Strenuous Physical Training, Physical Fitness, Body Composition and Bacteroides to Prevotella Ratio in the Gut of Elderly Athletes

Abstract

Regular physical activity seems to have a positive effect on the microbiota composition of the elderly, but little is known about the added possible benefits of strenuous endurance training. To gain insight into the physiology of the elderly and to identify biomarkers associated with endurance training, we combined different omics approaches. We aimed to investigate the gut microbiome, plasma composition, body composition, cardiorespiratory fitness, and muscle strength of lifetime elderly endurance athletes (LA) age 63.5 (95% CI 61.4, 65.7), height 177.2 (95% CI 174.4, 180.1) cm, weight 77.8 (95% CI 75.1, 80.5) kg, VO2max 42.4 (95% CI 39.8, 45.0) ml.kg^-1.min^-1 (n = 13) and healthy controls age 64.9 (95% CI 62.1, 67.7), height 174.9 (95% CI 171.2, 178.6) cm, weight 83.4 (95% CI 77.1, 89.7) kg, VO2max 28.9 (95% CI 23.9, 33.9), ml.kg^-1.min^-1 (n = 9). Microbiome analysis was performed on collected stool samples further subjected to 16S rRNA gene analysis. NMR-spectroscopic analysis was applied to determine and compare selected blood plasma metabolites mostly linked to energy metabolism. The machine learning (ML) analysis discriminated subjects from the LA and CTRL groups using the joint predictors Bacteroides 1.8E + 00 (95% CI 1.1, 2.5)%, 3.8E + 00 (95% CI 2.7, 4.8)% (p = 0.002); Prevotella 1.3 (95% CI 0.28, 2.4)%, 0.1 (95% CI 0.07, 0.3)% (p = 0.02); Intestinimonas 1.3E-02 (95% CI 9.3E-03, 1.7E-02)%, 5.9E-03 (95% CI 3.9E-03, 7.9E-03)% (p = 0.002), Subdoligranulum 7.9E-02 (95% CI 2.5E-02, 1.3E-02)%, 3.2E-02 (95% CI 1.8E-02, 4.6E-02)% (p = 0.02); and the ratio of Bacteroides to Prevotella 133 (95% CI -86.2, 352), 732 (95% CI 385, 1079.3) (p = 0.03), leading to an ROC curve with AUC of 0.94. Further, random forest ML analysis identified VO2max, BMI, and the Bacteroides to Prevotella ratio as appropriate, joint predictors for discriminating between subjects from the LA and CTRL groups. Although lifelong endurance training does not bring any significant benefit regarding overall gut microbiota diversity, strenuous athletic training is associated with higher cardiorespiratory fitness, lower body fat, and some favorable gut microbiota composition, all factors associated with slowing the rate of biological aging.

Keywords: VO2max; aging; body fat; exercise; microbiome.

FIGURE 1

ROC (receiver operating characteristic) curves with an area under the ROC curve (AUC) for the RFM-L algorithm with VO2max, fat content, and BMI as joint predictors/discriminators between LA and CTRL groups. FPR, false-positive rate; LA, lifetime elderly endurance athletes; CTRL, controls; RFM-L, random forest machine-learning; TPR, true positive rate.

FIGURE 2

The set of 30 bacterial genera identified as one of the 10 most abundant bacterial genera within at least one sample within both groups of samples. Core microbiome is typical for seniors. The minimum abundance must reach at least 1% of all bacteria and in at least one sample of both groups.

FIGURE 3

Microbial composition of significantly (p < 0.05) distinct bacterial genera within the healthy control group (CTRL) and elderly athletes (LA) seniors. Selected rows are centerd; unit variance scaling is applied to rows. Imputation is used for missing value estimation. Both rows and columns are clustered using correlation distance and average linkage. Numbers represent samples visualized by heat map.

FIGURE 4

Beta diversity of analyzed samples represented by significantly altered (p < 0.05) OTUs in elderly athletes and the control group visualized by PCA. SVD with imputation is used to calculate principal components. X and Y axes show principal component 1 and principal component 2 that explain 26.2 and 14.7% of the total variance, respectively. Prediction ellipses are such that with a probability of 0.95, a new observation from the same group will fall inside the ellipse (N = 22 data points).

FIGURE 5

ROC (receiver operating characteristic) curves with an area under the ROC curve (AUC) for the RFM-L algorithm with Bacteroides, Prevotella, Intestinimonas, Subdoligranulum, and the Bacteroides to Prevotella ratio as joint predictors/discriminators between the LA and CTRL groups. FPR, false-positive rate; LA, lifetime elderly endurance athletes; CTRL, controls; RFM-L, random forest machine-learning; TPR, true positive rate.

FIGURE 6

The plot depicting the ranking of predictors by graph depth (y-axis) vs. variable importance (VIMP, x-axis). VO2max is the best variable for both criteria. The Bacteroides to Prevotella ratio is the second most important by the VIMP criterion whereas it is the fourth most important by the graph depth. Overall, the two criteria are in good agreement in ranking the predictors.