ingestion of probiotics or prebiotics within two menstrual cycles before the biopsy. RIF was defined as failure to achieve a clinical pregnancy after ≥ 3 embryo transfer cycles. RPL was defined as ≥ 2 pregnancy losses before 22 weeks of gestation.
FIGURE 1.
Flow diagram of the 141 participants in the analysis.
2.2. Endometrial Biopsy and Sample Processing
Endometrial samples were obtained by Pipet Curet aspiration during the mid‐luteal phase (approximately 7–10 days after ovulation), as determined by the last menstrual period and transvaginal ultrasound. After cleansing the vaginal wall and perineum with 0.025% benzalkonium chloride, the aspirated endometrial tissue was divided. Because Pipelle aspiration yields fragmented tissue without identifiable anatomical orientation within the uterine cavity, site‐specific selection of tissue was not possible. Therefore, no intentional selection of a specific tissue portion was performed, and the obtained tissue fragments were directly transferred into the OMNIgene‐VAGINAL collection tube according to the standardized collection protocol. Microbiome analyzes were subsequently performed at a certified commercial laboratory using standardized processing procedures. A portion of the obtained tissue was transferred into an OMNIgene‐VAGINAL kit (DNA Genotek Inc., Ottawa, Canada) containing a DNA/RNA stabilizer for microbiome analysis, and the remaining tissue was fixed in 10% neutral‐buffered formalin for histopathological evaluation, including CD138 immunohistochemistry. Microbiome analyzes were performed at a certified commercial laboratory (Varinos Inc., Tokyo, Japan), as described previously [24].
2.3. Assessment of Chronic Endometritis
Histopathological evaluation was performed by an external diagnostic laboratory. CD138‐positive plasma cells were quantified using immunohistochemistry. Following Liu's criteria, CE was diagnosed when > 5.15 plasma cells per 10 mm2 were observed [25]. CE was treated as a binary variable (present/absent).
2.4. Endometrial Microbiome Analysis
The endometrial microbiome was profiled by sequencing the V4 region of the bacterial 16S rRNA gene. Following DNA extraction, the target region was amplified by polymerase chain reaction and libraries were generated using the Illumina 16S Metagenomic Sequencing Library Preparation workflow. Sequencing data were processed with a standardized bioinformatics pipeline to derive taxonomic determination and relative abundance profiles. To comprehensively evaluate microbiota imbalance, rather than relying on individual taxa alone, a composite outcome was defined. Specifically, the detection of any microorganisms previously associated with reproductive or perinatal disorders, including CE, BV, infertility, implantation failure, miscarriage, or preterm birth, was defined as “reproductive‐failure–related bacterial species positive.” These taxa included Gardnerella, Prevotella, Atopobium, Dialister, Anaerococcus, Ureaplasma, and Mycoplasma. Since Lactobacillus species are generally indicative of a healthy uterine and vaginal environment, they were evaluated separately on an individual basis rather than included in this category. A taxon was considered “present” when its relative abundance exceeded 0%, and it was analyzed as a binary variable.
2.5. Measurement of Anti‐β2GPI/HLA‐DR Antibodies
Serum levels of anti‐β2GPI/HLA‐DR antibodies were measured at AOI Biosciences Co. Ltd., Tokyo, Japan following previously described standardized methods. Specifically, anti‐β2GPI/HLA‐DR antibodies were measured using a standardized flow cytometry–based assay as previously described by Tanimura et al. [21]. In brief, cells expressing β2GPI/HLA‐DR complexes and control cells expressing HLA‐DR alone were incubated with diluted patient sera. Bound IgG was detected using fluorochrome‐conjugated anti‐human IgG antibodies and analyzed by flow cytometry. Specific binding was calculated by subtracting IgG binding to control cells from that to β2GPI/HLA‐DR–expressing cells, and antibody levels were quantified using standardized reference sera under quality‐controlled conditions, and all measurements were performed in duplicate, and the mean value was defined as the anti‐β2GPI/HLA‐DR antibody level of the sample. The cutoff value of 73.3 U was based on the normal range reported by Tanimura et al. [21], in which sera from 374 healthy controls (300 women and 74 men) were analyzed and the upper limit of normal (< 73.3 U) was defined as the 99th percentile of the antibody distribution after outlier removal. In addition, 61.9 U, corresponding to the 95th percentile within the same control, was also used to stratify antibody levels. In addition, the association between antiphospholipid antibodies (aPL), the uterine microbiota, and chronic endometritis (CE) was examined. aPL positivity was defined when at least one of the following antibodies of antiphospholipid syndrome criteria was detected: lupus anticoagulant (LA), anti‐cardiolipin antibody (aCL; IgG and/or IgM), or anti‐β2‐glycoprotein I antibody (anti‐β2GPI; IgG and/or IgM).
2.6. Statistical Analysis
Comparisons between antibody‐positive and antibody‐negative groups were conducted separately for patients with RIF and RPL, as well as in the combined cohort. Variables included age, body mass index (BMI), CE status, CD138‐positive cell counts, relative abundance of Lactobacillus species, and detection of individual taxa and composite microbiome abnormalities. Continuous variables were expressed as medians (minimum–maximum) and compared using the Mann–Whitney U test. Categorical variables were expressed as frequencies (%) and compared using Fisher's exact test or the chi‐square test. Multivariate logistic regression analysis was performed with positivity of anti‐β2GPI/HLA‐DR antibodies as the dependent variable and age, CE, the presence of Lactobacillus iners and reproductive‐failure–related microbiome abnormalities as independent variables. This analysis was restricted to the 95% cut‐off, because the number of antibody‐positive patients at the 99% was too small to permit reliable parameter estimation. Analyzes were conducted for patients with RIF only. Odds ratios (ORs) and 95% confidence intervals (CI) were calculated, and two‐tailed p‐values < 0.05 were considered statistically significant. Statistical analyzes were performed using JMP software, version 16 (SAS Institute Inc., Cary, NC, USA).
3. Results
3.1. Overall Analysis
As shown in Table 1, there were no significant differences between the antibody positive and negative groups for age, BMI, number of previous miscarriages, history of live births, total number of detected uterine bacterial species, the prevalence of CE, CD138‐positive plasma cell counts, the relative abundance of Lactobacillus species, or the frequency of a Lactobacillus‐dominant microbiome (> 90% relative abundance), irrespective of the cut‐off used. In contrast, analysis of individual Lactobacillus species revealed that the presence of Lactobacillus iners was significantly greater in the antibody‐positive group at the 95% cut‐off analysis (62.5% [15/24] vs. 33.7% [35/104], p = 0.017). The relative abundance of Lactobacillus gasseri tended to be higher in the antibody‐positive group in the 99% analysis (0 [0–96.2] vs. 0 [0–99.7], p = 0.062). Anaerococcus species were significantly more frequent in the antibody‐positive group using both cut‐offs (99%, 25.0% [5/20] vs. 5.6% [6/108], p = 0.016; 95%, 25.0% [6/24] vs. 4.8% [5/104], p = 0.005). Furthermore, the presence of Ureaplasma species tended to be higher in antibody‐positive patients in the 95% analysis (29.2% [7/24] vs. 12.5% [13/124], p = 0.086), and the frequency of species associated with reproductive failure tended to be higher as well (75.0% [18/24] vs. 52.9% [55/104], p = 0.081). The proportion of Lactobacillus iners within total Lactobacillus did not differ significantly between antibody‐positive and antibody‐negative groups at either cut‐off.
