In Utero 2015

[Edie Staniszewski]

Thankyou for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Recent scientific advances suggest that early gut microbiota perturbations and the ensuing aberrant immune and metabolic maturation act as a major contributing element in the development of non-communicable diseases and obesity1. Data from epidemiological studies link aberrant gut microbiota composition and factors known to disrupt intestinal microbial colonisation in early infancy with the development of disease in later life1. Delivery by caesarean section or early exposure to antibiotics and the resultant perturbation in the establishment of the gut microbiota have been associated with the development of asthma2,3,4, inflammatory bowel disease (IBD)2,5,6, and obesity7,8,9 in both epidemiological and experimental studies. The development of effective means to reduce disease risk by modulating early host-microbe interaction has been hindered by considerable gaps in our knowledge regarding the precise timing and origin of intestinal microbial colonisation.

The presence of microbes in the healthy human placenta10,11, umbilical cord12, and meconium13,14,15 suggests that foetal microbial contact is a physiological phenomenon, but the significance of intrauterine microbes for infant gut colonisation remains to be elucidated. Both epidemiological studies and experimental data indicate, however, that foetal microbial contact may be causally related to disease risk. The maternal microbial environment16,17 and antibiotic exposure during pregnancy18 are both reportedly associated with the risk of developing asthma later in childhood. After birth, breast milk modulates the infant gut microbiota by providing factors, which selectively promote the growth of specific microbes such as bifidobacteria1,19. There are data to suggest that human mammary gland tissue harbours bacteria20 and it has been suggested, that microbes in breast milk may provide a repeating colonising inoculum to the neonatal gut21. We have previously reported a diverse microbiota in human colostrum reflecting the health and nutritional status of the mother21,22. Our current understanding of the role of intrauterine and breast milk bacteria in human gut colonisation is scarce.

The purpose of this study was to characterise the microbial populations in the placenta, amniotic fluid and colostrum and to elucidate their role as the initial inoculum for the intestinal microbiota. The study is based on maternal faeces, placenta, amniotic fluid, colostrum, meconium and infant faeces samples obtained from mothers undergoing elective caesarean section delivery and their infants. Obtaining samples from several niches from the same mother-infant pairs and selecting only infants born by sterile elective caesarean section at full term without onset of labour, rupture of membranes or signs of maternal infection enabled us to reliably compare microbiota composition in these maternal compartments and correlate them with meconium and neonatal gut microbiota. The study design also allowed us to exclude the impact of bacterial transfer during vaginal delivery. We report here for the first time direct evidence suggesting that distinct amniotic fluid, placenta and colostrum microbiota contribute to perinatal human gut colonisation.

The microbial populations in amniotic fluid and placenta observed in this study were similar and highly consistent across individuals (Fig. 1A) with low abundance (Fig. 1E), low richness and low diversity (Fig. 1F,G). Consistently with a previous report11, Proteobacteria was the most prevalent phylum in amniotic fluid and placenta samples (Fig. 1A) with particularly high abundance of species belonging to Enterobacteriaceae (Fig. 1B). Enterobacter and Escherichia/Shigella were the most predominant Proteobacteria genera in amniotic fluid and placenta samples. These genera were also present in colostrum, meconium and infant faeces but in lower abundance. Propionibacterium was the second most predominant genus present in the amniotic fluid and placenta, and also detectable in meconium. Streptococcus genus was present in low abundance in the amniotic fluid, placenta and meconium (

The microbiota observed in colostrum samples was more varied with higher richness and diversity (Fig. 1A,B,F,G). Interestingly, sequences belonging to Proteobacteria and particularly Enterobacteriaceae were frequent also in colostrum (Fig. 1A,B). A relatively large number of unassigned sequences were detected in colostrum samples as well as in placenta samples by 16S rRNA gene pyrosequencing (Fig. 1A).

The microbiological findings from meconium samples obtained in this study shared features with the amniotic fluid and placenta samples. Of the 75 bacterial family level phylotypes detected in the meconium samples, 41 were also detected in both amniotic fluid and placenta samples and an additional 15 in amniotic fluid samples alone (Fig. 3A). Identification of microbes on the genus level revealed several microbes that were detected in amniotic fluid and placenta as well as in meconium (Supplementary Table 1). The shared genus level phylotypes between different sample types are presented in Supplementary Table 3. A core microbiome consisting of OTUs detected in 50% of samples from each compartment was found to include Streptococcus, unclassified Enterobacteriaceae (mostly Enterobacter and Escherichia), Propionibacterium, Lactobacillus and unclassified Bacillales (Supplementary Table 4). Bacteria belonging to the genera Escherichia/Shigella, Lactobacillus and Propionibacterium were present in all placenta, amniotic fluid and meconium samples. Moreover, we found specific de novo OTUs assigned to Bifidobacterium, Nitrobacter and Clostridium, which were consistently detected in the placenta, amniotic fluid and meconium samples from specific individual mother-infant pairs. Clear phylogenetic clustering of some of the meconium samples with amniotic fluid and placenta samples was also observed (Fig. 3B). Furthermore, specific bacteria present in amniotic fluid and placenta were also detected in the meconium samples by DGGE (Fig. 3C).

