Numerous studies have shown that prolonged intense physical exercise is associated with a transient depression of immune function in athletes. While moderate exercise beneficially influences the immune system [1
], a heavy schedule of training and competition can lead to immune impairment associated with an increased risk of upper respiratory tract infections (URTIs) due to altered immune function [2
]. It has been suggested that exhaustive exercise creates a potential ‘open window’ of decreased host protection, during which viruses and bacteria can gain a foothold, increasing the risk of developing an infection [4
]. During major competitions of 2–3 weeks duration, typically about 7% of athletes experience at least one episode of illness and about half of these are respiratory [5
]. Exercise immunological studies reported that infection episodes were preceded by declines in immunoglobulin A (IgA) in saliva [6
]. Furthermore, results suggest a possible mechanism for the increased incidence of infection during intensified training via modulation of type 1/type 2 T lymphocyte distributions [9
Physical exercise and sports influence immunoregulatory circuits which, as a primary response, involve the production of forward regulatory cytokines is followed by counter-regulation leading to an immunosuppressed state [3
]. Downstream biochemical events include changes in tryptophan (Trp) metabolism when T helper cell type 1 (Th1-type) cytokine interferon-γ (IFN-γ) is released and induces tryptophan-degrading enzyme indoleamine 2,3-dioxygenase (IDO-1). In turn, blood concentrations of Trp become reduced, leading to various potential consequences [12
]. The essential amino acid Trp is not only a precursor of the serotonin biosynthesis pathway but is also the key element for the formation of the energy carrier and coenzyme nicotinamide-adenine-dinucleotide NAD and its reduced form NADH via the so-called kynurenine (Kyn) pathway [13
]. Recently, exhaustive aerobic exercise in athletes was reported to significantly impact on Trp–Kyn metabolism [15
]. Results indicate an involvement of IDO-1 activation in enhanced Trp catabolism and Kyn production following demanding exercise [15
]. The close association of Trp metabolites with neuropsychopharmacologically relevant metabolites may have special consequences for athletes since it influences immunosurveillance and the development of infections as well training adherence because of disturbed neurotransmitter biochemistry [16
Trp is also an important target for the gut and brain interaction [17
]. In addition to its resorption from dietary components, the composition of gut bacteria—the microbiome—is of enormous importance in the regulation of Trp. Available data suggest a role for the gut microbiota in actually modulating Trp and hence having control over serotonin levels in the host [18
]. Recently, an inverse correlation of serum levels of Trp, tyrosine, and phenylalanine with concentration of fecal calprotectin, a marker for gut leakiness, has been reported in patients suffering from Alzheimer’s disease, thus indicating a close relationship between the intestinal barrier function and aromatic amino acid concentration in the blood [19
]. Furthermore, there is growing body of evidence indicating that the microbiota is sensitive to physiological changes associated with exercise [20
]. For example, acute aerobic exercise reduces the expression of toll-like receptors (TLRs) in the monocyte cell-surface, contributing to post-exercise immunodepression, while over the long-term, a decrease in TLR expression may represent a beneficial effect because it decreases the inflammatory capacity of leukocytes, thus altering whole body chronic inflammation [22
]. TLRs can activate dendritic cells, which are associated with the attenuation of immune activation and inflammation protection [20
]. Notably, IDO-1 has been identified in mucosal Cluster of Differentiation 103 -expressing dendritic cells and has already been claimed to be a possible therapeutic target for gut disorders [23
Dietary supplements containing probiotics can modify the population of the gut microflora and may provide a practical means of enhancing gut and systemic immune function, which was shown to be beneficial by reducing the infection frequency in sensible groups, e.g., elderly in group homes or children [24
]. However, studies in these subject groups might not be reflective of athletes who have different gut microbiota [26
]. Exercise and associated dietary extremes were shown to increase gut microbial diversity in comparison to sedentary people [27
]. Some studies have established that probiotic intake can improve low-grade inflammation [28
] and enhance resistance to URTI in athletes [30
]. In a previous study, Lamprecht and colleagues found that adequate probiotic supplementation composed of six strains consisting of Bifidobacterium bifidum
W23, Bifidobacterium lactis
W51, Enterococcus faecium
W54, Lactobacillus acidophilus
W22, Lactobacillus brevis
W63, and Lactococcus lactis
W58 could improve redox hemostasis and low-grade inflammation in men under sustained exercise stress [29
]. The mechanisms behind these observations have not been widely investigated but may include direct interaction with gut microbiota, interaction with mucosal immune system and modulation of lung macrophage and T cell functions [33
]. For example, one study observed that the IFN-γ response (a potent stimulus for IDO-1) was moderately higher with probiotic treatment than with placebo, associated with a significant reduction in the number of days of respiratory illness symptoms in highly trained distance runners [30
]. Since Trp availability is primarily regulated via the Kyn pathway, the catabolism of amino acid Trp via Kyn may play an important role on the risk of developing an infection.