TABLE 1.
Characteristics of RIF and RPL patients grouped by anti‐β2GPI/HLA‐DR antibody positivity at the 99th and 95th percentile cut‐offs.
| 99‰ (> 73.3 U) | p | 95‰ (> 61.9 U) | p | |||
|---|---|---|---|---|---|---|
| Positive for the antibody (n = 20) | Negative for the antibody (n = 108) | Positive for the antibody (n = 24) | Negative for the antibody (n = 104) | |||
| Age (years) | 39 (32–45) | 36.0 (25–44) | 0.289 | 38 (24–45) | 36 (25–44) | 0.17 |
| Body mass Index (kg/m2) | 21.2 (18.0–37.2) | 21.8 (16.2–35.9) | 0.603 | 22.2 (18.0–37.2) | 21.4 (16.2–35.9) | 0.168 |
| The number of previous miscarriages | 2 (0–3) | 2 (0–12) | 0.894 | 2 (0–3) | 2 (0–12) | 0.312 |
| The history of live births | 7 (35.0%) | 38 (35.2%) | 0.811 | 9 (37.5%) | 36 (34.6%) | 0.976 |
| The unit of anti‐β2GPI/HLA‐DR antibody (U) | 133.0 (75.5–162800) | 21.1 (0–69.5) | < 0.001 | 108.7 (65.1–162800) | 20.6 (0–60.7) | < 0.001 |
| Number of bacterial species | 5.5 (1–15) | 4 (1–17) | 0.35 | 5.5 (2–12) | 4 (1–17) | 0.245 |
| The prevlance of chronic endometritis (Liu's method) | 6 (30.0%) | 22 (20.4%) | 0.508 | 6 (25.0%) | 22 (21.2%) | 0.891 |
| Plasma cell count/10 mm2 with CD138 staining | 2.0 (0–84.3) | 0.68 (0–252.6) | 0.163 | 2.0 (0–84.3) | 0.67 (0–252.6) | 0.296 |
| Relative abundance of Lactobacilus species (%) | 95.5 (0–99.4) | 98.5 (0–100) | 0.73 | 95.5 (0–100) | 98.5 (0–100) | 0.797 |
| Lactobacilus‐ dominant microbiome (> 90%) | 11 (55.5%) | 61 (56.5%) | 0.903 | 14 (58.3%) | 60 (57.7%) | 0.863 |
| Presence of Lactobacilus crispatus | 10 (50.0%) | 51 (47.2%) | 0.988 | 12 (50.0%) | 49 (47.1%) | 0.799 |
| Presence of Lactobacilus gasseri | 6 (30.0%) | 22 (20.4%) | 0.508 | 8 (33.3%) | 20 (19.2%) | 0.218 |
| Presence of Lactobacilus iners | 11 (55.5%) | 39 (36.1%) | 0.18 | 15 (62.5%) | 35 (33.7%) | 0.017 |
| Presence of Lactobacilus jenseni | 2 (10.0%) | 18 (16.7%) | 0.675 | 3 (12.5%) | 17 (16.3%) | 0.876 |
| Proportion of Lactobacillus iners within total Lactobacillus (%) | 0.10 (0–100) | 0 (0–100) | 0.798 | 0.10 (0–100) | 0 (0–100) | 0.466 |
| Relative abundance of Lactobacilus crispatus (%) | 0.1 (0–99.9) | 0 (0–100) | 0.979 | 0.1 (0–100) | 0 (0–100) | 0.817 |
| Relative abundance of Lactobacilus gasseri (%) | 0 (0–96.2) | 0 (0–99.7) | 0.062 | 0 (0–96.2) | 0 (0–99.7) | 0.225 |
| Relative abundance of Lactobacilus iners (%) | 0.1 (0–98.8) | 0 (0–100) | 0.318 | 0.1 (0–99.4) | 0 (0–100) | 0.116 |
| Relative abundance of Lactobacilus jenseni (%) | 0 (0–81.3) | 0 (0–89.3) | 0.616 | 0 (0–81.3) | 0 (0–89.3) | 0.368 |
| Relative abundance of Lactobacillus species excluding iners (%) | 57.3 (0–99.9) | 20.2 (0–100) | 0.896 | 42.8 (0–100) | 31.7 (0–100) | 0.94 |
| Presence of Bifidobacterium species | 5 (25.0%) | 11 (10.2%) | 0.141 | 5 (20.8%) | 11 (10.6%) | 0.304 |
| Presence of Atopobium species | 5 (25.0%) | 22 (20.4%) | 0.867 | 7 (29.2%) | 23 (22.1%) | 0.64 |
| Presence of Prevotella species | 9 (45.0%) | 30 (27.8%) | 0.203 | 11 (45.8%) | 28 (26.9%) | 0.117 |
| Presence of Gardnerella species | 7 (35.0%) | 34 (31.5%) | 0.961 | 8 (33.3%) | 33 (31.7%) | 0.927 |
| Presence of Anaerococcus species | 5 (25.0%) | 6 (5.6%) | 0.016 | 6 (25.0%) | 5 (4.8%) | 0.005 |
| Presence of Dialister species | 7 (35.0%) | 19 (17.6%) | 0.14 | 8 (33.3%) | 18 (17.3%) | 0.14 |
| Presence of Ureaplasma species | 6 (30.0%) | 14 (13.0%) | 0.111 | 7 (29.2%) | 13 (12.5%) | 0.086 |
| Presence of Mycoplasma species | 1 (5.0%) | 1 (1.0%) | 0.713 | 1 (4.2%) | 1 (1.0%) | 0.819 |
| Presence of Escherichia species | 0 (0%) | 4 (3.7%) | 0.861 | 0 (0%) | 4 (3.8%) | 0.744 |
| Presence of Streptococcus species | 6 (30.0%) | 19 (17.6%) | 0.328 | 7 (29.2%) | 18 (17.3%) | 0.301 |
| Reproductive failure related species | 15 (75.0%) | 58 (53.7%) | 0.128 | 18 (75.0%) | 55 (52.9%) | 0.081 |
Note: Data are shown median (range) or n (%).
Abbreviations: RIF, recurrent implantation failure; RPL, recurrent pregnancy loss.
