Effects of exercise frequency on the gut microbiota in elderly individuals

Abstract Growing evidence has shown that exercise can affect the gut microbiota. The effects of exercise frequency on the gut microbiota in elderly individuals are still largely unknown. In the present study, samples from 897 elderly and 1,589 adult individuals (18–60 years old) from the American Gut Project were screened. Microbial diversity and composition were analyzed by QIIME2, and microbial function was predicted by PICRUSt2. The outcomes were further analyzed by STAMP. The analysis showed that the α‐diversity of gut microbiota increased with increasing age, and regular exercise reshaped the alterations in microbial composition and function induced by aging. Moreover, the α‐diversity of gut microbiota was higher in overweight elderly individuals than in normoweight elderly individuals, and regular exercise significantly affected the microbial composition and function in overweight elderly individuals. In conclusion, we revealed that regular exercise benefits elderly individuals, especially overweight elderly individuals, by modulating the gut microbiota.


| INTRODUC TI ON
In the intestine, hundreds of millions of bacteria play essential roles in host health (Bäckhed et al., 2004). The gut microbiota is composed of various microorganisms that form a complex ecological balance.
The complexity and diversification of the gut microbiota benefit the host in many ways, such as by providing resistance against potential pathogens and increasing immunity (Round & Mazmanian, 2009).
Metabolites produced by bacteria mediate host nutrition absorption and health. Thus, an imbalance in the gut microbiota could result in various disease pathogenicities, such as obesity and diabetes mellitus (Tremaroli & Bäckhed, 2012). The gut microbiota can be regulated by a range of factors, such as diet, antibiotic use, diseases, and exercise (Nicholson et al., 2012). Therefore, the gut microbiota is becoming a novel target for disease therapies.
Exercise is an important means to delay aging and prevent and manage diseases (Biddle & Batterham, 2015). Exercise can effectively reduce the risks of heart disease, stroke, hypertension, diabetes, cancer, and osteoporosis and improve depression (Biddle & Batterham, 2015). Exercise is beneficial to health by reducing body mass, the body mass index (BMI), the fat mass percentage, the fasting glucose level, and the fasting insulin level (Biddle & Batterham, 2015). Accumulating evidence shows that exercise affects the gut microbial composition, which might play a positive role in energy regulation (Mach & Fuster-Botella, 2017;Monda et al., 2017). Animal studies have shown that exercise improves the diversity of the gut microbiome and changes the composition of bacteria in the gut (Brandt et al., 2018;Feng et al., 2017).
Studies show that exercise could change the microbial composition at the phylum level, affecting the abundance of phyla such as Firmicutes, Proteobacteria, Prevotella, and Cyanobacteria (Lambert et al., 2015;Liu et al., 2015). The families Christensenellaceae and Coriobacteriaceae have shown increased abundance, suggesting a link between exercise and health improvement (Liu et al., 2015;Zhao et al., 2018). Furthermore, exercise training increases the production of short-chain fatty acids (Feng et al., 2017). In humans, following an exercise challenge, the gut microbiota diversity and microbial composition are altered (Taniguchi et al., 2018;Whisner, Maldonado, Dente, Krajmalnik-Brown, & Bruening, 2018). Similar results have been shown in athletes. Athletes have increased microbial diversity, which might be correlated with metabolic improvement and an inflammation reduction Clarke et al., 2014). Thus, the exercise-gut microbiota axis might play an important role in maintaining health.
Overweight (BMI > 25) is becoming a global health problem.
Exercise helps regulate body weight, as it is inversely associated with weight gain and contributes to weight loss (Jakicic, Rogers, Davis, & Collins, 2018). In high-fat diet (HFD)-induced obese mice, exercise counteracted the microbial imbalance, which was distinct from dietary effects, protected the intestinal barrier, and improved bile acid homeostasis (Carbajo-Pescador et al., 2019;Evans et al., 2014).
The gut microbiota undergoes aging-related changes that may affect health. An early study showed that γ-proteobacteria were enriched, while Ruminococcus obeum and its closely related phylotypes were hardly detected in elderly individuals (Hayashi, Sakamoto, Kitahara, & Benno, 2003). Studies have shown that the microbial community structure of elderly individuals is significantly different from that of young individuals (Claesson et al., 2011). Shen et al. (2018 found that the microbial changes caused by aging were characterized by reduced Bacteroidetes abundance. Ogawa, T., et al. suggested that adequate exercise was important for maintaining health because it manipulates the gut microbiota in elderly individuals (Shimizu, 2018). However, the effects of exercise frequency on the gut microbiota of elderly individuals remain largely unknown. The present study was carried out to characterize regular exercise-induced changes in the properties of the gut microbiota in elderly individuals and to further analyze regular exercise-induced changes that benefit overweight elderly (OE) individuals.

