In vitro metabolic capacity of carbohydrate degradation by intestinal microbiota of adults and pre-frail elderly

Abstract

Globally increased life expectancy strongly triggered interest to delay the onset of frailty, which has been associated with alterations in compositional and functional characteristics of intestinal microbiota. In the current study, we used an in vitro batch incubation model to compare the metabolic capacity of the faecal microbiota of adults (n = 6) versus pre-frail elderly (n = 6) to degrade various glycosidic carbohydrates, including galacto-oligosaccharides, 2'-fucosyllactose, chicory fructo-oligosaccharides and inulin, and isomalto/malto-polysaccharides. The in vitro metabolic capacity was also compared with an in vivo GOS intervention study based on the same subjects. Analysis of 16S rRNA gene sequences and metabolites revealed distinct portions of variation in overall microbiota and metabolite composition during incubation being explained by individuality of the subjects and carbon source. In addition, the age group of the subjects also had significant impact on microbiota variation, carbohydrate degradation and metabolite production. This was accompanied by elevated increase in the relative abundance of Bifidobacterium in the microbiota of adults compared to that of pre-frail elderly and significantly decreased effectiveness to degrade galacto-oligosaccharides by the latter group. Altogether, the carbohydrate degradation in elderly was different compared to adults, with some carbohydrates showing decreased degradation rates. Longer interventions periods may be required to enhance bifidobacterial abundance in the microbiota of pre-frail elderly and thereby to obtain associated prebiotic health benefits.

Fig. 1. Microbiota and metabolite variation in the dataset.

A RDA based on total metabolite data. PCoA based on B weighted UniFrac and C unweighted UniFrac distance matrices of all incubation samples. D variation in metabolites that can be explained by age group, subjects or carbohydrates at 0, 4, 10 and 24 h (E, F) Variation in microbiota composition that can be explained by age group, subjects or carbohydrates, based on (E) weighted and (F) unweighted UniFrac distance matrices at 0, 4, 10 and 24 h. Both duplicate samples were included for the analysis and demonstration. AD adult, EL elderly, PCoA principal coordinate analysis, RDA redundancy analysis.

Fig. 2. Degradation kinetics of different carbohydrates.

A GOS, B 2′-FL and C FOS. Data are expressed as fraction of residual substrate as compared to the initial concentration of oligosaccharides or 2′-FL. Concentrations per DP in initial GOS and FOS were set to 1.0. Mean ±  SD are shown (A and C). Error bar was included to demonstrate the variability between subjects. In (B) no error bar was added as individual data (per subject) was shown. Individual data for A and C are available in Figs. S9 and S10. AD adult, EL elderly, DP degree of polymerization, GOS galacto-oligosaccharides, FOS fructo-oligosaccharides, 2′-FL 2′-fucosyllactose, F fructose, G glucose, SD standard deviation.

Fig. 3. Degradation kinetics of Inulin and IMMP.

A HPAEC elution patterns of Inulin. B HPSEC patterns of IMMP before and after incubation with faecal microbiota of three adults and three elderly. AD adult, EL elderly, DP degree of polymerization, IMMP isomalto/malto-polysaccharides. Incubation lasted for 24 h. Samples were taken at 0 h (Black line), 4 h (Blue line), 10 h (Purple line) and 24 h (Brown line). As duplicate samples demonstrated very high reproducibility, hereby only the chromatography elution pattern of one sample (out of the duplicate) is shown.

Fig. 4. Comparative analysis between in vitro and in vivo [11] including both adults and elderly.

A First axis of the principal response curve showing alterations in microbial composition over time in response to GOS in vivo and in response to non-carbohydrate control (protein mix) and carbohydrates (GOS or 2′-FL) in vitro, while taking in vivo GOS intervention as reference. Both duplicate samples were included for the analysis. Genera for which the model best explained the observed variation between reference and treatments (weights > 0.05) are shown on the right side of the figure. B Relative abundance of different bacterial families (top 12, ranked based on the average relative abundance across the entire dataset) in the microbiota of six adults and six elderly. Averaged relative abundance of the duplicate samples was used here (B) for visibility. Top 12 microbial families are listed in the legend. Other families are summarized as “Other”. Each column represents the corresponding type of sample from one subject. Samples collected at 24 h were excluded from this comparative analysis as some carbohydrates were completely depleted within 10 h (see Figs. 2 and 3). AD adult, EL elderly, GOS galacto-oligosaccharides, 2′-FL 2′-fucosyllactose.