3.2. Analysis in Patients With RIF
In patients with RIF, no significant differences were observed in age, BMI, and numbers of bacterial species between antibody‐positive and antibody‐negative patients in both cut‐off analyzes. A detailed analysis of patients with RIF is shown in Table 2. The relative abundance of Lactobacillus species and the frequency of a Lactobacillus‐dominant microbiome (> 90%) did not differ between antibody‐positive and antibody‐negative groups at both percentile cut‐offs. However, the presence of Lactobacillus iners was significantly higher in the antibody‐positive group at both cut‐offs (99%, 71.4% [5/7] vs. 23.4% [11/47], p = 0.019; 95%, 70.0% [7/10] vs. 20.5% [9/44], p = 0.004). Similarly, its relative abundance was higher in the antibody‐positive group (99%, 0.1% [0–92.2] vs. 0% [0–100], p = 0.03; 95%, 0.1% [0–99.4] vs. 0% [0–100], p = 0.026). Furthermore, analysis of the proportion of Lactobacillus iners within total Lactobacillus demonstrated a higher proportion in antibody‐positive patients, reaching statistical significance at the 95% percentile cut‐off analysis (0.25% [0–100] vs. 0% [0–100], p = 0.016), while showing a similar trend at the 99% percentile cut‐off (0.25% [0–100] vs. 0% [0–100], p = 0.076). Prevotella species were significantly more frequent in the antibody‐positive group in the 99% analysis (71.4% [5/7] vs. 27.7% [13/47], p = 0.034), and the frequency of Anaerococcus species was significantly higher in the 95% analysis (30.0% [3/10] vs. 4.5% [2/44], p = 0.037). Ureaplasma species showed a trend toward a higher frequency in the antibody‐positive group in the 99% analysis (28.6% [2/7] vs. 4.3% [2/47], p = 0.077) and were significantly more frequent in the 95% percentile analysis (30.0% [3/10] vs. 2.3% [1/44], p = 0.016). No differences were observed between groups for Atopobium, Gardnerella, or Dialister species. However, the prevalence of reproductive failure–related bacterial species was significantly higher in patients who were antibody‐positive at both cut‐offs (99%, 100% [7/7] vs. 51.1% [24/47], p = 0.016; 95%, 90.0% [9/10] vs. 52.3% [22/47], p = 0.032). The prevalence of CE did not differ by antibody status (99%, 14.3% [1/7] vs. 12.8% [6/47]; 95%, 10.0% [1/10] vs. 13.6% [6/44]). These findings indicate that in patients with RIF, anti‐β2GPI/HLA‐DR antibody positivity is associated with increased Lactobacius iners prevalence and a higher frequency of endometrial dysbiosis, particularly an increased presence of reproductive failure–related bacteria species.
TABLE 2.
Characteristics of RIF patients grouped by anti‐β2GPI/HLA‐DR antibody positivity at the 99th and 95th percentile cut‐offs.
| 99‰ (> 73.3 U) | p | 95‰ (> 61.9 U) | p | |||
|---|---|---|---|---|---|---|
| Positive for the antibody (n = 7) | Negative for the antibody (n = 47) | Positive for the antibody (n = 10) | Negative for the antibody (n = 44) | |||
| Age (years) | 40 (37–45) | 38 (25–44) | 0.09 | 39.5 (37–45) | 38 (25–44) | 0.374 |
| Body mass Index (kg/m2) | 20.7 (20.3–28.1) | 21.4 (16.2–35.9) | 0.35 | 25.9 (20.3–35.4) | 21.6 (16.2–35.9) | 0.311 |
| The unit of anti‐β2GPI/HLA‐DR antibody (U) | 215.9 (79.5–162800) | 21.1 (0–69.5) | < 0.001 | 116.1 (65.1–1285.9) | 20.6 (16.2–55.5) | < 0.001 |
| Number of bacterial species | 7.0 (2–11) | 4 (1–16) | 0.41 | 5.5 (2–12) | 4 (1–16) | 0.403 |
| The prevlance of chronic endometritis (Liu's method) | 1 (14.3%) | 6 (12.8%) | 1 | 1 (10%) | 6 (13.6%) | 1 |
| Plasma cell count/10 mm2 with CD138 staining | 0.73 (0–6.98) | 0.56 (0–58.3) | 0.97 | 0.53 (0–6.98) | 0.56 (0–58.3) | 0.708 |
| Relative abundance of Lactobacilus species (%) | 94.4 (0–99.4) | 97.0 (0–100) | 0.88 | 95.5 (0.1–99.9) | 92.3 (0–100) | 0.895 |
| Lactobacilus‐ dominant microbiome (> 90%) | 4 (57.1%) | 24 (51.1%) | 1 | 6 (60.0%) | 22 (50.0%) | 0.825 |
| Presence of Lactobacilus crispatus | 3 (42.9%) | 22 (46.8%) | 1 | 4 (40.0%) | 21 (47.7%) | 0.93 |
| Presence of Lactobacilus gasseri | 3 (42.9%) | 11 (23.4%) | 0.36 | 3 (30.0%) | 11 (25.0%) | 0.71 |
| Presence of Lactobacilus iners | 5 (71.4%) | 11 (23.4%) | 0.019 | 7 (70.0%) | 9 (20.5%) | 0.004 |
| Presence of Lactobacilus jenseni | 1 (14.3%) | 6 (12.8%) | 1 | 1 (10.0%) | 6 (13.6%) | 1 |
| Proportion of Lactobacillus iners within total Lactobacillus (%) | 0.25 (0–100) | 0 (0–100) | 0.076 | 0.25 (0–100) | 0 (0–100) | 0.016 |
| Relative abundance of Lactobacilus crispatus (%) | 0 (0–98.2) | 0 (0–100) | 0.48 | 0 (0–98.2) | 0 (0–100) | 0.363 |
| Relative abundance of Lactobacilus gasseri (%) | 0 (0–96.2) | 0 (0–99.7) | 0.24 | 0 (0–96.2) | 0 (0–99.7) | 1 |
| Relative abundance of Lactobacilus iners (%) | 0.1 (0–92.2) | 0 (0–100) | 0.03 | 0.1 (0–99.4) | 0 (0–100) | 0.026 |
| Relative abundance of Lactobacilus jenseni (%) | 0 (0–98.3) | 0 (0–89.3) | 0.64 | 0 (0–98.3) | 0 (0–89.3) | 0.657 |
| Relative abundance of Lactobacillus species excluding iners (%) | 86.7 (0.1–99.4) | 71.2 (0–100) | 0.94 | 44.5 (0–99.9) | 57.8 (0–100) | 0.802 |
| Presence of Bifidobacterium species | 2 (28.6%) | 6 (12.8%) | 0.275 | 2 (20.0%) | 6 (13.6%) | 0.632 |
| Presence of Atopobium species | 1 (14.3%) | 9 (19.1%) | 1 | 2 (20.0%) | 8 (18.2%) | 1 |
| Presence of Prevotella species | 5 (71.4%) | 13 (27.7%) | 0.034 | 6 (60.0%) | 12 (27.3%) | 0.107 |
| Presence of Gardnerella species | 3 (42.9%) | 12 (25.5%) | 0.382 | 4 (40.0%) | 11 (25.0%) | 0.572 |
| Presence of Anaerococcus species | 2 (28.6%) | 3 (6.4%) | 0.12 | 3 (30.0%) | 2 (4.5%) | 0.037 |
| Presence of Dialister species | 2 (28.6%) | 9 (19.1%) | 0.621 | 3 (30.0%) | 8 (18.2%) | 0.408 |
| Presence of Ureaplasma species | 2 (28.6%) | 2 (4.3%) | 0.077 | 3 (30.0%) | 1 (2.3%) | 0.016 |
| Presence of Mycoplasma species | 0 (0%) | 0 (0%) | NA | 0 (0%) | 0 (0%) | NA |
| Presence of Escherichia species | 0 (0%) | 2 (4.3%) | 1 | 0 (0%) | 2 (4.5%) | 1 |
| Presence of Streptococcus species | 2 (28.6%) | 14 (29.8%) | 1 | 3 (30.0%) | 13 (29.5%) | 1 |
| Reproductive failure related species | 7 (100%) | 24 (51.1%) | 0.016 | 9 (90%) | 22 (52.3%) | 0.032 |
Note: Data are shown median (range) or n (%).
