There is growing awareness that microbial communities colonize different regions of the gastrointestinal tract, playing a major role in the health and disease of their host [1
]. In the healthy state, the commensal microbes help to digest and absorb nutrients, modulate the immune system and provide protection against enteropathogens [1
]. Gut microbial imbalance, followed by a state of dysbiosis, in turn, is associated with obesity [3
], type 2 diabetes [4
], cardiovascular disease [5
] and non-alcoholic fatty liver disease [6
]. Modulation of the gut microbiome has therefore become a topic of considerable interest.
Although using antibiotics is considered to be an efficient way to modify the microbiota, recent studies have suggested that dietary modification and regular exercise may offer cost-effective alternative means to achieve the same end [7
]. A cross-sectional study of professional rugby players showed that exercise is associated with gut microbial diversity, and that proportions of several microbial taxa were significantly higher in the rugby players compared with the control group [9
]. A recent study by Estaki et al. [10
] also showed a positive correlation between cardiorespiratory fitness and microbial diversity in a small group of young healthy adults. However, earlier studies were done in athletic men and young adults; a current study has shown that estrogen in women influences gut microbiota [11
], but overweight or obese women with relatively low physical fitness remain to be studied.
Physical fitness is considered one of the most objective measures of the level of physical activity [12
]. Cardiorespiratory fitness is the overall capacity of the cardiovascular and respiratory systems and the ability to carry out prolonged strenuous exercise. The maximal oxygen consumption (VO2max
) attained during a graded maximal exercise to voluntary exhaustion has long been considered by the World Health Organization as the single best indicator of cardiorespiratory fitness [13
]. Therefore, to study whether the level of physical fitness is associated with certain compositions of the gut microbiota, using VO2max
as a segregate variable is possible option in a cross-sectional study design. The aim of the present study was to examine the relationship between gut microbiota and aerobic fitness in premenopausal women with diverse body composition (mostly overweight or obese with a sedentary lifestyle). More specifically, we examined (1) the correlations between cardiorespiratory fitness and specific bacteria groups, and (2) whether these correlations between cardiorespiratory fitness and specific bacteria groups were independent of age, dietary intake and whole-body fat mass.
2. Materials and Methods
2.1. Study Design and Subjects
The study participants consisted of 71 Finnish women (aged 19 to 49 years) who resided in the city of Jyvaskyla, Central Finland, which has a population of approximately 150,000 [14
]. In our early study, 80% of the people who participated in our study were still living in Jyvaskyla after 10 years and there were no differences between people who were living inside the city compared to those who were living outside the city (unpublished data). However, duration of living in the city of Jyvaskyla was not a required inclusion criterion. The majority (80%) of the study participants were either overweight or obese; but had no diagnosed cardiovascular disease, type I or type II diabetes or serious musculoskeletal problems. None of the participants had been on antibiotics treatment during the past three months. The study protocol was approved by the ethical committee of the Central Finland Health Care District (No: 7/2011). Informed consent was obtained from each participant prior to the assessments.
2.2. Background Information
Background information, including medical history and current health status, was collected via self-administered questionnaires. The level of physical activity in terms of duration (exercise hours per week) and frequency (exercise times per week) was also collected by validated questionnaires [15
]. Food consumption and intakes of total energy and energy-yielding nutrients were assessed from food records that were kept for three days (including two weekdays and one weekend day). Food diary records were analyzed for total calorie, protein, carbohydrate, and fat intake by using the software Micro-Nutrica (v3.1), developed by the Social Insurance Institution of Finland and updated with a database for new foodstuffs by the study nutritionist [17
]. To minimize possible under-reporting, the proportion of total energy intake of energy-yielding nutrients (E%) was calculated and reported in this study.
2.3. Anthropometrical and Body Composition Assessments
Body height (cm) was measured using standardized protocols (a wall-fixed measuring device). Body weight (kg) and whole-body fat mass were assessed with light clothes and without shoes using bio-impedance (Inbody 720, Biospace Co. Ltd., Seoul, South Korea). Body mass index (BMI) was calculated according to weight/height2 (kg·m−2). Percent of body fat (fat%) was calculated as whole-body fat mass divided by weight.
2.4. Fitness Test
The maximum oxygen uptake (VO2max in mL/kg/min) was assessed by a bicycle ergometer under the supervision of a physician. The test began with a 2-min warm-up at 50 W. After that the intensity was increased by 25 W at 2-min intervals until exhaustion. Electrocardiography was monitored continuously and heart rate and maximal work load were recorded at the end of every load. Oxygen uptake was assessed by the breath-by-breath method using a respiratory gas analyzer (Sensor Medics Vmax, Yorba Linda, CA, USA). Maximal oxygen uptake was reached when the measured VO2 reached a plateau or started to decrease, or the subject felt she had reached her maximal level and wanted to stop the test. On the basis of the VO2max values, participants were divided into three groups by tertiles (low, moderate and high).
