The alteration of uterine microbiota participated in the activation of the decidual inflammatory response in early spontaneous abortion.

Spontaneous abortion is one of the most common clinically recognized pregnancy complications in the general population, with an incidence of approximately 15%, with more than 80% of cases considered early spontaneous abortion (ESA) as they occur before 12 weeks [ 1 ]. The consequences of miscarriage are both physical and psychological, such as infection, bleeding, depression, anxiety, and posttraumatic stress disorder [ 2 ]. Many factors may contribute to ESA, including embryo chromosomal abnormalities and maternal factors, such as immune dysfunctions, endocrine disorders, uterine dysplasia or malformation, antiphospholipid syndrome and thrombophilias [ 1 – 4 ]. With the wide application of preimplantation diagnosis technology, early screening and identification of maternal risk factors has become increasingly important to further investigate the potential aetiologies, prevention, and treatment of spontaneous abortion. Studies [ 5 – 8 ] on the role of the vaginal microbiota in early pregnancy miscarriage have shown that patients with ESA have reduced Lactobacillus spp.-dominated vaginal microbiota and greater diversity and richness of the bacterial community. The endometrial microbiota plays a crucial role in several obstetrical complications, increasing attention has been focused on the uterine microenvironment to explain the causes of miscarriage [ 9 , 10 ]. Altered uterine microbiome composition reportedly contributes to poor reproductive outcomes, such as failed embryo implantation [ 11 ], early spontaneous abortion [ 12 ] and recurrent spontaneous abortion [ 13 ]. However, clinical research about the upper genital tract with ESA is limited, for the difficulty in collecting samples and the possibility of contamination. Alterations in the composition of uterine microorganisms can induce an inflammatory response, which is closely linked to the adhesion of the blastocyst to the epithelial endometrial wall. Researchers have shown that successful pregnancy requires the involvement of an inflammatory response, with alterations in cytokines and chemokines [ 14 ]. However, over-activated immune response may be responsible for spontaneous abortion. Endometrial expression of the pro-inflammatory cytokines (IL-1β, TNF-α, IFN-γ and TGF-β1) and prostaglandin E2 (PGE2) are upregulated and the expression of anti-inflammatory cytokines (IL-4 and IL-10) and angiogenesis-associated cytokines (IL-2, IL-6, and IL-8) are downregulated in idiopathic recurrent spontaneous miscarriage women compared with normal fertile women [ 15 ]. Are changes in decidual inflammatory cytokines associated with alterations in the uterine microbiota? Do the changes in inflammatory cytokines and the uterine microbiota contribute to early spontaneous abortion? In this study, we analysed the clinical characteristics of ESA without embryo abnormalities compared with those of normal pregnancies to determine the effects of the composition of uterine microbial communities and inflammation-related indicators and explore the high-risk factors of ESA.

A total of 154 ESA patients and 128 AA patients were initially screened; however, after applying exclusion criteria, 28 ESA patients and 111 AA patients were ultimately included in the retrospective clinical trait analysis ( Fig 1 ). Compared with AA patients, ESA patients generally presented a greater inflammatory response, with an increased number of leukocytes, higher counts and percentages of neutrophils and a decreased percentage of lymphocytes ( Fig 2A , S1 Fig and Table 1 ). No obvious differences in age, foetal size, the number or percentage of monocytes, number of lymphocytes, blood platelet count, clarity of leucorrhoea, leukocytes in leucorrhoea, or the level of Lactobacillus in leucorrhoea were detected between the two groups ( S1 Fig and Table 1 ). The correlation analysis also indicated that the occurrence of ESA showed positive relation with the number of leukocytes and neutrophils, and negative relation with the number and percentage of lymphocytes ( Fig 2B ). A logistic regression model was used to assess the association between age, leucorrhea cleanliness, leucorrhea cell content and leucorrhea lactobacillus content (independent variables). Age (β = −0.235, p = 0.017) and leucorrhea cleanliness (β = −1.429, p = 0.006) were significantly negatively associated with the outcome, indicating that increases in these variables reduced the odds of the condition. Conversely, leucorrhea cell content (β = 2.559, p = 0.005) showed a significant positive association, suggesting that higher levels increased the odds of the condition. Leucorrhea bacterial content (β = 0.077, p = 0.802) was not significantly associated with the outcome (S1 Table in S1 Text ). ESA, early spontaneous abortion; AA, artificial abortion. (A) Comparison of the number of leukocytes, percentage of lymphocytes, the number and percentage of neutrophils between ESA and AA groups. Each dot represents