Abbreviations: NA, not applicable; RIF, recurrent implantation failure.
Multivariable logistic regression analysis was performed in patients with RIF to evaluate whether reproductive‐failure‐related bacterial species were independently associated with antibody positivity. In 95% cut‐off analyzes, reproductive‐failure–related bacterial species and Lactobacillus iners were identified as independently associated factors, with odds ratios of 9.64 (95% CI, 1.23–216.0; p = 0.029) and 13.1 (95% CI, 2.32–122.0; p = 0.003), respectively (Table 3).
TABLE 3.
Multivariable logistic regression analysis of anti‐β2GPI/HLA‐DR antibody positivity (> 61.9 U; 95th percentile) in 56 patients with RIF.
| Variables | Odds ratio (95% confidence interval) | p |
|---|---|---|
| Age | 0.81 (0.56–1.10) | 0.352 |
| Reproductive failure related species | 9.64 (1.23–216.0) | 0.029 |
| Lactobacillus iners | 13.1 (2.32–122.0) | 0.003 |
| Chronic endometritis | 0.45 (0.22–3.49) | 0.473 |
3.3. Analysis in Patients With RPL
In patients with RPL, no significant differences were observed between antibody positive and antibody‐negative groups regarding the relative abundance of Lactobacillus species, individual Lactobacillus species, or CE prevalence at either the 99% or 95% cut‐off levels (99%, 31.3% [5/16] vs. 23.9% [17/71]; 95%, 29.4% [5/17] vs. 24.3% [17/70], Table 4). Similarly, the proportion of Lactobacillus iners within total Lactobacillus did not differ between antibody‐positive and antibody‐negative groups at either percentile cut‐off. CD138‐positive plasma cell counts did not differ between groups. Although only the presence of Anaerococcus species was higher in antibody‐positive patients at both cut‐offs (99%, 25.0% [4/16] vs. 4.2% [3/71], p = 0.020; 95%, 23.5% [4/17] vs. 4.3% [3/70], p = 0.034), unlike in patients with RIF, the frequency of reproductive‐failure–related bacterial species did not differ in either percentile cut‐off analysis (99%, 62.5% [10/16] vs. 57.7% [41/71]; 95%, 64.7% [11/17] vs. 51.1% [40/70], p = 0.95, Table 4).
TABLE 4.
Characteristics of RPL patients grouped by anti‐β2GPI/HLA‐DR antibody positivity at the 99th and 95th percentile cut‐offs.
| 99‰ (> 73.3 U) | p | 95‰ (> 61.9 U) | p | |||
|---|---|---|---|---|---|---|
| Positive for the antibody (n = 16) | Negative for the antibody (n = 71) | Positive for the antibody (n = 17) | Negative for the antibody (n = 70) | |||
| Age (years) | 37 (24–45) | 35 (25–44) | 0.228 | 37 (24–45) | 35 (25–44) | 0.139 |
| Body mass Index (kg/m2) | 21.7 (18.0–37.2) | 21.4 (16.2–35.9) | 0.347 | 22.2 (18.0–37.2) | 21.4 (16.2–35.9) | 0.22 |
| The number of previous miscarriages | 2 (2–3) | 2 (2–12) | 0.257 | 2 (2–3) | 2 (2–12) | 0.19 |
| The history of live births | 7 (43.8%) | 27 (38.0%) | 0.672 | 8 (47.1%) | 26 (37.1%) | 0.452 |
| The unit of anti‐β2GPI/HLA‐DR antibody (U) | 124.6 (75.5–1285.9) | 21.1 (0–69.5) | < 0.001 | 116.2 (75.5–1285.9) | 21.2 (16.2–35.9) | < 0.001 |
| Number of bacterial species | 3.0 (1–15) | 4 (1–17) | 0.801 | 4.0 (1–15) | 4 (1–17) | 0.661 |
| The prevlance of chronic endometritis (Liu's method) | 5 (31.3%) | 17 (23.9%) | 0.773 | 5 (29.4%) | 17 (24.3%) | 0.9 |
| Plasma cell count/10 mm2 with CD138 staining | 1.88 (0–84.3) | 1.32 (0–252.6) | 0.642 | 2.00 (0–84.3) | 15.3 (0–252.6) | 0.546 |
| Relative abundance of Lactobacilus species (%) | 98.8 (0–100) | 98.8 (0–100) | 0.881 | 92.8 (0–100) | 98.8 (0–100) | 0.732 |
| Lactobacilus‐ dominant microbiome (> 90%) | 9 (56.3%) | 43 (60.6%) | 0.97 | 9 (52.9%) | 43 (61.4%) | 0.72 |
| Presence of Lactobacilus crispatus (%) | 6 (37.5%) | 36 (50.7%) | 0.5 | 8 (47.1%) | 34 (48.6%) | 0.87 |
| Presence of Lactobacilus gasseri (%) | 5 (31.3%) | 15 (21.1%) | 0.59 | 6 (35.3%) | 14 (20.0%) | 0.31 |
| Presence of Lactobacilus iners (%) | 6 (37.5%) | 30 (42.3%) | 0.95 | 9 (52.9%) | 27 (38.6%) | 0.42 |
| Presence of Lactobacilus jenseni (%) | 2 (12.5%) | 13 (18.3%) | 0.73 | 3 (17.6%) | 12 (17.1%) | 1 |
| Relative abundance of Lactobacilus crispatus (%) | 0 (0–100) | 0.1 (0–100) | 0.605 | 0 (0–100) | 0.1 (0–100) | 0.858 |
| Relative abundance of Lactobacilus gasseri (%) | 0 (0–86.7) | 0 (0–80.8) | 0.189 | 0 (0–86.7) | 0 (0–80.8) | 0.259 |
| Relative abundance of Lactobacilus iners (%) | 0 (0–98.8) | 0 (0–99.9) | 0.956 | 0.05 (0–98.8) | 0 (0–99.9) | 0.875 |
| Relative abundance of Lactobacilus jenseni (%) | 0 (0–98.3) | 0 (0–89.3) | 0.816 | 0 (0–98.3) | 0 (0–89.3) | 0.991 |
| Proportion of Lactobacillus iners within total Lactobacillus (%) | 0.05 (0–100) | 0 (0–100) | 0.91 | 0.1 (0–100) | 0 (0–100) | 0.964 |
| Relative abundance of Lactobacillus species excluding iners (%) | 59.0 (0–99.9) | 25.2 (0–100) | 0.733 | 57.3 (0–100) | 31.7 (0–100) | 0.858 |
| Presence of Bifidobacterium species | 4 (25.0%) | 6 (8.5%) | 0.15 | 4 (23.5%) | 6 (13.3%) | 0.19 |
| Presence of Atopobium species | 5 (31.3%) | 15 (21.1%) | 0.589 | 6 (35.3%) | 14 (20.0%) | 0.292 |
| Presence of Prevotella species | 7 (43.8%) | 21 (28.0%) | 0.424 | 7 (41.2%) | 21 (30.0%) | 0.348 |
| Presence of Gardnerella species | 5 (31.3%) | 24 (33.8%) | 0.922 | 5 (29.4%) | 24 (34.3%) | 0.924 |
| Presence of Anaerococcus species | 4 (25.0%) | 3 (4.2%) | 0.02 | 4 (23.5%) | 3 (4.3%) | 0.034 |
| Presence of Dialister species | 6 (37.5%) | 13 (18.3%) | 0.18 | 6 (35.3%) | 13 (18.6%) | 0.242 |
| Presence of Ureaplasma species | 4 (25.0%) | 12 (16.9%) | 0.69 | 4 (23.5%) | 12 (17.1%) | 0.794 |
| Presence of Mycoplasma species | 1 (6.3%) | 1 (1.4%) | 0.27 | 1 (5.9%) | 1 (1.4%) | 0.354 |