The aim of the present study was to examine the effect of a probiotic supplement on the incidence of URTI and Trp metabolism after exhaustive aerobic exercise in trained athletes during three months of winter training We hypothesized that daily supplementation with probiotics is beneficial in reducing the incidence of URTI in athletes during training periods in winter and is associated with modulation of the Trp—Kyn metabolic pathways.
2. Materials and Methods
Thirty-three healthy and trained volunteer athletes (mean age 26.7 years; average body mass index 22 kg/m2
; average peak oxygen uptake 51.4 mL/kg/min) participated in this study that was conducted at the Department of Sport Science at the Leopold Franzens University of Innsbruck, Austria. Individuals were invited to participate if they were 20–35 years of age, non-smokers, had no previous history of muscle disorders and were free of heart, kidney, lung, neurologic, and psychiatric diseases. Athletes with a cardiorespiratory response and fitness of ≥150% of reference values during maximal exercise [34
] were included. A questionnaire about medical history and previous training was filled out by each participant. In total, 33 individuals were enrolled with 29 participants (13 men 16 women) completing the study. Baseline characteristics of the subjects are presented in Table 1
Subjects who met the inclusion criteria of the study were randomly assigned to the treatment or placebo group. The randomization code was held by a third party and handed over for statistical analyses after collection of all data. All of the participants were informed of the risks and potential discomforts associated with the investigation and signed a written consent to participate. The study was approved by the Board for Ethical Questions in Science Ethics at the Leopold Franzens University of Innsbruck according to the principles expressed in the Declaration of Helsinki.
2.2. Study Intervention
Subjects randomized to probiotics (PRO, n
= 17) received boxes with sachets containing multi-species probiotics composed of six strains consisting of Bifidobacterium bifidum
W23, Bifidobacterium lactis
W51, Enterococcus faecium
W54, Lactobacillus acidophilus
W22, Lactobacillus brevis
W63, and Lactococcus lactis
Performance, Winclove B.V., Amsterdam, The Netherlands). The total cell count was adjusted to 2.5 × 109
colony forming units (CFU) per gram. The candidate strains were selected upon their survival in the gastrointestinal tract, activity, intestinal barrier function, and anti-inflammatory properties and were used in a previous study on immune health in athletes [29
]. The matrix consisting of cornstarch, maltodextrin, vegetable protein, MgSO4
and KCl. Subjects were instructed to take 1 sachet of 4 g per day, which is equivalent to 1 × 1010
CFU/day, with 100–125 mL of plain water, one hour prior to breakfast and throughout the 12 weeks. Those subjects assigned to the placebo group (PLA, n
= 16) received identical boxes and sachets with the same instructions for use.
2.3. Study Protocol
During the three-month intervention period (January 2015 to March 2015) subjects were asked to maintain their normal diet and to continue with their normal training programs. In addition, participants agreed to avoid taking medicine including anti-inflammatory drugs (e.g., aspirin, ibuprofen, voltaren) and antibiotics, additional probiotics and dietary supplements such as fish oil, vitamins (vitamin C, vitamin E) and minerals (selenium). Consumption of alcohol (>10 and 20 g for women and men, respectively, per day), or any fermented dairy products (e.g., yoghurt) was not permitted during this period. During the first visit to the laboratory, measures of participants’ weight and height were obtained using standardized methods and used to calculate body mass index (BMI, kg/m2). Prior to and at the end of the study, all subjects were tested for body fat (in percent of body weight), body cell mass (kg), and resting energy expenditure (kcal/day) using the bioelectrical impedance analysis (BIA) method (BIA-2000-M, Data Input, Pöcking, Germany). Prior to the first blood draw and after 12 weeks of supplementation, participants were asked to complete a three-day food record to evaluate energy and nutrient intake. Diet records were analyzed for total calories, protein, carbohydrate, fat, alcohol, and water intake using “nut.s science” nutritional software (dato Denkwerkzeuge, Vienna, Austria). Weekly training (modality, frequency, intensity, volume) and illness (URTI symptoms and gastrointestinal GI complaints symptoms) logs were kept.
The illness symptoms listed on the self-constructed questionnaire, modified according to Gleeson et al. (2011) [31
] were sore throat, runny nose, cough, fever, and weakness. Subjects were asked to rate the severity of their symptoms (very light, light, moderate, severe, very severe). The GI discomfort symptoms listed on the questionnaire were abdominal pain, diarrhea, loss of appetite, vomiting, and others. The incidence score relates to the number of participants who reported symptoms in each arm of the study. One or more symptoms on at least two consecutive days were defined as an episode of illness. Symptoms with an interval of only one day were counted as the same episode.
