It is postulated that women suffered from recurrent implantation failure (RIF) have different endometrial receptivity compared to those who experienced with idiopathic recurrent miscarriage (RM). In this study, expression of common endometrial markers Leukemia inhibitor factor (LIF), mucin 1 (MUC1) and integrin β3 were studied and compared. Fourteen women with RIF, 25 with RM and 20 fertile controls were recruited for endometrial biopsy during implantation window on day LH + 7. Spatial and temporal expression of MUC1, LIF and Integrin β3 were compared using semi-quantitative immunohistochemistry. Association of MUC1, LIF and integrin β3 expression levels with demographic and clinical characteristics were determined. MUC1 expression in both luminal and glandular epithelium in women with RIF were significantly lower than that in women with RM and fertile controls. There were no differences in LIF and Integrin β3 expression in endometrial epithelium among three groups. Decreased MUC1 expression were not significantly associated with age, BMI, gravidity, parity, cycle length, progesterone level and previous miscarriage. Deceased expression of MUC1 is an independent marker for endometrial receptivity in RIF women, suggesting MUC1 may contribute to the reproductive failure in RIF women.

Background: A recent study has reported that the microbiota in endometrial fluid of patients receiving in vitro fertilization and embryo transfer (IVF-ET) may predict implantation and pregnancy rates. However, studies are lacking that simultaneously compare the microbiota between endometrial fluid and tissue samples. Whether the microbiota composition in endometrial fluid reflects that in the endometrial tissue remains unclear. Methods: We systematically profiled the microbiota in endometrial fluid and tissue samples of IVF-ET patients using massively parallel sequencing. The bacterial 16S ribosomal RNA gene (V4 region) was PCR-amplified. Sequencing reads with >98% nucleotide identity were clustered as a bacterial taxon. To account for the different number of reads per sample, we normalized the read counts of each taxon before comparing its relative abundances across samples. Results: Thirteen taxa, including Verrucomicrobiaceae, Brevundimonas, Achromobacter, Exiguobacterium, and Flavobacterium, were consistently detected only in endometrial tissue samples but not fluid samples. Eight taxa were detected in fluid but not tissue. Twenty-two taxa were differentially abundant between fluid and tissue samples (adjusted P values, 4.1 × 10⁻²⁵ to 0.025). The numbers of taxa identified per 1000 sequencing reads, diversity, and evenness in fluid samples were smaller than those in tissue samples. Conclusions: Our data suggest that the microbiota composition in endometrial fluid does not fully reflect that in endometrial tissue. Sampling from both endometrial fluid and biopsy allows a more comprehensive view of microbial colonization. Further efforts are needed to identify the preanalytical effects, including sampling sites, methods, and sequencing depth, on profiling endometrial microbiota.