| Presence of Escherichia species | 0 (0%) | 2 (2.8%) | 1 | 0 (0%) | 2 (2.9%) | 1 |
| Presence of Streptococcus species | 5 (31.3%) | 7 (9.9%) | 0.086 | 5 (29.4%) | 7 (10.0%) | 0.091 |
| Reproductive failure related species (%) | 10 (62.5%) | 41 (57.7%) | 0.946 | 11 (64.7%) | 40 (51.1%) | 0.769 |
Note: Data are shown median (range) or n (%).
Abbreviation: RPL, recurrent pregnancy loss.
3.4. Analysis of RIF and RPL Patients With and Without aPLs
Among women with RIF, although the sample size was limited, the aPL positive group (n = 3) showed a higher trend toward anti‐β2GPI/HLA‐DR antibody positivity compared with the aPL‐negative group (n = 51), with positivity rates of 66.7% (2/3) vs. 15.7% (8/51), respectively (p = 0.08). However, no significant differences were found between aPL positive and negative groups in numbers of bacterial species (4.0 [4–11] vs. 4 [1–16]), CE prevalence (0% [0/0] vs. 13.7% [7/51]), CD138‐positive cell count (0.73 [0–3.96] vs. 0.51 [0–58.3]), or relative abundance of Lactobacillus species (94.4 [0–99.5] vs. 98.8 [0–100]). Also, there were no differences between the groups regarding the prevalence of reproductive‐failure‐related bacterial species (100% [3/3] vs. 54.9% [28/51], p = 0.25) (Table S1).
Among women with RPL, there were no significant differences between aPL‐positive (n = 6) and aPL‐negative (n = 81) patients regarding the percentage of anti‐β2GPI/HLA‐DR antibody positivity (0% [0] vs. 21.0% [17/81]), numbers of bacterial species (3.5 [2–13] vs. 4.0 [1–17]), CE prevalence (16.7% [1/6] vs. 25.9% [21/81]), CD138‐positive cell counts (1.58 [0–8.45] vs. 1.3 [0–252.6]), or relative abundance of Lactobacillus species (96.2 [0–99.8] vs. 98.8 [0–100]). Additionally, there were no differences between the groups regarding the prevalence of reproductive failure–related bacterial species (50.0% [3/6] vs. 59.3% [48/81], p = 0.69) (Supporting Information S1).
4. Discussion
To the best of our knowledge, this is the first study to examine, in patients with RIF and RPL, the relationships among anti‐β2GPI/HLA‐DR antibodies, the uterine endometrial microbiome, and CE. Our findings in patients with RIF showed that although the frequency of CE did not differ between the antibody‐positive and antibody‐negative groups, the overall prevalence of reproductive‐failure‐related bacteria was significantly higher in the antibody‐positive group. In contrast, in patients with RPL there was no difference in the prevalence of either the reproductive‐failure‐related bacteria or CE between groups, suggesting that the two conditions may reflect distinct pathophysiological backgrounds.
Regarding specific bacterial species, Lactobacillus iners was significantly more frequent in the antibody‐positive group not only in the RIF group but also in the overall analysis, and in the multivariable model using the 95% cut‐off in RIF analysis, Lactobacillus iners remained an independently associated factor for anti‐β2GPI/HLA‐DR antibody positivity. Lactobacillus iners is considered a transitional Lactobacillus species with low lactic acid–producing capacity and is known to predominate more readily under inflammatory conditions such as bacterial vaginosis (BV) and chronic endometritis (CE) [8, 26]. Furthermore, dominance of Lactobacillus iners has been associated with lower implantation rates and poorer in vitro fertilization outcomes as compared with dominance of Lactobacillus crispatus [27]. Accordingly, the increased presence of Lactobacillus iners in antibody‐positive cases may reflect an early shift toward dysbiosis in the endometrial microbiome, even when apparent Lactobacillus dominance is maintained. An important point is that the biological significance of Lactobacillus iners should not be interpreted solely on the basis of small differences in relative abundance. Rather, its significance lies in the higher detection frequency observed in the antibody‐positive group and its independent association with antibody positivity in multivariable analysis. Because 16S rRNA sequencing evaluates relative abundance rather than absolute bacterial load, values close to zero may represent levels below the detection limit rather than true absence. Therefore, detection of Lactobacillus iners is not interpreted as evidence that minute quantities directly exert biological effects, but rather as a marker of an early shift toward dysbiosis, even when apparent Lactobacillus dominance seems to be maintained. In addition, we demonstrated that, in patients with RIF, the proportion of Lactobacillus iners within total Lactobacillus was also higher in antibody‐positive women, reaching statistical significance at the 95% percentile cut‐off. This finding further supports the notion that antibody‐positive RIF cases are characterized not only by increased detection of Lactobacilus iners, but also by a relative expansion of this species within the Lactobacillus community. In contrast, such differences were not observed in patients with RPL, suggesting that this microbial shift may be more specifically related to implantation failure rather than pregnancy loss. Functional validation, including absolute quantification and assessment of inflammatory markers, will be required to clarify causality in future studies. These findings suggest that assessment of anti‐β2GPI/HLA‐DR antibodies in RIF women with Lactobacillus iners positivity may be useful for the early detection of abnormalities in the endometrial environment.
