Mother-to-Infant Microbial Transmission from Different Body Sites Shapes the Developing Infant Gut Microbiome

Abstract

The acquisition and development of the infant microbiome are key to establishing a healthy host-microbiome symbiosis. The maternal microbial reservoir is thought to play a crucial role in this process. However, the source and transmission routes of the infant pioneering microbes are poorly understood. To address this, we longitudinally sampled the microbiome of 25 mother-infant pairs across multiple body sites from birth up to 4 months postpartum. Strain-level metagenomic profiling showed a rapid influx of microbes at birth followed by strong selection during the first few days of life. Maternal skin and vaginal strains colonize only transiently, and the infant continues to acquire microbes from distinct maternal sources after birth. Maternal gut strains proved more persistent in the infant gut and ecologically better adapted than those acquired from other sources. Together, these data describe the mother-to-infant microbiome transmission routes that are integral in the development of the infant microbiome.

Keywords: infant microbiome; microbiome transmission; shotgun metagenomics; strain-level profiling.

Graphical abstract

Figure 1

Longitudinal Metagenomic Sequencing of the Microbiome of Mother-Infant Pairs (A) Samples were collected from 25 mother-infant pairs and metagenomically sequenced. Samples were taken from the stool (FE), skin (SK), vagina (VA), and tongue dorsum (TD) of the mothers and from the stool and tongue dorsum of the infants. Sampling of the infant started within 24 hr from delivery and continued for up to 4 months (STAR Methods). All samples were shotgun sequenced and the average depth (in Gbases) of the quality-controlled and human DNA-free samples are reported. (B) Alpha-diversity distributions for each sample type and time point (^∗p < 0.05, ^∗∗∗p < 0.001). (C) Ordination plot (MDS) of all the samples that passed preprocessing based on the Bray-Curtis distance between samples highlights the spatial clustering of samples with respect to both different body sites and longitudinal time points. (D) Beta diversities (Bray-Curtis on log-scaled relative abundances) between samples within each infant body site (gut and tongue dorsum) across time points.

Figure 2

Microbial Species Common to the Mothers and Their Infants (A and B) Average taxonomic composition of the infant stool (A) and tongue dorsum (B) microbiomes over time. The colored sectors indicate species that are found in the infant and his/her mother. White portions refer to species not found in the maternal body sites. External rings show the cumulative abundance of bacterial species per maternal body site. (C) Relative abundances of the most abundant vaginal bacteria in the mothers and in the gut of their infants. Each line represents a mother-infant pair. (D) Number, percentage over total number, and cumulative abundance of identified microbial species that are shared between each mother and her own infant (intra-pair), and between each mother and unrelated infants (inter-pair) (^∗p < 0.05 and ^∗∗p < 0.01, t test).

Figure 3

Strain Transmission between Mothers and Their Infants (A and B) The distribution of the normalized strain intra-pair distances (STAR Methods) (A) and the number of vertical transmitted strains for each maternal source and each infant recipient body site and time point (B). (C) Escherichia coli strain-specific gene content as identified by PanPhlAn. Strains with clear evidence of vertical transmission are indicated with boxes. (D and E) Mother-infant phylogenies for Bacteroides uniformis and Bacteroides vulgatus as inferred by StrainPhlAn. Maternal body sites are represented by squares and infant body sites by circles. Mother-infant pairs with at least two samples are labeled with the pair ID in the trees (black circles otherwise).

Figure 4

Strain Persistence, Strain Replacement Events, and Strain Heterogeneity (A) Map of the strain dynamics in longitudinal infant stool (FE) samples for selected species (for full map, see Figure S6). The tongue dorsum (TD) column shows the species for which at least one of the strains found in stool was also present on the tongue dorsum. Blue circles represent the first strain of the species identified in the infant, whereas orange and green circles denote the second and third longitudinally identified strain, respectively. Empty circles refer to species for which strain profiling was not possible in the specific sample. Missing samples and samples lacking the species are not reported. The total number of infant replacement events observed in each species is shown in parentheses. (B) Mean percentages of polymorphic sites and average frequency of the dominant alleles in polymorphic sites for each body site and time point (“M” indicates maternal samples). Color coding is as per Figure 1. p values are reported in Table S4. Error bars refer to 95% confidence intervals.

Figure 5

Infants Have More Shared Species and Strains between the Oral and Gut Microbiome Than Their Mothers (A) Number of shared species normalized by the total number of species present in the gut and the tongue dorsum of each subject, and the cumulative abundances of shared species in the two body sites. The number of samples considered for the analysis is reported in parentheses. (B and C) Transmission trees for Streptococcus salivarius (B) and Rothia mucilaginosa (C). Only pairs with more than two samples present in the tree are shown (black circles otherwise).

Figure 6

Phylogenetic Placement of 1,132 Metagenomically Reconstructed Genomes and Mother-to-Infant Transmission of Taxonomically Uncharacterized Strains (A) We used PhyloPhlAn2 (Segata et al., 2013) to place the genomes reconstructed with metaSPAdes (Nurk et al., 2017) and binned with MetaBAT2 (Kang et al., 2015) (STAR Methods) on the microbial “Tree of Life” (Ciccarelli et al., 2006, Segata et al., 2013), which encompasses 4,000 species with available reference genomes. Leaf nodes without circles refer to reference genomes from known species, white circles indicate reconstructed genomes that are close (>95% identity) to a known species, and red circles show reconstructed genomes that cannot be assigned (<95% identity) to known species. The eight events of mother-to-infant transmission of strains from species yet to be described are called out on the top right, and the external ring of the phylogeny reports the percent identity of each leaf node against the closest genomes from known species (values below 95% are shown in red). (B) The reconstructed genomes with completeness >50% from each body site are plotted with the corresponding completeness and genome size.