The influence of early life exposures on the infant gut virome

Abstract

The factors influencing the establishment of the gut bacterial community in early life are fairly well studied. However, the factors shaping the infant gut virome remain elusive. Interestingly, early life gut virome imbalances have recently been linked with increased risk of developing diseases like type 1 diabetes and asthma. We utilized the deeply phenotyped COPSAC2010 cohort to investigate how environmental factors influence the gut virome at one year age. We demonstrate that the presence of older siblings as well as residential location (urban or rural) had the strongest impact on gut virome composition at 1 year of age. A total of 16,118 species-level clustered viral representative contigs (here termed viral Operational Taxonomic Units - vOTUs) were identified and of these 2105 vOTUs varied in abundance with environmental exposures. Of these vOTUs 94.1% were phages mainly predicted to infect Bacteroidaceae, Prevotellaceae, and Ruminococcaceae. Strong co-abundance of phages and their bacterial hosts was confirmed underlining the predicted phage-host connections. Furthermore, we found some gut viruses affected by environmental factors encode enzymes involved in the utilization and degradation of major dietary components, potentially affecting infant health by influencing the bacterial host metabolic capacity. These findings provide a valuable insights for understanding the early life factors that predispose to autoimmune and metabolic disorders.

Keywords: Environmental exposures; bacterial host; gut microbiota; infant; metabolism; phage; virus.

Figure 1.

Virome structure of the infants enrolled in the COPSAC₂₀₁₀ cohort. (a) Distribution of the 16,118 vOTUs identified colored by their taxonomic class annotation. (b) Cumulative frequency of viral genomes (kb) identified by their taxonomic class annotation. (c) Circular diagram showing the distribution of vOTUs colored by their targeted bacterial hosts (at phylum and family levels), viral class and lifestyle. (d,e) t-Stochastic Neighbor Embedding (t-SNE) plots clustering tetra-mer vOTUs profiles identified according to host family (d) and viral class (e). (f–h) Percentage of vOTUs that appear at a specific prevalence (f), and vOtus’ distribution colored by their taxonomic class (g) and host family (h). (i) Relative abundance of vOTUs across all samples at the class level. Samples were sorted by Malgrandaviricetesabundance.

Figure 2.

Virome diversity and composition covariates with early life exposures. (a–d) Barplot showing the strength of associations (-log10 p-value) of the alpha diversity metrics observed vOTUs (a) and Shannon index (b) across different exposures (linear mixed model) as well as beta diversity using distance-based redundancy analysis (db-RDA) on bray-curtis dissimilarity (c) and sorensen-dice distance (d) matrices. Statistical significance levels are represented as p < 0.001 (***), p < 0.01 (**) and p ≤ 0.05 (*). P-values were adjusted by Benjamini – Hochberg method and shown as q-values (only shown for significant exposures). (e,f) distribution of observed vOTUs for weight at birth and siblings (e) and Shannon index for siblings and dietary introduction of egg (f). Dietary introduction of egg is indicated in days. (g,h) db-RDA constrained-components based on bray-curtis distances for location and siblings (g), and sorensen-dice distances for weight at birth and dietary introduction of fish (h).

Figure 3.

Viral host family, relative abundance and lifestyle associate with environmental exposures at one year of age. Visualization of differential abundance analysis of 2105 vOTUs across the nine exposures significantly associated with virome diversity and composition. Log₂ Fold change panel displays the change in abundance between the two groups for each exposure. The viral families to which vOTU belongs, surrounded by red boxes, are labeled. Adjusted p ≤ 0.001 and Log₂-Fold changes ≥ |1| were used to select differentially abundant vOTUs.

Figure 4.

Abundance of phage accessory genes differ in connection with exposures. (a) Abundance of genes (3^rd level KEGG pathway) in the virome of infants with significant (p ≤ 0.05) enzymatic enrichments that are associated with the presence of siblings and residential location. (b) Viral host families that contribute to metabolism pathways. the outermost circle represents the vOTU class. The second circle represents the viral lifestyle (virulent vs. temperate). In the innermost circle, red represents metabolic pathways and green represents the vOtus. The pathway and the family of the vOTU host are labeled. (c–e) Bacterial points extracted from the procrustes analysis (e). The points are colored according to the abundance of the specific genus in each sample.