| Data sources
The data for this study were obtained from the American Gut Project (AGP) (McDonald et al., 2018

| FASTQ format conversion
The SRA files were downloaded to a computer, and the sratoolkit tools were used to convert the SRA data into the fastq format. The fastq-dump.exe command was used to convert the format.

| Data processing
QIIME version 2 was utilized to process the data (Brandt et al., 2018).
First, the Deblur plugin was used for sequence quality control. The operational taxonomic units (OTUs) resulting from Deblur in QIIME2 were created by grouping unique sequences that had the equivalent of 100% similarity to OTUs in QIIME1 (Brandt et al., 2018). The feature-table-table seqs command mapped feature IDs to sequences.
The QIIME diversity alpha-rarefaction visualizer was utilized to explore alpha diversity. Next, the taxonomic composition of the samples was explored. The naive Bayes classifier and the q2-featureclassifier plugin with the Greengenes 13.8 database were used to assign taxonomy to the sequences and map the sequences. The relative abundances at the phylum and family levels were determined based on OTU tables.
PICRUSt2 was used to predict metagenomic functions based on the normalized OTU tables (Douglas et al., 2020).

| Statistical analysis
STAMP was used to calculate the level of significance (Parks, Tyson, Hugenholtz, & Beiko, 2014

| Patient and public involvement statement
Patients were not included in the sampling for this study. Thus, the microbial α-diversity significantly increased with age.

| Microbial composition was altered by age
Since α-diversity significantly changed with age, we detected the gut microbiota composition in elderly individuals and compared it with that of adults 18-60 (Figure 1c,d

| Microbial composition and function were altered by exercise frequency
To determine the effect of exercise frequency on microbial α-diversity in elderly individuals, elderly individual samples were divided into 5 groups (never, rare (a few times/month), occasional (1-2 times/week), regular (3-5 times/week), and daily). The analysis showed that the OTU numbers were 194.9, 195.5, 196.3, 186.7, and 191.9 (p > .05), while the Shannon indices were 5.478, 5.466, 5.489, 5.351, and 5.332, respectively (p > .05; Figure 2a,b). Thus, these results suggested that the microbial α-diversity was almost unaffected by exercise frequency in elderly individuals.
To define the benefit of exercise in elderly individuals, the microbial abundances of the never, rare, occasional, regular, and daily exercise groups were compared with that of adults 18-60 (Figure 2c,d).
At the phylum level, the relative abundances of Actinobacteria, of those pathways were significantly higher, while the abundances of 5 of those pathways were significantly lower in the DRE group than in the NRE group. Therefore, these results suggest that regular exercise significantly modulated microbial function in elderly individuals.
F I G U R E 1 Alteration of the microbial composition in elderly individuals. Both (a) OTU number and (b) Shannon index showed that the microbial α-diversity increased with age in 2,011 people with a normal BMI. Comparing the microbial composition in elderly individuals to that in adults 18-60 , (c) three phyla were significantly changed. (d) Twenty-seven families were significantly changed in the microbial composition in elderly individuals with different exercise frequencies compared with that in adults 18-60

| Microbial α-diversity was altered by regular exercise in OE individuals
The analysis showed that the OTU numbers were 165.5, 184.3, and 200.6 (p < .001), while the Shannon indices were 5.054, 5.244, and 5.508 (p < .001) in the underweight elderly, NE, and OE groups, respectively ( Figure 4a,b). Thus, microbial α-diversity significantly increased with BMI.
The α-diversity analysis showed that the OTU numbers were 207.2 and 195.2 (p < .001), while the Shannon indices were 5.681 and 5.508 (p < .001) in the DROE and NROE groups, respectively ( Figure 4c,d). Thus, the microbial α-diversity was significantly affected by exercise in OE individuals.