an individual ESA (red), AA (blue). (B) Spearman correlation and hierarchical clustering of indicated features for ESA and AA groups. Significance determined by spearman correlation test and student’s t test: * p <  0.05, **p <  0.01 and ***p <  0.001. ESA, early spontaneous abortion; AA, artificial abortion. Since the ESA group was sampled with no information on whether embryo chromosome abnormalities were present or not. Therefore, when the embryonic chromosome results were available, some samples in the ESA group were excluded because of embryonic chromosome abnormalities. To further explore the inflammatory changes of ESA patients, decidual tissues of 11 ESA and 19 AA patients were collected. We applied a cytokine array comprising 40 human inflammatory cytokines to detect alterations in the inflammatory status of the decidua of ESA patients. Correlation analysis was performed to investigate the potential association between the clinical index and cytokine expression. According to our results, ESA was associated with increased expression of multiple cytokines, including IL-17, TNF-β, IL-1α, IL-1β, IL-7, IFN-γ, IL-2, IL-4, IL-11, and IL-16 ( Fig 3 ). The clustering heatmap shows representative cytokine expression in the two groups. Compared with those in the AA group, I-309, IL-1α, IL-1β, IL-12p40 and MIG were relatively upregulated, whereas MIP-1δ, BLC and TNF RII were downregulated in the ESA group ( Fig 4A ). Further statistical analysis revealed significantly increased expression of IL-1β in the ESA group ( Fig 4B ), which was confirmed by quantitative PCR ( S2A Fig ). Significance was assessed using Spearmans correlation test and Bayesian analysis. For Bayesian tests, the Benjamini-Hochberg correction was applied to control the false discovery rate across multiple comparisons. Corrected results indicate the strength of evidence after adjusting for multiple hypotheses: * p <  0.05, **p <  0.01 and ***p <  0.001. ESA, early spontaneous abortion; AA, artificial abortion. (A) Heatmap of the representative cytokines expression in decidua between ESA and AA groups. (B) Volcano plot showed the significantly upregulated or downregulated cytokines in decidua between ESA and AA groups. (C) The immune-histochemistry staining of IL-1β in decidua of ESA and AA group. (D) The comparison of IL-1β integrate optical intensity (IOD) between ESA and AA group. (E) The double-immunofluorescence staining of IL-1β (green), prolactin (red) and DAPI (blue). A representative image of IL-1β staining in the decidua is shown in Fig 4C , and the mean fluorescence intensity (MFI) of IL-1β was calculated and compared between the two groups. Consistent with the results of the cytokine test and RNA quantification, the expression of IL-1β was significantly greater in the ESA group (P < 0.05, Fig 4D ). IL-1β-positive cells were identified as decidual stromal cells by double immunofluorescence staining for IL-1β and prolactin ( Fig 4E ). To confirm whether the increased expression of decidual IL-1β was associated with chronic deciduitis, we selected CD138 and MUM1 for histological evaluation. However, no significant difference in CD138- or MUM1-positive staining was found between these two groups ( S2D-E Fig ). To further investigate whether changes in the uterine microbiome were related with inflammatory response, 16S rRNA gene sequencing was performed. There were 11 uterine swab samples in the ESA group and 19 uterine swab samples in the AA group. Four uterine swab samples from the ESA group and nine uterine swab samples from the AA group were excluded because of DNA extraction failure. A total of 118,522 CCS sequences were produced from the 17 samples (7 ESA and 10 AA). Each sample generated at least 4,224 CCS sequences, and an average of 6,972 CCS sequences were obtained. After clustering, 1,512 and 572 OTUs were obtained from the ESA and AA groups, respectively (S2 Table in S1 Text ), and 413 OTUs overlapped between the two groups ( Fig 5A ). Six OTUs were shared among all the samples in the ESA group, whereas no common OTUs were found across the samples in the AA group (S3 Table in S1 Text , S3A-B Fig ). (A) Venn diagram between ESA patient and AA groups at OTUs level. (B) Hierarchical clustering heatmap of samples by weighted-unifrac algorithm. (C) The comparison of alpha diversity between ESA patients and AA groups showed by Chao1 index. (D) The comparison of alpha diversity between ESA patients and AA patients showed by Shannon index. (E) The top10 relative abundance of different bacteria in the two groups at genus level. The middle is the difference in the abundance of bacterial groups within the 95% confidence interval; the right side is the P value, P < 0.05 means the difference is statistically significant. (F) The top5 linear discriminant analysis (LDA) scores obtained from the linear discriminant analysis effect size (LEfSe) analysis of the uterine microbiota in the two groups. Significance determined by Student t test or Wilcox rank sum test: * p <  0.05, **p <  0.01 and ***p <  0.001. ESA, early spontaneous abortion; AA, artificial