Cardiorespiratory fitness as a predictor of intestinal microbial diversity and distinct metagenomic functions

Abstract

Background: Reduced microbial diversity in human intestines has been implicated in various conditions such as diabetes, colorectal cancer, and inflammatory bowel disease. The role of physical fitness in the context of human intestinal microbiota is currently not known. We used high-throughput sequencing to analyze fecal microbiota of 39 healthy participants with similar age, BMI, and diets but with varying cardiorespiratory fitness levels. Fecal short-chain fatty acids were analyzed using gas chromatography.

Results: We showed that peak oxygen uptake (VO2peak), the gold standard measure of cardiorespiratory fitness, can account for more than 20 % of the variation in taxonomic richness, after accounting for all other factors, including diet. While VO2peak did not explain variation in beta diversity, it did play a significant role in explaining variation in the microbiomes' predicted metagenomic functions, aligning positively with genes related to bacterial chemotaxis, motility, and fatty acid biosynthesis. These predicted functions were supported by measured increases in production of fecal butyrate, a short-chain fatty acid associated with improved gut health, amongst physically fit participants. We also identified increased abundances of key butyrate-producing taxa (Clostridiales, Roseburia, Lachnospiraceae, and Erysipelotrichaceae) amongst these individuals, which likely contributed to the observed increases in butyrate levels.

Conclusions: Results from this study show that cardiorespiratory fitness is correlated with increased microbial diversity in healthy humans and that the associated changes are anchored around a set of functional cores rather than specific taxa. The microbial profiles of fit individuals favor the production of butyrate. As increased microbiota diversity and butyrate production is associated with overall host health, our findings warrant the use of exercise prescription as an adjuvant therapy in combating dysbiosis-associated diseases.

Keywords: Butyrate; Community diversity; Dysbiosis; Exercise; Intestinal microbiota; Metagenome; Microbial ecology; Physical activity.

Fig. 1

Dietary patterns amongst fitness groups. Scores of the two first components of the PCA of dietary data for all 39 subjects are presented. Each circle represents one participant, colored based on their CRF fitness levels. A lack of distinct clustering amongst groups suggests comparable dietary patterns amongst groups

Fig. 2

Correlation between VO₂peak and species richness (SR). Result of a multiple regression model showing a significant association between VO₂peak and SR when holding all other variables constant. Shaded area represent 95 % confidence intervals

Fig. 3

Beta diversity amongst fitness groups. PCoA plot of genus abundance data based on Bray-Curtis dissimilarity measure shows no clear clustering when grouped according to CRF levels

Fig. 4

Bacterial abundance RDA correlation biplots constrained by selected explanatory variables. The sites and explanatory variables (a) and genera (b) plots are presented separately for clarity; however, they are derived from the same RDA model, note the difference in axes scales. RDA1 and RDA2 which explain over 10 % of total variation in beta diversity are plotted. The global model’s P value was calculated using the Monte Carlo Permutation Procedure (MCPP). In plot A, subjects are color coded according to their CRF levels for illustrative purposes only as groupings were not included in the model. Black circles represent centroids for the categorical variable sex

Fig. 5

RDA correlation biplots of predicted metagenomics functions constrained by selected explanatory variables. The sites and explanatory variables (a) and genera (b) plots are presented separately for clarity; however, they are derived from the same RDA model, note the difference in axes scales. RDA1 and RDA2 which explain over 13 % of the total variation in data are plotted. The global model’s P value was calculated using the Monte Carlo Permutation Procedure (MCPP). In plot A, subjects are color coded according to their CRF for illustrative purposes only as groupings were not included in the model. Black circles represent centroids for the categorical variable sex

Fig. 6

Correlation between VO₂peak and fatty acid biosynthesis. Spearman correlation plot showing a positive relationship between VO₂peak and the functional category “fatty acid biosynthesis.” rho Spearman’s correlation coefficient

Fig. 7

RDA correlation triplot of SCFA abundance data constrained by selected explanatory variables. RDA1 and RDA2 which explain over 29 % of the total variation in SCFA data are plotted. Subjects are color coded according to their CRF for illustrative purposes only as groupings were not included in the model. Black circles represent centroids for the categorical variable sex. The global model’s P value was calculated using the Monte Carlo Permutation Procedure (MCPP)