Diversity, compositional and functional differences between gut microbiota of children and adults

Abstract

The gut microbiota has been shown to play diverse roles in human health and disease although the underlying mechanisms have not yet been fully elucidated. Large cohort studies can provide further understanding into inter-individual differences, with more precise characterization of the pathways by which the gut microbiota influences human physiology and disease processes. Here, we aimed to profile the stool microbiome of children and adults from two population-based cohort studies, comprising 2,111 children in the age-range of 9 to 12 years (the Generation R Study) and 1,427 adult individuals in the range of 46 to 88 years of age (the Rotterdam Study). For the two cohorts, 16S rRNA gene profile datasets derived from the Dutch population were generated. The comparison of the two cohorts showed that children had significantly lower gut microbiome diversity. Furthermore, we observed higher relative abundances of genus Bacteroides in children and higher relative abundances of genus Blautia in adults. Predicted functional metagenome analysis showed an overrepresentation of the glycan degradation pathways, riboflavin (vitamin B2), pyridoxine (vitamin B6) and folate (vitamin B9) biosynthesis pathways in children. In contrast, the gut microbiome of adults showed higher abundances of carbohydrate metabolism pathways, beta-lactam resistance, thiamine (vitamin B1) and pantothenic (vitamin B5) biosynthesis pathways. A predominance of catabolic pathways in children (valine, leucine and isoleucine degradation) as compared to biosynthetic pathways in adults (valine, leucine and isoleucine biosynthesis) suggests a functional microbiome switch to the latter in adult individuals. Overall, we identified compositional and functional differences in gut microbiome between children and adults in a population-based setting. These microbiome profiles can serve as reference for future studies on specific human disease susceptibility in childhood, adulthood and specific diseased populations.

Figure 1

Effect of ambient temperature on individual OTUs. Regression analysis of individual OTUs with time in mail (TIM) for samples in GenR (A) and RS (B). At each TIM. the initial OTU table was sub-sampled to contain only samples up to that TIM. Red bars indicate bacteria that decreased in abundance and green bars bacteria that increased upon increasing TIM. Q-values are indicated; only significantly abundant OTUs are presented.

Figure 2

Characteristics of the final datasets of the two cohorts. GenR (A) and RS (B). Number of observed taxa at each taxonomy level Top: indicates the number of unique OTUs identified in each taxonomic clade, top A: RS cohort, top B: GenR cohort. Bottom: Donut plots indicate the average relative abundances of the top major phyla in each cohort. Donut plots of the COPSAC cohort (children aged 6 years) and doughnut plots of FGFP and LLD cohorts (adults) are plotted for comparison with the abundance in GenR and RS.

Figure 3

Comparison of the gut microbiome diversity and composition between adults (RS) and children (GenR). (A) boxplots of the Shannon diversity Index. (B) ordination plot of the gut microbiome composition in the two cohorts based on Bray-Curtis dissimilarities. The centroid and dispersion of each cohort is represented by the cohort name and ellipses, respectively. Clustering of RS and GenR was tested for significance using PERMANOVA. (C) Circular representation of the taxonomic tree of the microbiome compositions of the two cohorts. Each node represents one taxon at different taxonomic level. Orange nodes are the taxa that were observed with higher abundance in the GenR cohort and green nodes represent the taxa that were higher abundant in the RS cohort. (D) The genera represented the most in each cohort. On the x-axis the arcsine squared root transformed coefficients of the most significantly abundant genera in each cohort are shown. Orange bars represent GenR and green bars represent RS. Minus signs in the x-axis are used only for visualization.

Figure 4

Predicted functional composition of metagenomes based on 16S rRNA gene sequencing data from GenR and RS cohorts. LEfSe based on the PICRUSt dataset revealed differentially enriched metabolic pathways associated with GenR (orange) or RS (green).