abortion. A simple overview of the overall sample distribution was obtained via a clustering heatmap and principal component analysis (PCA) ( Fig 5B and S3C Fig ). In brief, samples were roughly clustered by group, and samples within groups presented slight differences from each other. The Chao1 index and Shannon index were calculated to estimate the alpha diversity. The Chao1 index (474.69 ± 111.96 for the ESA group and 113.02 ± 26.85 for the AA group) and Shannon index (5.57 ± 0.83 for the ESA group and 1.07 ± 0.54 for the AA group) were significantly greater in the ESA group than in the AA group ( Fig 5C - D ). ANOSIM, which estimates beta diversity, revealed that the differences between groups were greater than the differences within groups (R =  0.56, P = 0.001) ( S3D Fig ). To investigate whether specific uterine bacteria contribute to the increased risk of ESA, we performed a metastats analysis of the relative abundance between the two groups at the genus level (S4 Table in S1 Text ). Among the top ten genera, the ESA group presented a significantly greater abundance of nine genera, including Acidibacter, uncultured rumen bacteria , Lachnoclostridium, Alistipes, Desulfovibrio, Rhizobacter, unclassified Longimicrobiaceae, Fimbriimonas and unclassified Gaiellales ( Fig 5E ). However, Lactobacillus was predominant in the AA group. The relative abundance of each individual sample is also shown in S2E Fig . Furthermore, linear discriminant analysis effect size (LEfSe) analyses with linear discriminant analysis (LDA) at the species level were applied to both groups. The bar plot shows the top five most enriched bacterial species in each group ( Fig 5F ). Bacteroidota, Bacteroidia, Bacteroidales, Burkholderiales and Acidobacteriota were significantly enriched in the ESA group. The Lactobacillaceae, Lactobacillus, Lactobacillales, Bacilli and Firmicutes abundances in the AA group were significantly greater than those in the ESA group ( Fig 5F ). Functional prediction analysis was conducted on both groups based on Clusters of Orthologous Groups of proteins (COG) functional analysis. COG analysis revealed that the predictive functions of the ESA group were mainly related to amino acid transport metabolism, cell wall/membrane/envelope biogenesis, modification, energy generation and conversion, whereas those of the AA group were related to carbohydrate transport metabolism, replication recombination and repair, translation, ribosomal structure and biogenesis ( S3F Fig ).

In this research, we performed a detailed study on the uterine microbiomes of patients with ESA and women with AA. Our results highlight the complex interplay of IL-1β at the fetomaternal interface and its potential role in ESA. Changes in the diversity and composition of the uterine microbiota may be correlated with the inflammatory response and imbalance of the immunological response in ESA rather than chronic deciduitis. The study of the uterine microbiome provides insight into the pathogenesis of ESA. The mechanic interactions between microbiome, inflammatory factors and ESA remains unclear. Therefore, a larger sample of the population and investigations of the underlying causal mechanism of ESA are needed in further research to confirm our conclusions. The more investigation of the mechanisms need to be studied in future animal experiments.

This single-centre clinical observational study was conducted at West China Second University Hospital of Sichuan University. Ethical approval for the study was granted by the Ethics Committee of West China Second University Hospital of Sichuan University (No. 2023-350). The patients/participants provided their written informed consent to participate in this study. The study consisted of two parts. The first part was a retrospective analysis, in which the clinical characteristics and laboratory test results of ESA (early spontaneous abortion) and AA (artificial abortion) patients were collected, compared, and analysed. The requirement for informed consent was waived by the ethics committee for the retrospective analysis of the medical records, and the data were anonymised to remove any personally identifying information. The second part was an observational study, where decidual tissues and uterine secretions from ESA and AA patients were collected for histological analysis, cytokine detection, and microbial analysis. Written informed consent was obtained from all patients who provided samples. The retrospective analysis was started from 26/04/2023. Patients who met the following inclusion/exclusion criteria and underwent outpatient surgical termination of pregnancy between 01/01/2023 and 31/03/2023 were recruited. The included patients were divided into two groups according to the survival status of the embryos at the time of termination. Patients with a nonviable intrauterine pregnancy, such as an empty gestational sac or a gestational sac containing an embryo or foetus with no foetal heart activity, were included in the ESA group. The AA group included patients who chose to terminate unwanted pregnancies with normally developing embryos. Clinical information such as the patient’s age, maternal history, gestational sac size, medication use during pregnancy, routine blood, routine leucorrhoea and ultrasound findings were collected. The inclusion criteria were as follows: 1) Patients fully understood the potential risks and benefits of this research and agreed to participate. 