2.4. Exercise Tests
In the morning of the exercise test a standardized breakfast was provided 2 h prior to strenuous exercise tests (379 kcal; 88 energy percent carbohydrates, 11 energy percent proteins, and 1 energy percent fat). The composition of this standardized breakfast is shown in Table 2
For eligibility testing all subjects performed an incremental cycle ergometer exercise test until exhaustion. Cycle ergometry was performed on an electronically braked ergometer (Ergometrics 900, Ergoline, Germany) and started at a workload of 50/75 W (women/men) for 5 min (warm up) with a following increase in workload of 25 W per minute until exhaustion. Exhaustion was defined as the state when the pedaling rate dropped below 60 rpm. Heart rate and ventilatory parameters were monitored continuously (Oxycon mobile, Jaeger, Germany). Peak power output (Pmax
) was defined as the last completed workload rate plus the fraction of time spent in the final uncompleted work rate multiplied by 25 W [35
]. Peak oxygen uptake (VO2max
) was defined as the highest 30-s average during the test.
After a 20 min resting period, athletes performed a 20-min maximal time-trial on a cycle ergometer (RBM Cyclus 2, Leipzig, Germany) as described by Faulhaber and colleagues [35
]. Briefly, the cycle ergometer was shifted to a fixed pedal force in which power output was dependent on the pedaling rate. Pedal force for each participant was set so that pedaling at 100 rpm would produce about 70% (rounded to 5 W) of peak power output, which was determined by the incremental cycle ergometry. During the test, cyclists were strongly encouraged to choose a maximal pedaling rate that could be maintained for the respective test duration. The main outcome measurement was mean power output during the 20-min test, which was automatically calculated by the software of the ergometer. The participants were allowed to drink water ad libitum. Three months later this procedure was repeated on the same cycle ergometer and with the same investigator.
2.5. Blood Measurements
We conducted blood collections from the participants in the supine position from a medial cubital vein at baseline and after 12 weeks at rest and within 5 min after exercise (four blood draws per study participant). After centrifugation for 10 min cells were removed and plasma samples were frozen at −20 °C until analysis. Serum concentrations of Trp and Kyn as well as concentrations of phenylalanine (Phe) and tyrosine (Tyr) were determined by high-performance liquid chromatography (HPLC), as previously described [36
]. The ratios of Kyn/Trp and Phe/Tyr were calculated as indexes of Trp degradation and phenylalanine 4-hydroxylase (PAH) activity, respectively. Pro-inflammatory cascades were found to be associated with disturbed PAH activity [37
]. Serum neopterin concentrations were measured by ELISA (BRAHMS Diagnostics, Hennigsdorf, Germany) following the manufacturer’s instructions [38
2.6. Statistical Analysis
Per protocol analyses were performed using SPSS (IBM SPSS Statistics Version 22, IBM Corp., Armonk, NY, USA). Normality in the distribution of data was tested using the Kolmogorov-Smirnov’s test and Boxplots. In the case of Gaussian distribution, baseline characteristics, performance data, nutrient and biological markers were compared by unpaired Student’s t
-test or Mann-Whitney-U
-Test. Changes in variables during the study were analyzed by univariate analysis of variance (ANOVA) for parametric variables. The Wilcoxon-signed rank and Friedman test were applied to non-parametric data. Spearman’s rank correlation was used to assess the association between two variables. Partial eta-squared values were calculated to estimate the effect of any statistically significant differences found. Using the guidelines of Cohen [39
], 0.01 = small effect, 0.06 = moderate effect, and 0.14 = large effect. A p
-value of less than 0.05 (two-tailed) was considered to indicate statistical significance. Data are presented as mean values ± standard deviation (SD) or by mean values ± standard error of the mean (SEM).
Sample size calculation was based on changes in exercise-induced Trp levels [40
] from baseline to the end of the 12-week intervention between the PRO group and the control. We estimated between 10 and 12 subjects per group—depending on SD and effect size—to reach a probability of error (alpha/2) of 5% and 80% power. Allowing for a drop-out rate of 30%, 16 subjects per group were recruited.
3.1. Study Population
Twenty-nine of the 33 randomized subjects completed the full program and entered statistical analyses. Three withdrew because of injury or persistent illness with antibiotic medications, one because of a longer training interruption. Returned sachet count after the treatment period revealed a compliance rate >95% in both groups (97.6% in the probiotics group, 98.8% in the control group). The lowest level of compliance for a subject was 86.9%. A CONSORT (Consolidated Standards of Reporting Trials) diagram outlining participant recruitment is depicted Figure 1
At baseline, a significant gender-dependent difference (females were overrepresented in the control group), VO2max and Trp was observed between groups (p < 0.05). Females had a lower BMI, VO2max, and mean power output during the 20-min test (PTT) compared to male athletes, as Kyn levels were lower in females (p = 0.019). None of the other parameters were influenced by gender.