Furthermore, Anaerococcus species were more frequently observed in the antibody positive group in the overall cohort, as well as the RIF and RPL analyzes. Previous reports indicate that the Anaerococcus genus is often present in intra‐uterine or vaginal dysbiosis associated with adverse reproductive outcomes such as CE, pelvic inflammatory disease, and endometrial cancer [28, 29]. Additionally, vaginal or lower genital tract microbial communities dominated by the Anaerococcus genus have been linked with elevated concentrations of inflammatory cytokines including IL‐6 and TNFα [30]. Taken together, these findings suggest that the Anaerococcus genus could contribute to the formation of a pro‐inflammatory, intra‐uterine environment via cytokine induction, thereby potentially playing a role in reproductive failure. Given the biological pathway by which β2GPI/HLA‐DR antigens can be induced via inflammatory cytokines [31], this hypothesis is coherent; however, at present there is no evidence demonstrating that this organism directly induces β2GPI/HLA‐DR antigen expression. This remains a topic for future investigation.
Most notably, in patients with RIF, anti‐β2GPI/HLA‐DR antibody positivity emerged as an independent associated factor for the presence of reproductive‐failure‐related bacterial species. This suggests that anti‐β2GPI/HLA‐DR antibodies are closely linked to intra‐uterine dysbiosis in RIF and may serve as a novel marker for the evaluation of the intra‐uterine environment in these patients. In RIF, compromised endometrial epithelial barrier function and disruption of mucosal immunity have been reported [32, 33]. It is conceivable that mild inflammatory stimuli could induce expression of the β2GPI/HLA‐DR complex. A biological framework in which epithelial stress caused by inflammation or dysbiosis contributes to the expression of β2GPI/HLA‐DR antigen and subsequent autoantibody production is coherent with the associations observed in this study and warrants investigation in the future.
One plausible biological mechanism underlying our findings is that local inflammatory responses associated with endometrial dysbiosis induce expression of the β2GPI/HLA‐DR complex in endometrial epithelial cells, thereby promoting neoself antigen presentation and subsequent production of anti‐β2GPI/HLA‐DR antibodies. In our previous studies, coordinated expression of β2GPI and HLA‐DR antigens has been demonstrated in endometrial and decidual tissues of patients with recurrent implantation failure and recurrent pregnancy loss [18, 23]. Because neoself antigens are thought to be induced by infection or inflammation [31, 34], dysbiosis‐associated local inflammation and inflammatory cytokines may enhance neoself antigen presentation pathways in the endometrium. Future studies should evaluate inflammatory cytokine expression and epithelial barrier function in the endometrium, analyze longitudinal changes in microbiome composition and antibody levels, and utilize experimental models in which endometrial epithelial cells are exposed to dysbiosis‐associated inflammatory stimuli to determine whether β2GPI/HLA‐DR expression is directly induced. Such approaches will help clarify the causal relationships among dysbiosis, neoself antigen expression, and antibody production.
In contrast, in RPL, no overall increase in reproductive‐failure‐related microbiome was observed in antibody‐positive cases. A possible explanation for the different associations observed between RIF and RPL is the difference in timing and anatomical location at which neoself antigen expression becomes relevant. Previous studies have shown that β2GPI/HLA‐DR complexes are expressed in decidual and placental endothelial cells in patients with obstetric antiphospholipid syndrome, where binding of autoantibodies can lead to complement activation, endothelial injury, and thrombosis, ultimately contributing to pregnancy loss [18]. Thus, in RPL, neoself antigen expression and antibody‐mediated pathology may primarily occur after implantation during early placentation and may not necessarily depend on preconception endometrial microbiome abnormalities. In contrast, RIF mainly reflects implantation failure within the preconception endometrial environment. In this setting, dysbiosis‐associated local inflammation may induce HLA‐DR expression in endometrial epithelial cells, promoting formation of β2GPI/HLA‐DR complexes and subsequent neoself antigen presentation, which may contribute to antibody production and impaired implantation. Our findings therefore suggest that anti‐β2GPI/HLA‐DR antibodies may be linked to dysbiosis‐driven mechanisms in RIF, whereas in RPL the antibody may contribute through post‐implantation vascular and inflammatory pathways that are independent of preconception dysbiosis. Future longitudinal studies evaluating microbiome composition, inflammatory responses, and neoself antigen expression before and during early pregnancy are needed to clarify these mechanisms.
Interestingly, in the present study, conventional aPLs were not significantly associated with either the composition of the uterine microbiome or the prevalence of reproductive‐failure‐related bacterial species in patients with either RIF or RPL. In contrast, the anti‐β2GPI/HLA‐DR antibody was associated with endometrial dysbiosis only in patients with RIF. This discrepancy suggests that, unlike conventional aPLs, the anti‐β2GPI/HLA‐DR antibodies may more sensitively reflect abnormalities in the local intra‐uterine environment specific to RIF. However, the pathophysiological significance and causal relationship of this association remain to be elucidated and require further investigation. Collectively, these findings indicate that anti‐β2GPI/HLA‐DR antibodies may serve as a potential marker for detecting endometrial dysbiosis in patients with RIF, and they could be a promising novel biomarker of implantation failure.