| Microbial abundance was partially restored by regular exercise in OE individuals
Next, the abundances of gut microbiota constituents were detected at the phylum and family levels in OE individuals ( Figure 5)

| Microbial functions were partially restored by regular exercise in OE individuals
Microbial functions were detected in OE individuals, and it was found that the relative abundances of 129 pathways were significantly changed (Appendix 1: Table A1). The relative abundances of 79 pathways were significantly increased, while those of 50 pathways were significantly decreased in the overweight group compared with the normoweight groups. Moreover, 25 pathways were identified by comparing the DROE and NROE groups (Appendix 2: Table A2).
Among them, 19 pathways were significantly higher and 6 pathways were significantly lower in the DROE group than in the NROE group.
Notably, 13 common pathways were identified ( Figure 6). Among them, 12 pathways that were changed by overweight were restored by frequent exercise. The abundances of purine nucleotides de F I G U R E 3 Alteration of microbial function in the DRE group. Twenty-nine pathways were significantly changed in the DROE group compared with the NROE group. The 29 pathways were associated with vitamin, nucleotide, glucose, and amino acid metabolism

| D ISCUSS I ON
The gut microbiota plays a major role in age-related diseases.
Moreover, changes in gut microbiota diversity have profound impacts on host metabolism. In the present study, we detected the effect of exercise on the gut microbiota in elderly individuals by using data obtained from the American Gut Project. The microbial α-diversity increased with age in the normal BMI population, and the gut microbiota was restored by exercise in elderly individuals.
A recent study showed that microbial α-diversity is positively associated with age in populations such as the United States, the United Kingdom, and Colombia (de la Cuesta-Zuluaga et al., 2019). Kasai et al. (2015) showed that bacterial diversity was significantly higher in obese individuals than in nonobese individuals in a Japanese population. The Shannon index was increased in those with an obese BMI in a randomly selected Alabama resident study (Davis, Yadav, Barrow, & Robertson, 2017). Infants who were obese at 6 months of age had higher levels of alpha diversity than nonobese infants (Ville, Levine, Zhi, Lararia, & Wojcicki, 2020). However, some studies showed that obesity/overweight was associated with decreased α-diversity in animals or humans (Chen et al., 2020;Da Silva, Monteil, & Davis, 2020;van der Merwe et al., 2020). Moreover, a recent study showed no differences in the α-diversity in obese individuals in an Asian population (Koo et al., 2019). In this study, we analyzed the association of gut microbial α-diversity with age and BMI. The analysis showed that the microbial α-diversity increased with increasing age.