Despite these similarities between amniotic fluid, placenta and meconium microbiota, the overall microbial composition and activity in meconium appears unique. Significant differences were observed between the activity of the amniotic fluid and meconium microbiota (Fig. 2B). In addition, the LDA Effect Size (LEfSe: Linear Discriminant Analysis Effect Size) algorithm used to identify taxa with differing abundance in amniotic fluid and meconium samples revealed that while Enterobacteriaceae are the predominant bacteria in the placenta, meconium samples differ from the placenta particularly with regard to high numbers of Bacillaceae and Streptococcaceae (Fig. 4A,B). On the phylum level, the meconium microbiota is dominated by Firmicutes (Figs 1A and 3D) and Staphylococcaceae was the most frequent bacterial family detected in meconium samples (Fig. 1B). These findings are consistent with a previous report on microbial composition of meconium in preterm infants24.

(A) Bacterial taxa that were differentially abundant in placenta and meconium samples visualised using a cladogram generated from LEfSe analysis. (B) Differences in key OTUs identified as differentiating between placenta and meconium samples. (C) Bacterial taxa that were differentially abundant in placenta and colostrum samples visualized using a cladogram generated from LEfSe analysis. (D) Differences in key OTUs identified as differentiating between placenta and colostrum samples.

Our data demonstrate that human amniotic fluid and placenta harbour unique microbial communities, which may provide the initial inoculum for gut colonisation, the single most important determinant of host-microbe interaction modulating the risk of non-communicable disease. The distinct microbial populations in placenta, amniotic fluid and colostrum and the consistency of the findings across individuals observed in this study suggest that there may be active and selective mechanisms by which bacteria are transported to these maternal compartments.

The origin of the intrauterine microbiota is currently not known. Similarities between human oral and placenta microbial communities have previously been suggested11 but the study relied on oral microbiota data from a previous report based on non-pregnant individuals and the mechanism of possible bacterial transport between these confined maternal compartments remains unknown. In contrast, there are mechanistic data consistent with the notion that both breast milk and intrauterine microbes may originate in the maternal gut. We have previously reported that the composition of the intestinal microbiota changes dramatically during pregnancy in humans25 and increased intestinal bacterial translocation has been reported in experimental animals during pregnancy and lactation26. Specific labelled bacteria introduced to the gut of pregnant mice have been detected in the placenta in an experimental animal model12. There are data indicating that maternal intestinal microbes may be actively transported to breast milk by immune cells in the systemic circulation in both lactating mice and humans26. Consequently, we hypothesise that maternal intestinal microbes may be selectively transported to the mammary gland and to the foeto-placental interface.

Microbes in the amniotic cavity have previously been investigated primarily as potential pathogens causing infection or premature labour, but recently the dogma of sterile foetal life has been challenged1. Several investigators have reported the presence of microbes in meconium13,14,15 but their origin has not been known. Recently, the bacteria in meconium have been reported to reflect maternal health27 and also to be associated with the development of disease in the offspring15. Combining observations from separate studies, it has been speculated that the meconium microbiota may be derived from swallowed amniotic fluid28. We present here for the first time comprehensive analyses of the microbiota in the placenta, amniotic fluid and meconium from the same mother-infant pairs. Only infants born by elective caesarean section were included in this study. While it is important to recognize that this population does not reflect the physiological continuum of microbial contact during pregnancy, vaginal delivery and breastfeeding, the study design allowed us to exclude the impact of bacterial transfer during labour and delivery. Mothers, who presented with rupture of membranes, labour, or signs of infection were excluded from the study to minimise contamination of the placenta and amniotic fluid samples. Contamination of the meconium samples with environmental microbes cannot be ruled out since the specimens were collected noninvasively from diapers, but the procedures were designed to avoid contamination as much as possible. Based on detection of both live microbes and microbial DNA, the intrauterine compartment appears to harbour a unique microbiota with a distinct composition and activity. Shared features and specific bacteria as well as clustering of microbial findings in meconium on the one hand and placenta and amniotic fluid on the other suggest that colonisation of the foetal intestine may indeed be initiated in utero by microbes in the placenta and amniotic fluid. This notion is consistent with data from an experimental animal model according to which specific bacteria introduced to the gut of pregnant animals may be recovered in the meconium of the offspring after sterile caesarean section29.

3a8082e126