2.5. Blood and Biochemicla Measurements
Venous blood samples were taken in standardized fasting condition (12 h) in the morning (7–9 a.m.). Serum samples were stored frozen at −80 °C until analyzed. Serum triglycerides, total cholesterol and high-density lipoprotein (HDL) were determined by using KONELAB 20XTi analyzer (Thermo Fischer Scientific Inc, Waltham, MA, USA) and described previously [18
]. The intra- and inter-assay correlation coefficients (CVs%) were 3.4% and 2.9% for triglycerides. Serum leptin was assessed using human leptin (ELISA; Diagnostic Systems Laboratories, Inc., Webster, TX, USA). The inter- and intra-assay coefficients of variation (CVs%) were 2.2% and 2.7% for leptin, respectively.
2.6. Fecal Samples
Fecal samples were taken from evacuated stool by the subjects with detailed guidance and frozen immediately and stored at −70 °C until processing. The bacterial cells were separated and analyzed with a previously described method using 16S rRNA hybridization, DNA-staining, and flow cytometry [19
]. The following five 16S rRNA-targeted oligonucleotide probes labelled at the 5′-end with Cy5 indocarbocyanine (Ex/Em 646/662 nm; Molecular Probes, Eugene, OR, USA) were used: Bacto1080 for Bacteroides
group, Bif164 for Bifidobacterium
species, Enter1432 for Enterobacteria
(Ent), EreC482 for Eubacterium rectale-Clostridium coccoides
group) and Fprau645 for Faecalibacterim prausnitzii
2.7. Statistical Analyses
Data analyses were carried out using SPSS Statistics 22 for Windows (SPSS Inc., Chicago, IL, USA). All data was checked for normality using the Shapiro–Wilk W-test. If data were not normally distributed, their natural logarithms were used. Multiple comparisons were applied by using analysis of variance (ANOVA) with the Sidak post-hoc test.
Linear regression was used to assess the relationship between cardio-respiratory fitness and gut microbiota. The analyses were performed with and without adjusting for age, dietary intake and fat%. The p-values were adjusted by multiple comparisons in FDR (false discovery rate). Statistical significance was set at p < 0.05 with 2-tails.
In addition, multivariable least square regressions were performed in assessing the contribution of the variables (i.e., EreC, age, energy yield nutrients of fat, carbohydrates, protein and alcohol, as well as fat% of the whole body) and the outcome variables (i.e., VO2max, leptin, HDL and TG). To make each outcome comparable, we standardized each column by means of z-scoring so that each column has mean value 0 and standard deviation 1. The Pearson correlation coefficients between the regressed outcome and the observed outcome were calculated and regression weights for each variable were provided. The higher the absolute weight, the higher the contribution.
In this study, we found that gut microbiota was associated with cardiorespiratory fitness in young and middle-aged women. The study participants with low aerobic fitness had higher Eubacterium rectale-Clostridium coccoides and Enterobacteria, but lower Bacteroides. These differences were independent of age and macronutrient intake, but appeared to be confounded by adiposity, since all differences between the groups disappeared after adjusting for the percent of body fat.
There is growing interest in gut microbiota and their role in health and disease. However, research in this field is still in the beginning phase, therefore, the results are inconclusive but illustrative. It has been shown that obese mice had significantly lower Bacteroidetes
but a higher proportion of Firmicutes
compared to lean mice [21
]. A similar observation was also found in a human study when comparing obese to lean twins [22
]. Previous studies also showed that the EreC
group is associated with obesity and related to metabolic disorders [23
]. Our current results are in agreement with these studies by showing that the low VO2max
group (which had a high BMI) had significantly lower Bacteroides
but higher EreC
) and Enterobacteria
(Phylum Proteobacteria). EreC
contributes the most to the regression of VO2max
. However, Goodrich et al. found in the large TwinsUK population study that Bacteroidales
) and family Clostridiaceae
) correlated negatively with BMI and triglycerides [27
]. Similar results were reported in another large population-based cohort by Fu et al. [28
] The discrepancy between these studies may partly be due to differences in sample size, study population and the methods used to assess the bacteria, it might also be because different bacterial species (within the same phyla) have different functions and thus also different associations. In addition, the intestinal microbiota is known to regulate host energy homeostasis and can be influenced by diet and other environmental factors which are still not well characterized [28
Lifestyle factors such as diet and exercise contribute largely to obesity and other cardio-metabolic disorders. Several studies have shown that a low carbohydrate diet can decrease EreC
group proportions [29
]. A decreased Firmicutes
ratio was found in an intervention with a calorie-restricted diet [31
]. Our result showed that the EreC
group was negatively correlated with carbohydrate intake but positively with fat intake. No association was found between the EreC
group and fiber intake. Earlier studies have indicated that a dietary-resistant (or “non-digestive”) carbohydrate—through microbial conversion to short-chain fatty acids (SCFA)—is less effective than the equivalent amount of sugar absorbing directly in the small intestine for harvesting the energy [29
]. SCFAs serve as energy substrates for the epithelial cells of the gut and provide part of the dietary energy [29
], while other studies have shown that microbial pathways which generate SCFAs were enriched in metagenomics studies of obese subjects, and levels of SCFAs were elevated in overweight or obese people and animal models [25
]. Hence the changes in gut microbiota composition that result from different dietary intakes are still inclusive.