There are several limitations to this study. First, the number of antibody‐positive cases was relatively small, and validation of the observed associations in a larger cohort is warranted. Second, as this study was observational in nature, causality between anti‐β2GPI/HLA‐DR antibodies and microbiome abnormalities cannot be established, and potential interactions and bidirectionality should be further investigated. Third, antiphospholipid antibodies (aPLs) were evaluated based on a single measurement, and persistence of positivity was not confirmed by repeat testing at ≥ 12 weeks. Therefore, transient positivity and possible misclassification of antibody status cannot be excluded. Fourth, as follow‐up data on ART outcomes and pregnancy prognosis were not presented in this study, the impact of antibody positivity and microbiome abnormalities on actual reproductive outcomes remains to be clarified. Despite these limitations, this study suggests that anti‐β2GPI/HLA‐DR antibodies may be positioned as surrogate markers of the uterine endometrial environment, providing novel insight into the pathophysiology and therapeutic stratification of RIF. From a clinical perspective, in the diagnostic workup of patients with RIF, if an endometrial microbiome dysbiosis is detected, testing for anti‐β2GPI/HLA‐DR antibodies should be considered. Conversely, in patients where the presence of the antibody is identified, evaluation of the uterine endometrial microbiome may be warranted, given the potential bidirectional association observed in this study. This reciprocal assessment may help identify patients with subtle intra‐uterine inflammatory changes who could benefit from early intervention. Although the present study was observational and not designed to evaluate treatment efficacy, our findings may have potential implications for future clinical management strategies. In patients showing both detection of Lactobacillus iners or reproductive‐failure–related bacterial species and anti‐β2GPI/HLA‐DR antibody positivity, impaired implantation may reflect the combined effects of abnormalities in the preconception endometrial environment and antibody‐associated inflammatory and thromboinflammatory mechanisms. Previous studies have reported that treatment of chronic endometritis or dysbiosis with antibiotics may improve the endometrial environment and reproductive outcomes, suggesting that correction of microbial imbalance prior to pregnancy or fertility treatment may be beneficial [9, 10, 14]. In addition, our recent study demonstrated that, among anti‐β2GPI/HLA‐DR antibody–positive patients undergoing ART, administration of low‐dose aspirin combined with heparin starting before embryo transfer was associated with improved pregnancy rates, suggesting that suppression of antibody‐related inflammatory and coagulation abnormalities during the implantation period may also be important. Taken together, a stepwise approach may be considered in the future, in which dysbiosis is addressed prior to pregnancy using antimicrobial or probiotic interventions, followed by anticoagulant therapy during the implantation period. However, whether anticoagulant therapy itself can directly modify the endometrial microbiome or reduce the presence of Lactobacillus iners remains unknown, and prospective interventional studies are required to clarify these effects.
Recent reports have demonstrated that in patients with anti‐β2GPI/HLA‐DR antibody positivity, both among those with RPL and those undergoing ART, treatment with low‐dose aspirin or heparin was associated with a lower miscarriage rate and/or a higher clinical pregnancy rate [20, 23]. Future studies should investigate whether, in antibody‐positive patients with RIF, combining antimicrobial interventions, such as antibiotics, probiotics, or intra‐uterine lavage, with low‐dose aspirin and heparin therapy may be effective. This could potentially lead to an individualized treatment strategy based on anti‐β2GPI/HLA‐DR antibody status in infertile patients.
5. Conclusion
This study is the first, to the best of our knowledge, to demonstrate an association between anti‐β2GPI/HLA‐DR antibodies and abnormalities of the uterine endometrial microbiome, particularly in patients with RIF. This suggests that antibody positivity may serve as an indicator of an impaired uterine environment. Further large‐scale prospective and interventional studies are needed to aid in the development of preconception care strategies utilizing anti‐β2GPI/HLA‐DR antibodies.
Funding
This research was funded by the Japan Agency for Medical Research and Development (Grants JP21gk0110047, JP22gk0110061, JP23fk0108682, and JP25gn0110094 to H.Y.) and the Japan Society for the Promotion of Science (Grants 20K09642, 23K08888, and 24K02691 to H.Y.).
Ethics Statement
This observational study was conducted in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects issued by the Ministry of Health, Labour and Welfare, Japan. The study protocol was approved by the ethics committee of Teine Keijinkai Hospital, and the University of Yamanashi approved this prospective cohort study (Approval number: 020387‐00, 2‐023209‐00).
Consent
Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Clinical characteristics of RIF and RPL patients with and without aPLs.
Acknowledgments
We thank all members of the two medical institutions for their contributions to this study and the AMED‐funded research group (Grant Number JP24gn0110085) for their valuable scientific input; this study received no direct financial support from the grant.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Ma J., Gao W., and Li D., “Recurrent Implantation Failure: A Comprehensive Summary From Etiology to Treatment,” Frontiers in Endocrinology 13 (2023): 1061766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Pirtea P., Cicinelli E., De Nola R., de Ziegler D., and Ayoubi J. M., “Endometrial Causes of Recurrent Pregnancy Losses: Endometriosis, Adenomyosis, and Chronic Endometritis,” Fertility and Sterility 115 (2021): 546–560. [DOI] [PubMed] [Google Scholar]
- 3. Pluchino N., Drakopoulos P., Wenger J. M., Petignat P., Streuli I., and Genazzani A. R., “Hormonal Causes of Recurrent Pregnancy Loss (RPL),” Hormones (Athens, Greece) 13 (2014): 314–322. [DOI] [PubMed] [Google Scholar]
- 4. ESHRE Working Group on Recurrent Implantation Failure , Cimadomo D., de Los Santos M. J., et al., “ESHRE Good Practice Recommendations on Recurrent Implantation Failure,” Human Reproduction Open 2023 (2023): hoad023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. de Assis V., Giugni C. S., and Ros S. T., “Evaluation of Recurrent Pregnancy Loss,” Obstetrics and Gynecology 143 (2024): 645–659. [DOI] [PubMed] [Google Scholar]
