Birth mode is associated with earliest strain-conferred gut microbiome functions and immunostimulatory potential

Abstract

The rate of caesarean section delivery (CSD) is increasing worldwide. It remains unclear whether disruption of mother-to-neonate transmission of microbiota through CSD occurs and whether it affects human physiology. Here we perform metagenomic analysis of earliest gut microbial community structures and functions. We identify differences in encoded functions between microbiomes of vaginally delivered (VD) and CSD neonates. Several functional pathways are over-represented in VD neonates, including lipopolysaccharide (LPS) biosynthesis. We link these enriched functions to individual-specific strains, which are transmitted from mothers to neonates in case of VD. The stimulation of primary human immune cells with LPS isolated from early stool samples of VD neonates results in higher levels of tumour necrosis factor (TNF-α) and interleukin 18 (IL-18). Accordingly, the observed levels of TNF-α and IL-18 in neonatal blood plasma are higher after VD. Taken together, our results support that CSD disrupts mother-to-neonate transmission of specific microbial strains, linked functional repertoires and immune-stimulatory potential during a critical window for neonatal immune system priming.

Fig. 1

Curation of metagenomic data. a Schematic representation of the workflow for removal of artefacts introduced during genomic extraction or preparation of sequence libraries in the low-biomass neonatal samples. b Sample-wise bioinformatic workflow for removal of artefactual sequences from metagenomic data, extraction of taxonomic and functional profiles, and reconstruction of genomes and strain-resolved analyses. The resulting data sets used for inter-sample comparisons are highlighted in grey. mOTU, metagenomic operational taxonomic unit

Fig. 2

Maternal and neonatal gut microbiome functional profiles. a Heatmap of relative abundance of gut microbial orthologous gene groups with significant differential abundances in neonates born by vaginal delivery (VD) compared to either caesarean section delivery (CSD) or CSD with small for gestational age (SGA) status (CSD + SGA) groups and having the same direction of log₂ fold change (calculated with the R package DESeq2 ; false-discovery-rate (FDR)-adjusted P < 0.05). Colour key indicates row-wise z-scores. The six significantly enriched pathways are indicated (FDR-adjusted P < 0.05). b Spearman correlation coefficients of functional profiles of neonatal and maternal gut microbiomes in VD and CSD (±SGA) neonates. c Cumulative relative abundance of enriched pathways indicated in a for VD and CSD (±SGA) groups. b, c Comparison by Wilcoxon rank-sum test for two-group comparisons with multiple testing adjustment; *FDR-adjusted P < 0.05, **FDR-adjusted P < 0.01, ***FDR-adjusted P < 0.001; Boxplots: centre line – median, bounds – first and third quartile, whiskers <= 1.5 × interquartile range

Fig. 3

Transmission of functions by distinct microbial strains. a Taxa which were detected in gut microbiomes of mothers (diagonal line) and neonates (on postnatal day 3 (below the line) and/or day 5 (above the line), indicated by shading) in vaginal delivery (VD), caesarean section delivery (CSD) and CSD with small for gestational age (SGA) status (CSD + SGA) groups. The level of evidence of transmission is indicated by the shading colour, with darker shading for stronger evidence. A taxon without link describes a taxon that was found in the maternal samples, but not shared between mother and neonate. P based on PhyloPhlAn; S based on StrainPhlAn. Neonates C115 and C116 are twins. b Inter-population fixation indices (F_ST) comparing maternal (M) and neonatal (days 1, 3, 5) faecal samples. Phylum-level colour key is given in a. Encircled symbols highlight strains that are shared with the respective mother. c Intra-population diversity index (π). Circles and triangles represent maternal and neonatal faecal samples, respectively. d, e Relative abundance of the metagenomic operational taxonomic units (mOTU) belonging to Bacteroides dorei/vulgatus (d) and Staphylococcus epidermidis (e) in maternal faecal (M), maternal vaginal (MV) and neonatal faecal (days 1, 3, 5) samples from VD, CSD and CSD + SGA groups; *false discovery rate (FDR)-adjusted P < 0.05 in Wilcoxon rank-sum test for two-group comparisons; boxplots: centre line – median, bounds – first and third quartile, whiskers <= 1.5 x interquartile range. f, g Strain-level phylogenetic trees of B. vulgatus (f) and S. epidermidis (g); black bordered and borderless symbols represent genome reconstructions and read-based strain-level identity, respectively; genome reconstructions marked with a red asterisk are represented in h and i. h, i Genome reconstructions of B. vulgatus from neonatal faeces (C117; VD; day 3; bin P2.2.1) (h) and S. epidermidis from neonatal faeces (C112; CSD; day 5; bin P2.2) (i). Circular tracks represent: assembled contigs (1), single-nucleotide variants (black), shared between mother and neonate (red) (2), positions of strain markers (3), abundance fold-changes between VD and CSD (±SGA) neonates for functionally annotated genes in forward (4) and reverse directions (5); long spokes highlight genes affiliated with enriched pathways as depicted and colour-coded in Fig. 2a

Fig. 4

Cytokines of monocyte-derived dendritic cells after stimulation with LPS from neonatal stool and in neonatal plasma. a Lipopolysaccharide (LPS) was isolated from faecal samples collected on day 3 postpartum from neonates in groups of vaginal delivery (VD), caesarean section delivery (CSD) and CSD with small for gestational age (SGA) status (CSD + SGA), and incubated for 24 h with human monocyte-derived dendritic cells (MoDCs) isolated from a total of 12 adult donors. MoDCs were stimulated with the exact same LPS volume that was extractable from the same initial amount of faecal material from each neonate sample (Methods). Exact numbers of donors used per sample are given in the plot. Positive control: LPS isolated from E. coli overnight culture. Neonates C115 and C116 are twins. b Plasma levels of TNF-α and IL-18 in samples collected at day 3 after birth from VD and CSD (±SGA) neonates. Comparison by Wilcoxon rank-sum test with multiple testing adjustment; *false discovery rate (FDR)-adjusted P < 0.05 and **FDR-adjusted P < 0.01. Circles correspond to neonates with metagenomic data, crosses represent neonates without metagenomic data. Boxplots: centre line – median, bounds – first and third quartile, whiskers <= 1.5 × interquartile range