The underlying mechanisms by which cardiorespiratory fitness might be associated with microbiota are yet to be fully understood. A recent study by Estaki et al. [10
] showed that cardiorespiratory fitness is associated with increased gut microbial diversity. The association between cardiorespiratory fitness and certain bacteria in our study was confounded by adiposity, which raised the question of whether cardiorespiratory fitness plays a role in altering gut microbiota. One of the possible links between cardiorespiratory fitness and certain bacteria may be that physical activity decreases total colonic transit time. Long-term regular physical activity (resulting in high cardiorespiratory fitness) has a positive effect on both constipation indices and rectosigmoid transit time [34
]. An earlier study has showed that a short gastrointestinal tract (GIT) resulting from gastric bypass enriched the gut microbiota of phylum level Bacteroidetes, Verrucomicrobia, and Proteobacteria after the surgery [36
]. It is possible that the speed of colonic transit might be higher in those subjects who had higher VO2max
. Furthermore, physically active individuals are more likely to follow a healthy lifestyle and to be more exposed to their environmental biosphere, and this may also contribute to a different composition of microbiota. Simultaneously, adaptation and the acute effect of endurance training can lead to changes in the GIT, such as increased transit and absorptive capacity, tissue hypoxia, and decreased blood flow [37
]. These and other potential mechanisms, such as changes in gut pH, may create an environmental setting which limits the growth of the EreC
group. On the other hand, controlling for fat%, the association between the EreC
group and VO2max
disappeared and fat% contributes the most in the regression model. Hence, whether exercise or fitness level has significant impact on the time of chyme staying in the intestinal tract, altering the composition of gut microbiota, deserves further study.
An earlier study has reported that the gut microbiome contributes to a substantial proportion of the variation in blood lipids [28
]. Gut microbiota have been linked with lipid metabolism through their role in bile acid metabolism [40
]. Fu et al. used 16s rRNA gene sequencing and found that members of family Ruminococcaceae
(OTU 175962, 1703711, 295743, 178385) were negatively correlated with triglycerides, but certain members of the genus Eubacterium
(OTU 49837) positively correlated with triglycerides [28
]. However, Karlsson et al. used shotgun sequencing to find that the Clostridium
species and Eubacterium rectale
were correlated positively with triglycerides and leptin, and negatively with HDL [41
]. We found that a high proportion of EreC
was associated with high triglycerides, but low HDL. Controlling for fat%, the significant associations between the EreC
group and triglycerides remained, but disappeared between the EreC
group and HDL. Since the EreC
group includes members of Clostridium oroticum
, Clostridium nexile
, Ruminococcus hansenii
, Ruminococcus productus
, and Eubacterium rectale
, it indicates the complexity of the bacterial in relation to the clinical outcomes. Nerveless, this could partly explain why different associations were reported for the EreC
group with triglycerides and HDL.
Our study had some limitations. First, the study used a cross-sectional design and therefore no cause and effect relationship could be identified. Second, the method we used to assess gut microbiota was based on the detection of conserved regions of the 16S ribosomal RNA gene [42
]. Though the method we used can detect the specific groups and clusters of gut microbiota [43
], it cannot reach the deeper taxonomical levels such as species and strains, which might be important in the assessment of the functional role of intestinal microbiota. Thus, a more detailed structural and functional analysis of gut microbiota based on, for example, next generation sequencing methods in a larger scale of the population is required in order to explore the entire gut microbiota community [47
]. One other limitation is that most of our participants were overweight or obese with relatively low fitness levels, this may have resulted in the underestimation of the association between VO2max
and bacteria. Finally, only premenopausal women were included in this study, thus the effect of estrogen on the gut microbiota [11
] cannot be verified. Nerveless, our results provide new information that there is a moderate association between physical fitness and certain gut microbiota in premenopausal women.