3.2. Training Loads
Analysis of training loads indicated that the weekly training of the aerobic system, mainly continuous endurance training at moderate intensity (60% to 80% VO2max
), varied significantly between the groups over the 12-week treatment period (Figure 2
). The means were significantly higher in the probiotics group as compared to the placebo group: 8.0 ± 2.3 and 6.6 ± 4.3 h per week endurance training, respectively (U
= 2.597, p
3.3. Body Composition, Nutrition, and Performance
After 12 weeks of treatment, there was no significant difference between probiotic supplementation groups and placebo groups in anthropometric characteristics, body composition, and food intake (p
> 0.05). Performance (VO2max
) remained unchanged over time and still differed significantly between groups in week 12 (p
< 0.05). Resting energy expenditure (REE, kcal/day) was significantly different between groups after 12-weeks of the study (mean ± SEM: 1617 ± 57 kcal/day and 1518 ± 56 kcal/day for PRO and PLA, respectively; p
< 0.05, η2
= 0.13; Figure 3
3.4. Amino Acids
At the beginning of the study, exhaustive exercise induced a decrease in Trp levels in both the probiotic and the placebo group (Table 3
). At the end of the experimental protocol, the exercise-induced Trp shift was comparable to the shift in week 0 in subjects who ingested probiotics but was more pronounced in in the placebo group (approximately 10% lower than in week 0, p
< 0.05) (Figure 4
These data indicate reduced Trp degradation rates in subjects supplemented with probiotics, although this effect was not significant (p = 0.13, η2 = 0.08). It should be mentioned that baseline Trp concentrations were slightly but significantly lower in the placebo group compared to the probiotics group, most probably due do the different percentage of female athletes in the groups. In parallel to Trp decrease, Kyn/Trp and neopterin levels were increased after exercise in both study groups at both time points.
Further, at the beginning of the study, VO2max correlated significantly with baseline concentrations of Trp (rs = 0.562, p = 0.001) and this relation remained significant after 12 weeks of treatment (r = 0.497, p = 0.006) but was no longer present after intense exercise.
Tyrosine levels significantly increased and Phe/Tyr significantly decreased with exhaustive exercise (p
= 0.018 and p
< 0.001, respectively), but there were no significant time-dependent differences between groups. Serum concentrations of Phe were not significantly affected, either by exercise or by supplementation (Table 3
3.5. Immune System Biomarkers
Exhausting exercise was associated with a strong increase in neopterin levels up to +61% (U = 4.420, p < 0.001) and +63% of pre-exercise values (U = 4.660, p < 0.001), before and after 12 weeks of treatment, respectively, with no significant differences between and within groups over time. However, this increase was significantly influenced by endurance training volume with a strong inverse correlation between the athletes’ training status and the concentrations of neopterin at exhaustion (rs = −0.502, p < 0.01).
Kyn concentrations were slightly increased with exercise by 7% (U = 2.671, p < 0.01) before and by 3% (U = 0.923, n.s.) after 12 weeks of intervention, contributing to the elevation of the Kyn/Trp ratios by 22% (U = 4.544, p < 0.001) and by 21% (U = 4.433, p < 0.001), respectively. Exercise induced a change in Kyn levels with time (∆Kyn), with a significant decline being overserved in the PLA group (p = 0.04), whereas an increase was seen in the PRO group, but this effect was not significant between groups (p = 0.05, η2 = 0.13). At baseline, neopterin and Kyn/Trp ratios correlated significantly (rs = 0.490, p < 0.01), with the association even becoming slightly stronger upon exercise (rs = 0.512, p < 0.01). After 12 weeks there was no longer a significant relationship between pre-exercise neopterin and Kyn/Trp levels (rs = 0.280, n.s.), but it became again significant after exercise (rs = 0.583, p = 0.001). At the same time, higher neopterin levels correlated with lower Trp concentrations (rs = −0.384, p < 0.05).
3.6. Infection Incidence
Only one participant on the placebo experienced GI-discomfort symptoms during the study period. Analysis of the URTI-symptom questionnaires indicated that 55% (16 subjects) of the cohort experienced an URTI episode during the 12-week study period. Thirteen subjects did not experience any URTI episode during the study period. Before supplementation, 10 subjects on placebo and 12 subjects on probiotics experienced one or more URTI symptoms over the prior three months. After 12 weeks of treatment, 11 subjects on placebo and 5 subjects on probiotics experienced one or more URTI symptoms during the study period (Figure 5
). The proportion of subjects who experienced one or more URTI symptoms during the study period was 2.2-fold higher in the placebo group than in the probiotics group (PLA 0.79, PRO 0.35; p
Individuals who developed URTI had higher degradation rates of Trp before exercise compared to those without URTI (Table 4
). Additionally, a running nose, but not cough was associated with higher Kyn/Trp ratios compared to those individuals without such symptoms.