2) Surgical termination of pregnancy up to 10 weeks of gestation, including artificial abortion and early spontaneous abortion. 3) Maternal age between 20 and 35 years. 4) The number of pregnancies was less than or equal to 3 (including the present time). 5) No COVID-19 virus infection during pregnancy. The exclusion criteria were as follows: 1) early spontaneous abortion without embryo chromosome examination or chromosomal abnormality in the embryo. 2) Patients with abnormalities in the immune system or the use of immunosuppressive drugs, such as hydroxychloroquine, prednisone, tacrolimus, etc. 3) Patients with acute genital tract inflammation, such as all kinds of vaginosis or HPV infection. 4) Two or more consecutive spontaneous abortions. 5) Any pathological disease affecting the endometrial cavity, such as endometrial polyps, submucosal myomas, intramural myomas and adenomyosis. The sample collection of the observational study was conducted between 01/06/2023 and 31/07/2023, and patients who met the inclusion and exclusion criteria were recruited during this period. Prior to the acquisition of the embryo chromosomal results, all samples were properly stored. Sample collection was conducted in a sterile operating room during surgical abortion procedures. To prevent contamination from the vagina and cervix, the operator proceeded to sterilise the vagina and cervix, wipe away any cervical secretions, and dilate the cervical canal in accordance with the surgical procedure. Prior to the collection of the microbiota sample, the cervical canal was disinfected with iodophor and the residual iodophor was removed with a sterile dry cotton swab. Two sterile dry cotton swabs were separately inserted into the uterine cavity to collect uterine secretions for microbiota analysis. These swabs were placed in sterilized tubes and stored at − 80°C until DNA extraction. After collecting the microbial samples, the surgical procedure was completed, and decidual endometrial tissues were harvested and stored. The decidual endometria were stored at −80°C for cytokine analysis or fixed in 4% paraformaldehyde, embedded in paraffin and sectioned for the subsequent haematoxylin and eosin (H&E) staining, immunohistochemical (IHC) staining, and immunofluorescence staining. IHC staining was conducted using antibodies against CD138 (Zhongshan Jinqiao Biotechnology Co., ZA-0584), MUM1 (Chengdu Zen-Bioscience Co., Ltd., R50091 ), and IL-1β (Proteintech, 16806-1-AP). Four nonoverlapping areas per section were photographed at 400x magnification. Image analysis was performed using Fuji ImageJ software to calculate the integrated optical density (IOD) and the area of positively stained cells. The mean density of IL-1β-positive protein expression was computed for quantitative analysis. To evaluate the inflammatory status, we counted the number of CD138- and MUM1-positive cells in four nonoverlapping high-magnification fields per section. Based on the number of CD138-positive or MUM1-positive cells in each field, sections were categorized into grades 0 – 2, following a previously established protocol [ 16 ]. Specifically, Grade 0 was assigned to sections with no positive cells, Grade 1 for sections with 1 – 4 positive cells, and Grade 2 for sections with 5 or more positive cells. Decidual endometria were collected to measure the levels of inflammatory cytokines using a G-Series Human Inflammation Array (Raybiotech), which profiles 40 cytokines. Following the manufacturer’s protocol, 100 µl of each sample was added to the wells of the cytokine array and incubated at room temperature for 1–2 hours. After incubation, the samples were washed thoroughly and subjected to subsequent incubation steps as per the provided instructions. Fluorescence detection was performed using a laser scanner equipped with a Cy3 wavelength (Axon GenePix). The resulting data were analyzed using microarray analysis software (GenePix) and RayBio analysis tools to quantify the levels of the inflammatory cytokines. Decidual RNA was extracted from samples via the QIAamp RNA Kit (Qiagen) according to the manufacturer’s instructions. The sequences of the primers used for IL-1β were as follows: forward, 5’-TGGCTTATTACAGTGGCAATGAGG-3’; reverse, 3’-AGTGGTGGTCGGAGATTCGTAG-5’. The amplifications were performed on a Droplet Digital PCR system (Bio-Rad) with the following 50°C for 15 minutes, 95°C for 3 minutes, followed by 45 cycles of 95°C for 15 seconds and 60°C for 30 seconds. Double-immunofluorescence staining was performed to characterize the IL-1β-expressing cells in the decidua, and the decidual stromal cells were identified with prolactin. The decidual endometrium slices were incubated with primary