Characterisation of the human uterine microbiome in non-pregnant women through deep sequencing of the V1-2 region of the 16S rRNA gene

Abstract

Background. It is widely assumed that the uterine cavity in non-pregnant women is physiologically sterile, also as a premise to the long-held view that human infants develop in a sterile uterine environment, though likely reflecting under-appraisal of the extent of the human bacterial metacommunity. In an exploratory study, we aimed to investigate the putative presence of a uterine microbiome in a selected series of non-pregnant women through deep sequencing of the V1-2 hypervariable region of the 16S ribosomal RNA (rRNA) gene. Methods. Nineteen women with various reproductive conditions, including subfertility, scheduled for hysteroscopy and not showing uterine anomalies were recruited. Subjects were highly diverse with regard to demographic and medical history and included nulliparous and parous women. Endometrial tissue and mucus harvesting was performed by use of a transcervical device designed to obtain endometrial biopsy, while avoiding cervicovaginal contamination. Bacteria were targeted by use of a barcoded Illumina MiSeq paired-end sequencing method targeting the 16S rRNA gene V1-2 region, yielding an average of 41,194 reads per sample after quality filtering. Taxonomic annotation was pursued by comparison with sequences available through the Ribosomal Database Project and the NCBI database. Results. Out of 183 unique 16S rRNA gene amplicon sequences, 15 phylotypes were present in all samples. In some 90% of the women included, community architecture was fairly similar inasmuch B. xylanisolvens, B. thetaiotaomicron, B. fragilis and an undetermined Pelomonas taxon constituted over one third of the endometrial bacterial community. On the singular phylotype level, six women showed predominance of L. crispatus or L. iners in the presence of the Bacteroides core. Two endometrial communities were highly dissimilar, largely lacking the Bacteroides core, one dominated by L. crispatus and another consisting of a highly diverse community, including Prevotella spp., Atopobium vaginae, and Mobiluncus curtisii. Discussion. Our findings are, albeit not necessarily generalizable, consistent with the presence of a unique microbiota dominated by Bacteroides residing on the endometrium of the human non-pregnant uterus. The transcervical sampling approach may be influenced to an unknown extent by endocervical microbiota, which remain uncharacterised, and therefore warrants further validation. Nonetheless, consistent with our understanding of the human microbiome, the uterine microbiota are likely to have a previously unrecognized role in uterine physiology and human reproduction. Further study is therefore warranted to document community ecology and dynamics of the uterine microbiota, as well as the role of the uterine microbiome in health and disease.

Keywords: 16S ribosomal RNA; Bacterial vaginosis; Endometrium; Human microbiome; Illumina; Microbiota; Reproduction; Uterine microbiome; Uterus; Vaginal microbiome.

Figure 1. Endometrial bacterial community structure in subjects included (n = 19).

Endometrial bacterial community structure in each subject, showing those 60 phylotypes which exhibited an abundance of at least 1% in at least one of the samples. The colour code indicates the relative abundance of these phylotypes per sample (see colour code in the upper left corner of the figure).

Figure 2. Similarity between bacterial communities in endometrial samples (n = 19).

The dendrogram was constructed by agglomerative hierarchical clustering (group-average) based on a relative abundance matrix of phylotypes. The percentage of similarity between the communities was calculated using the Bray-Curtis similarity algorithm.