Studying Vertical Microbiome Transmission from Mothers to Infants by Strain-Level Metagenomic Profiling

Abstract

The gut microbiome becomes shaped in the first days of life and continues to increase its diversity during the first months. Links between the configuration of the infant gut microbiome and infant health are being shown, but a comprehensive strain-level assessment of microbes vertically transmitted from mother to infant is still missing. We collected fecal and breast milk samples from multiple mother-infant pairs during the first year of life and applied shotgun metagenomic sequencing followed by computational strain-level profiling. We observed that several specific strains, including those of Bifidobacterium bifidum, Coprococcus comes, and Ruminococcus bromii, were present in samples from the same mother-infant pair, while being clearly distinct from those carried by other pairs, which is indicative of vertical transmission. We further applied metatranscriptomics to study the in vivo gene expression of vertically transmitted microbes and found that transmitted strains of Bacteroides and Bifidobacterium species were transcriptionally active in the guts of both adult and infant. By combining longitudinal microbiome sampling and newly developed computational tools for strain-level microbiome analysis, we demonstrated that it is possible to track the vertical transmission of microbial strains from mother to infants and to characterize their transcriptional activity. Our work provides the foundation for larger-scale surveys to identify the routes of vertical microbial transmission and its influence on postinfancy microbiome development. IMPORTANCE Early infant exposure is important in the acquisition and ultimate development of a healthy infant microbiome. There is increasing support for the idea that the maternal microbial reservoir is a key route of microbial transmission, and yet much is inferred from the observation of shared species in mother and infant. The presence of common species, per se, does not necessarily equate to vertical transmission, as species exhibit considerable strain heterogeneity. It is therefore imperative to assess whether shared microbes belong to the same genetic variant (i.e., strain) to support the hypothesis of vertical transmission. Here we demonstrate the potential of shotgun metagenomics and strain-level profiling to identify vertical transmission events. Combining these data with metatranscriptomics, we show that it is possible not only to identify and track the fate of microbes in the early infant microbiome but also to investigate the actively transcribing members of the community. These approaches will ultimately provide important insights into the acquisition, development, and community dynamics of the infant microbiome.

Keywords: infant microbiome; metagenomics; microbial ecology; microbial genomics; vertical microbiome transmission.

FIG 1

Microbial composition of mother and infant samples and shared bacteria within mother-infant pairs. (A) Quantitative microbial taxonomic composition of the metagenomic samples from milk and fecal samples of mothers and infants as estimated by MetaPhlAn2 analysis (77) (only the 20 most abundant species are indicated). Milk samples present low microbial richness compared to fecal samples. (B) Ordination plot of microbiome composition showing clustering of the three different sample types: mother feces, infant feces, and breast milk samples. The two infant samples close to the cluster of mother feces and in between the clusters of mothers and infants are from later time points, denoting the convergence of the infant microbiome toward an adult-like one. (C) The abundances of the 10 microbial species detected (>0.1% abundance) in at least one infant and the respective mother (shared species have been identified on the basis of samples from time point 1 [T1] only).

FIG 2

Strain-level phylogenetic trees for microbes present in both the mother and infant. Phylogenetic trees were built by the StrainPhlAn method using species-specific markers confirming the presence of the same strain in the mother and infant intestinal microbiomes, thus suggesting vertical transmission. Available reference genomes were included in the phylogenetic trees. Here we report three bacterial species, namely, (A) Bifidobacterium bifidum, (B) Coprococcus comes, and (C) Ruminococcus bromii, and the most abundant viral species found in pair 4, (D) pepper mild mottle virus. Other species-specific phylogenetic trees (B. adolescentis, B. breve, and B. longum) are reported in Fig. S3.

FIG 3

Functional potential analyses. (A) HUMAnN2 heat map reporting the 25 most abundant pathways in the fecal samples of mothers and infants. Specific pathways of interest (sugars, mucin, and folate metabolism) are added at the bottom. The asterisk (*) near the heat map highlights statistically significant pathways. (B) Multidimensional scaling (MDS) result from functional potential profiles, showing the differences between fecal samples of mothers and infants and milk samples. In particular, the infant feces point in the mother feces cluster corresponds to time point 3 of pair 5, showing a shift from the infant microbiome toward an adult-like microbiome. (C) HUMAnN2 results for the 25 most abundant pathways found only in the milk samples. TCA, tricarboxylic acid.

FIG 4

Transcription levels of metabolic pathways and genes in mother and infant pair 4 at time point 2. (A) Scatterplots showing the transcription rates of metabolic pathways of shared and nonshared species and genera of interest for both the mother and infant of pair 4 at time point 2. (B) Comparison between transcription rates of gene families in mother and infant gut microbiomes.