F I G U R E 6 Alteration of microbial function in OE individuals. Thirteen pathways showed opposite trends in the NE/OE and DROE/NROE groups
In addition, the microbial α-diversity increased with increasing BMI in elderly individuals. We also showed that the microbial α-diversity was decreased in the overweight group of whole AGP samples (data not shown). Therefore, the analysis results based on AGP are in accordance with the finding that microbial α-diversity is positively associated with increased age and obesity in elderly individuals.  (Jami, Ghanbari, Kneifel, & Domig, 2015).
Bifidobacteria species (phylum: Actinobacteria) are widely used as probiotics and have demonstrated beneficial effects under a variety of pathological conditions (Shen et al., 2018). Studies have indicated that Proteobacteria may be a characteristic microorganism of diseases, including metabolic disorders and inflammatory bowel disease. Proteobacteria can cause inflammation and lead to disease development (Rizzatti, Lopetuso, Gibiino, Binda, & Gasbarrini, 2017).
The increase in intestinal Desulfovibrionaceae abundance is an important feature of colitis (Leonardi et al., 2017). Enterobacteriaceae are important pathogens because of their capacity to produce endotoxins (Xie et al., 2017). Therefore, our results suggested that aging induced an increase in harmful bacteria and a decrease in beneficial microbes.
The microbial function analysis predicted that 25 pathways were significantly different between elderly individuals and adults 18-60 (Appendix 2: Table A2). The amino acid metabolism-related pathways were altered, which might result in decreased biosynthesis of essential amino acids (threonine, phenylalanine, lysine, and tryptophan) and aromatic amino acids (tyrosine, phenylalanine, and tryptophan).
Vitamin biosynthesis-related pathway abundances were decreased.
A previous study showed that the gut microbiota was involved in essential amino acid homeostasis (Lin, Liu, Piao, & Zhu, 2017). The gut microbiota can synthesize K and B vitamins, including biotin, riboflavin, cobalamin, folic acid, nicotinic acid, pantothenic acid, pyridoxine, and thiamine (Rowland et al., 2018). It is estimated that vitamins synthesized by gut microbes can provide more than a quarter of the intake (Rowland et al., 2018). Therefore, our results suggest that age induced a decrease in the production of essential amino acids and vitamins by the gut microbiota, thereby reducing these substances in elderly individuals.
Studies have shown that exercise is linked to microbial composi-  . We compared the microbial composition of each exercise frequency group with that of adults 18-60 and found that the abundances of 3 phyla and 14 families gradually approached those in adults 18-60 with an increase in exercise frequency. The abundance of Actinobacteria gradually increased and was similar to that in adults 18-60 . Additionally, the abundance of Cyanobacteria gradually decreased and approached that in adults 18-60 . Cyanobacteria are associated with diseases and, in some cases, with human and animal death (Lange et al., 2018). Cyanobacteria release their toxins, such as lipopolysaccharides, in B cells in the gut (Swanson-Mungerson et al., 2017). Furthermore, the evolutionary analysis showed that the microbial composition in the daily exercise group was closest to that in adults 18-60 . We also revealed that daily exercise shifts the gut microbiota to a younger phenotype.
In OE individuals, the abundances of Actinobacteria and Bacteroidetes were decreased, while the abundances of Cyanobacteria and Firmicutes were increased at the phylum level. The Bacteroidetes/ Firmicutes ratio was significantly decreased. At the family level, the abundances of S24-7 and Lachnospiraceae were increased, while the abundances of Christensenellaceae, Barnesiellaceae, and Bacteroidaceae were decreased in OE individuals. It was reported that S24-7 enrichment was associated with a high-fat diet in diabetes-sensitive mice (Serino et al., 2012). Kameyama et al showed that intestinal colonization by a Lachnospiraceae bacterium promoted the development of diabetes in obese mice (Kameyama & Itoh, 2014). It was reported that Christensenellaceae bacteria were beneficial for human health and had an increased abundance in lean people (Requena, Martínez-Cuesta, & Peláez, 2018). Barnesiellaceae could serve as a marker to discriminate lean and obese individuals (Rodriguez, Benninghoff, Aardema, Phatak, & Hintze, 2019). Hakkak, Korourian, Foley, & Erickson (2017) showed that lean rats exhibited much lower Firmicutes to Bacteroidetes ratios than obese rats. A recent report showed that Bacteroidaceae might play a role in bamboo shoot fiber-mediated suppression of high-fat diet-induced obesity (Li, Guo, Ji, & Zhang, 2016). Therefore, our results suggest an increase in harmful bacteria and a decrease in beneficial microbes in OE individuals.
Next, we revealed that the microbial α-diversity was significantly Proteobacteria abundance (Munukka et al., 2018). Therefore, our results suggest that regular exercise might play a role in decreasing harmful bacteria and increasing beneficial microbes in OE individuals.
Thirteen common pathways were isolated and showed opposite trends in the OE/NE and DROE/NROE groups. Ten pathways related to nucleotide biosynthesis were significantly inhibited in OE individuals and enhanced after regular exercise. Thus, these results suggest that regular exercise might modulate nucleotide biosynthesis, which is important for microorganism growth.
This study relied on the participants' accurate self-reporting of the frequency of exercise without tracking the intensity of exercise.
Studies have shown that insufficient or excessive exercise intensity has a significant impact on the structure and function of the intestinal flora. Moreover, the main research participants in this study were Caucasian, and the amount of data from other ethnic groups is small. Future research should specifically study other ethnic groups for comparative analysis.

| CON CLUS ION
In conclusion, our results revealed that microbial diversity increased with increasing age. The gut microbiota composition of daily exercising elderly individuals approached that of adults 18-60 , and regular exercise resulted in an increased relative abundance of bacterial functional pathways related to nucleotide metabolism, glucose metabolism, and lipid metabolism. We also showed that OE individuals had increased microbial diversity and significant changes in the microbial composition, which responded with alterations in bacterial functional pathways related to vitamin, nucleotide, and glucose metabolism. Furthermore, regular exercise partially reshaped the microbial composition. Altogether, our findings support the role of regular exercise in maintaining the stability of the gut microbiota in elderly individuals and reveal that regular exercise benefits OE individuals.

ACK N OWLED G M ENTS
This study was funded by the National Natural Science Foundation of China (No. 81960672) (Guankui Du).

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
Qiwei Zhu: Data curation (lead). Shangfei Jiang: Software (lead).

E TH I C S S TATEM ENT
None required.