antibodies overnight at 4°C and then with fluorescent secondary antibodies for 1 h at room temperature. The primary and secondary antibodies used for IL-1β immunofluorescence staining were an anti-IL-1β rabbit polyclonal antibody (Proteintech, 16806-1-AP) and a fluorescein isothiocyanate (FITC)-labelled antibody (Chengdu Zen-Bioscience Co., Ltd., 550037), respectively. For prolactin immunofluorescence staining, a prolactin mouse monoclonal antibody (Chengdu Zen-Bioscience Co., Ltd., 222677) and a Cy5-labelled secondary antibody was used (Chengdu Zen-Bioscience Co., Ltd., 550047). The slides were embedded in DAPI containing mounting buffer (Vector Laboratories) and then observed under a Leica DM6B fluorescence microscope. IL-1β appeared green, whereas prolactin appeared red. Total DNA was extracted from the samples using the PowerSoil® DNA Isolation Kit (Qiagen) following the manufacturer’s instructions. For the amplification of the full-length 16S rRNA gene, specific primers were designed based on the conserved regions 27F and 1492R. The forward primer (27F) sequence was AGRGTTTGATYNTGGCTCAG, and the reverse primer (1492R) sequence was TASGGHTACCTTGTTASGACTT. PCR amplification of the target region was performed in a 10 μL reaction system using the Solexa PCR kit. The quality of the resulting sequencing library was inspected to ensure high-quality data. Following quality control, barcode recognition was performed on the high-quality circular consensus sequencing (CCS) sequences. The optimized CCS sequences were clustered at a 97% similarity level using USEARCH (version 10.0) to generate operational taxonomic units (OTUs). Species classification was based on the sequence composition of these OTUs [ 16 , 17 ]. Species annotation and taxonomic analysis were conducted using the Silva database and the RDP classifier. These tools were employed to analyze the diversity of the gut microbiota within the samples. Alpha diversity metrics, including the Chao1 index and Shannon index, were calculated to assess the richness and evenness of species within each sample. Beta diversity was evaluated using the Bray-Curtis dissimilarity metric to compare the differences in microbial community composition and structure between samples. Beta diversity analysis was utilized to compare the community composition and structure across different samples. Metastats analysis was performed to identify significant differences in the relative abundance of microbial genera between the ESA and AA groups. The functional potential of the microbial communities was predicted using PICRUSt2, with functional annotation based on the KEGG database ( https://www.kegg.jp/). Statistical analyses and visualizations were performed using GraphPad Prism version 9.0 (GraphPad Software), SPSS version 27.0, and R version 4.2. Continuous data were compared using Student’s t-test, while categorical data were analysed using the chi-square test, with Fisher’s exact test applied when necessary. Data are presented as means with standard deviations (SD) or as counts with percentages. Spearman’s correlation analysis was conducted to assess relationships between multiple factors. A logistic regression model was used to assess the association between multiple factors. Coefficients (β) were estimated using the maximum likelihood method, with standard errors (SE), z-values, p-values, and 95% confidence intervals (CI) computed to evaluate the significance and effect size of each predictor. A p-value of less than 0.05 was considered statistically significant.

Comparison of the number of mean age, gestational sac, fetus size, number of lymphocytes, number of platelets, the number and percentage of monocyte between ESA and AA groups. Each dot represents an individual ESA (red), AA (blue). Significance determined by Student’s t test: * p <  0.05. ESA, early spontaneous abortion; AA, artificial abortion. (TIF) (A) Comparison of IL-1β RNA expression between ESA and AA by qPCR. (B) Immunohistochemical (IHC) staining and Grading of CD138 in decidua. (C) Distribution of different grade of CD138 staining between ESA and AA. (D) Distribution of different grade of MUM1 staining between ESA and AA. (E) Immunohistochemical (IHC) staining and Grading of MUM1 in decidua. Significance determined by chi 2 test and Student’s t test: * p <  0.05. ESA, early spontaneous abortion; AA, artificial abortion. (TIF) (A) and (B) OTUs analysis of each sample and group by petal diagram, the central circle represented the number of OTUs common in the same group, and the non-overlapping part was the number of OTUs unique to each sample. (C) PCoA plot of weighted-unifrac distance between ESA and AA groups. (D) Beta diversity by Anosim analysis of weighted-unifrac distance within and between ESA and AA groups. (E) Relative abundances of the uterine microbiota of each sample at the genus level. (F) Clusters of Orthologous Groups of proteins (COG) functional comparison of significantly differentially abundant bacterial in ESA and AA group. ESA, early spontaneous abortion; AA, artificial abortion. (TIF) (XLSX)