- 6. Günther V., Allahqoli L., Deenadayal‐Mettler A., et al., “Molecular Determinants of Uterine Receptivity: Comparison of Successful Implantation, Recurrent Miscarriage, and Recurrent Implantation Failure,” International Journal of Molecular Sciences 24 (2023): 17616. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Zhang H., Zou H., Zhang C., and Zhang S., “Chronic Endometritis and the Endometrial Microbiota: Implications for Reproductive Success in Patients With Recurrent Implantation Failure,” Annals of Clinical Microbiology and Antimicrobials 23 (2024): 49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Takimoto K., Yamada H., Shimada S., Fukushi Y., and Wada S., “Chronic Endometritis and Uterine Endometrium Microbiota in Recurrent Implantation Failure and Recurrent Pregnancy Loss,” Biomedicine 11 (2023): 2391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cicinelli E., Matteo M., Tinelli R., et al., “Prevalence of Chronic Endometritis in Repeated Unexplained Implantation Failure and the IVF Success Rate After Antibiotic Therapy,” Human Reproduction 30 (2015): 323–330. [DOI] [PubMed] [Google Scholar]
- 10. Cheng X., Huang Z., Xiao Z., and Bai Y., “Does Antibiotic Therapy for Chronic Endometritis Improve Clinical Outcomes of Patients With Recurrent Implantation Failure in Subsequent IVF Cycles? A Systematic Review and Meta‐Analysis,” Journal of Assisted Reproduction and Genetics 39 (2022): 1797–1813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Liu F. T., Yang S., Yang Z., et al., “An Altered Microbiota in the Lower and Upper Female Reproductive Tract of Women With Recurrent Spontaneous Abortion,” Microbiology Spectrum 10 (2022): e0046222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Shi Y., Yamada H., Sasagawa Y., Tanimura K., and Deguchi M., “Uterine Endometrium Microbiota and Pregnancy Outcome in Women With Recurrent Pregnancy Loss,” Journal of Reproductive Immunology 152 (2022): 103653. [DOI] [PubMed] [Google Scholar]
- 13. McQueen D. B., Perfetto C. O., Hazard F. K., and Lathi R. B., “Pregnancy Outcomes in Women With Chronic Endometritis and Recurrent Pregnancy Loss,” Fertility and Sterility 104 (2015): 927–931. [DOI] [PubMed] [Google Scholar]
- 14. Gay C., Hamdaoui N., Pauly V., et al., “Impact of Antibiotic Treatment for Chronic Endometritis on Unexplained Recurrent Pregnancy Loss,” Journal of Gynecology Obstetrics and Human Reproduction 50 (2021): 102034. [DOI] [PubMed] [Google Scholar]
- 15. Practice Committee of the American Society for Reproductive Medicine , “Evaluation and Treatment of Recurrent Pregnancy Loss: A Committee Opinion,” Fertility and Sterility 98 (2012): 1103–1111. [DOI] [PubMed] [Google Scholar]
- 16. Papadimitriou E., Boutzios G., Mathioudakis A. G., Vlahos N. F., Vlachoyiannopoulos P., and Mastorakos G., “Presence of Antiphospholipid Antibodies Is Associated With Increased Implantation Failure Following In Vitro Fertilization Technique and Embryo Transfer: A Systematic Review and Meta‐Analysis,” PLoS One 17 (2022): e0260759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Sauer R., Roussev R., Jeyendran R. S., and Coulam C. B., “Prevalence of Antiphospholipid Antibodies Among Women Experiencing Unexplained Infertility and Recurrent Implantation Failure,” Fertility and Sterility 93 (2010): 2441–2443. [DOI] [PubMed] [Google Scholar]
- 18. Tanimura K. T., Jin H., Suenaga T., et al., “β2‐Glycoprotein I/HLA Class II Complexes Are Novel Autoantigens in Antiphospholipid Syndrome,” Blood 125 (2015): 2835–2844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tanimura K., Saito S., Nakatsuka M., et al., “The β2‐Glycoprotein I/HLA‐DR Complex Is the Major Autoantibody Target in Obstetric Antiphospholipid Syndrome,” Arthritis and Rheumatology 72 (2020): 1882–1891. [DOI] [PubMed] [Google Scholar]
- 20. Tanimura K., Saito S., Tsuda S., et al., “Low‐Dose Aspirin and Heparin Treatment Improves Pregnancy Outcome in Recurrent Pregnancy Loss Women With Anti‐β2‐Glycoprotein I/HLA‐DR Autoantibodies: A Prospective, Multicenter, Observational Study,” Frontiers in Immunology 15 (2024): 1445852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Tanimura K., Saito S., Tsuda S., et al., “Anti‐β2‐Glycoprotein I/HLA‐DR Antibody and Adverse Obstetric Outcomes,” International Journal of Molecular Sciences 24 (2023): 10958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Ono Y., Wada S., Ota H., et al., “Anti‐β2‐Glycoprotein I/HLA‐DR Antibody in Infertility,” Journal of Reproductive Immunology 158 (2023): 103955. [DOI] [PubMed] [Google Scholar]
- 23. Ono Y., Wada S., Kobayashi Y., et al., “Pregnancy Outcomes in Women With Anti‐β2‐Glycoprotein I/HLA‐DR Autoantibodies Receiving Assisted Reproductive Technology: A Prospective Cohort Study,” Frontiers in Immunology 16 (2025): 1626862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Ono Y., Kobayashi Y., Shimada S., et al., “Uterine Endometrium Microbiome in Women With Repeated Implantation Failure Complicated by Endometriosis,” Journal of Clinical Medicine 13 (2024): 4605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Liu Y., Ko E. Y., Wong K. K., et al., “Endometrial Microbiota in Infertile Women With and Without Chronic Endometritis as Diagnosed Using a Quantitative and Reference Range‐Based Method,” Fertility and Sterility 112 (2019): 707–717.e1. [DOI] [PubMed] [Google Scholar]
- 26. Zheng N., Guo R., Wang J., Zhou W., and Ling Z., “Contribution of Lactobacillus iners to Vaginal Health and Diseases: A Systematic Review,” Frontiers in Cellular and Infection Microbiology 11 (2021): 792787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Guan W., Dong S., Wang Z., Jiao J., and Wang X., “Impact of a Lactobacillus‐Dominant Cervical Microbiome, Based on 16S‐FAST Profiling, on the Reproductive Outcomes of IVF Patients,” Frontiers in Cellular and Infection Microbiology 13 (2023): 1059339, 10.3389/fcimb.2023.1059339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Medina‐Bastidas D., Camacho‐Arroyo I., and García‐Gómez E., “Current Findings in Endometrial Microbiome: Impact on Uterine Diseases,” Reproduction 163 (2022): R81–R96. [DOI] [PubMed] [Google Scholar]
- 29. Kuźmycz O., Kowalczyk A., Bolanowska A., et al., “A Comprehensive Analysis of the Uterine Microbiome in Endometrial Cancer Patients – Identification of Anaerococcus as a Potential Biomarker and Carcinogenic Cofactor,” Frontiers in Cellular and Infection Microbiology 15 (2025): 1511625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Jimenez N., Norton T., Diadala G., et al., “Vaginal and Rectal Microbiome Contribute to Genital Inflammation in Chronic Pelvic Pain,” BMC Medicine 22 (2024): 283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Jin H., Kishida K., Arase N., et al., “Abrogation of Self‐Tolerance by Misfolded Self‐Antigens Complexed With MHC Class II Molecules,” Science Advances 8 (2022): eabj9867. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Robertson S. A., Moldenhauer L. M., Green E. S., Care A. S., and Hull M. L., “Immune Determinants of Endometrial Receptivity: A Biological Perspective,” Fertility and Sterility 117 (2022): 1107–1120. [DOI] [PubMed] [Google Scholar]
- 33. Wang F., Qualls A. E., Marques‐Fernandez L., and Colucci F., “Biology and Pathology of the Uterine Microenvironment and Its Natural Killer Cells,” Cellular & Molecular Immunology 18 (2021): 2101–2113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Jin H., Arase N., Hirayasu K., et al., “Autoantibodies to IgG/HLA Class II Complexes Are Associated With Rheumatoid Arthritis Susceptibility,” Proceedings of the National Academy of Sciences of the United States of America 111 (2014): 3787–3792. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Clinical characteristics of RIF and RPL patients with and without aPLs.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.