Functional metagenomic profiling of intestinal microbiome in extreme ageing

Abstract

Age-related alterations in human gut microbiota composition have been thoroughly described, but a detailed functional description of the intestinal bacterial coding capacity is still missing. In order to elucidate the contribution of the gut metagenome to the complex mosaic of human longevity, we applied shotgun sequencing to total fecal bacterial DNA in a selection of samples belonging to a well-characterized human ageing cohort. The age-related trajectory of the human gut microbiome was characterized by loss of genes for shortchain fatty acid production and an overall decrease in the saccharolytic potential, while proteolytic functions were more abundant than in the intestinal metagenome of younger adults. This altered functional profile was associated with a relevant enrichment in "pathobionts", i.e. opportunistic pro-inflammatory bacteria generally present in the adult gut ecosystem in low numbers. Finally, as a signature for long life we identified 116 microbial genes that significantly correlated with ageing. Collectively, our data emphasize the relationship between intestinal bacteria and human metabolism, by detailing the modifications in the gut microbiota as a consequence of and/or promoter of the physiological changes occurring in the human host upon ageing.

Figure 1. Centenarians and elderly differ in both microbiota composition (a) and urine metabotypes (b)

a, Principal coordinates analysis (PCoa) of Euclidean distances between HITChip profiles [8]. PC1 represents 36.46% of the variability, PC2 19.91%. b, Two centenarian subjects versus elderly group; R² = 0.394, Q² = 0.596, two-component model. The ellipses represent the Hotelling's T2 with 95% confidence. Subjects are colored according to the age groups: red for centenarians, blue for elderly people, green for the young adult.

Figure 2. Taxonomic fingerprint of ageing

Genus-node fingerprints were obtained using Euclidean PCoA. Each circle represents a bacterial genus colored on the basis of the phylum classification. Circle diameter is proportional to the average relative abundance of the bacterial genus in the corresponding age group.

Figure 3. Metagenome function analysis separates centenarians from the other subjects in agreement with genus and urine metabolite clustering

a, Hierarchical Ward-linkage clustering based on the Pearson correlation coefficients of the abundance of KO genes, filtered for KO gene subject presence ≥ 1 in at least 8/9 subjects. KO genes are clustered in the vertical tree and color-coded according to the first level of KEGG classification or the second level for functions concerning metabolism. 2719 KO genes confidently classified in the KEGG database are visualized. The bottom panel shows the relative abundance of the KEGG categories. b, Procrustes analysis combining Euclidean PCoA of functional microbiota (non-circle end of lines) with either Euclidean PCoA based on the genus dataset (circle-end of lines; upper graph) [8], or Euclidean PCoA based on the spectra of urine metabolites (circle-end of lines; lower graph) [22]. In both graphs color codes are per age group as in Figure 1.

Figure 4. Age-related trajectory of gut microbiome functions

a, PCoA of Euclidean distances between KEGG KO profiles. Subjects are colored as in Figure 1. b, The position of each fecal microbiome along PC1, which described the largest amount of variation (82.5%) in the dataset, was plotted against age. Each circle is a subject colored as in Figure 1. PC1 was significantly related with age. c, Average (± SEM, error bar) PC1 coordinates obtained for functional cluster of KO genes, corresponding to KEGG pathways, used for building the PCoA. In this way pathways are ordered for: (1) negative value of PC1 coinciding with high concordance with ageing profile; (2) value of PC1 close to 0 coinciding with constant presence in the dataset; (3) positive value of PC1 coinciding with high concordance with the healthy adult profile.