Effects of oral cystine and glutamine on exercise-induced changes in gastrointestinal permeability and damage markers in young men

After approval from the Institutional Ethical Advisory Committee (Approval number: 2018–270), 16 young, Japanese men gave written informed consent to participate in this study. This study was registered in advance using the University Hospital Medical Information Network Center, a system for registering clinical trials (ID: UMIN000026008). Exclusion criteria were as fllows: being a current smoker; taking any medication or supplement known to affect lipid and carbohydrate metabolism; major illness, including gastrointestinal disorders or physical problems (acute or chronic), limiting the ability to perform physical activity. Habitual information regarding physical activity were obtained via interviews, while additional questionnaires indicated that all participants met the current international public health guidelines for physical activity (i.e., moderate-intensity physical activity for at least 150 min per week, vigorous-intensity physical activity for 75 min per week or an equivalent combination of moderate and vigorous-intensity physical activity) [14]. The physical characteristics of the participants (mean ± standard deviation (SD)) were as follows: age, 22.7 ± 2.6 years; height, 1.71 ± 0.05 m; body mass, 67.0 ± 8.0 kg; body mass index, 22.8 ± 2.5 kg/m²; and maximum oxygen uptake (measured while running on a treadmill) 54.3 ± 8.8 mL/kg/min.

Body mass was measured to the nearest 0.1 kg using a digital scale (Inner Scan 50, Tanita Corporation, Tokyo, Japan). Height was measured to the nearest 0.1 cm using a stadiometer (YS-OA, AS One Corporation, Osaka, Japan). Body mass index was calculated as weight in kilogrammes divided by the square of height in metres.

Approximately 14 days before the first main trial, participants underwent two preliminary exercise tests performed on a motorized treadmill (Jog Now 700, Technogym, Cesena, Italy). A 16-min, four-stage, submaximal treadmill test was conducted to determine the relationship between running speed and oxygen uptake. Initial running speed was set between 6.0 and 8.0 km/h depending on each participant’s physical activity level (as determined by the initial screening interview and questionnaire). The treadmill was level throughout the test period, and speed was increased by 1.0 or 1.5 km/h every 4 min, depending on each participant’s physical activity level. Next, maximum oxygen uptake was directly measured using an incremental uphill protocol at a constant speed, until participants reached volitional fatigue [15]. The initial inclination of the treadmill was set at 3.5% for the test. Thereafter, the gradient was increased by 2.5% every 3 min. Heart rate was monitored throughout these tests using short-range telemetry (Polar RCX3, Polar Electro, Kempele, Finland). Oxygen uptake, carbon dioxide production and respiratory exchange ratio were measured using a stationary gas analyzer (Quark RMR, COSMED, Rome, Italy). Ratings of perceived exertion were periodically assessed during tests using the Borg scale [16]. Data generated from these two tests were used to determine the running workload (i.e., running speed) corresponding to 75% of each participant’s maximum oxygen uptake, and this workload was used for the main trials.

Seven days after completing the preliminary exercise tests, one baseline gastrointestinal permeability evaluation was performed (further details are provided under “Evaluation of gastrointestinal permeability”). This baseline evaluation was conducted in a resting state (i.e., without performing exercise) throughout the 6-h evaluation period.

A randomized, double-blind, placebo-controlled crossover design was used in the present study. The primary outcomes of the present study were plasma L:M and I-FABP concentrations. Each participant underwent two, 6-day trials in a randomized, counterbalanced order: placebo or cystine and glutamine supplementation. The interval between trials was 14–22 days. A schematic illustration of the study protocol is presented in Fig. 1. Participants ingested 0.23 g of l-cystine, 1.00 g of L-glutamine and 1.23 g of maltodextrin three times (i.e., at 1000, 1500 and before sleep) per day for 5 days or 2.46 g of maltodextrin three times per day (i.e., at 1000, 1500 and before sleep) for 5 days. This glutamine dose and regimen were chosen since previous studies observed a suppressive effect of glutamine on oxidative stress [17] and no adverse effects in healthy adults when the individuals consumed a higher dose of glutamine than the present study [6]. Also, this cystine dose and regimen were chosen because of its high safety profile and no adverse effects in healthy adults (i.e., 2100 mg/day for 4 weeks) [18]. Participants were instructed to maintain their current physical activity level during the 5-day supplementation period and to refrain from taking any medications or supplements on days 3, 4 and 5. Participants were also instructed to refrain from exercising and to consume three standardized meals (i.e., provided by investigators) on day 5. Furthermore, all participants were informed of the foodstuffs (fruits, fruit juice, dairy products, chocolate, honey, mushrooms and chewing gum) that affect the gastrointestinal permeability evaluation before the main trials and they were asked to avoid consuming these food items on day 5 in the present study. On the morning of day 6, participants reported to the laboratory at 0845 after a 12-h overnight fast (no food or drink after 2100 on day 5) in both trials. No drinking water was permitted after 0645 on the morning of day 6 in either trial. Drinking water and a standardized meal was provided to each participant during a 6-h laboratory-based experiment on day 6 and this was standardized between trials as it affects gastrointestinal permeability evaluations (further details are provided under “Evaluation of gastrointestinal permeability”). Body mass was measured, and a rectal probe was inserted to measure the rectal temperature during exercise using a thermistor (401 J, Nikkiso-Therm Co., Ltd., Tokyo, Japan). After 15 min, a fasting venous blood sample was collected in a seated position by venepuncture at 0900. Further venous blood samples were collected immediately post-exercise (1045), and 30 min (1115), 1 h (1145), 2 h (1245) and 4 h (1445) post-exercise. 15 min after collecting fasting blood samples (at 0915), participants remained seated in a chair and ingested their designated supplements [i.e., placebo (2.46 g of maltodextrin) or cystine and glutamine (0.23 g of L-cystine, 1.00 g of l-glutamine, and 1.23 g of maltodextrin)] with 200 mL of warm water (37 °C). After 30 min of rest, participants performed running exercises at a speed corresponding to 75% of maximal oxygen uptake for 1 h. The atmospheric temperature and relative humidity during the 1-h exercise session was maintained at 25 °C and 60%, respectively, in both trials. Heart rate was monitored throughout exercise using short-range telemetry (Polar RCX3, Polar Electro, Kempele, Finland), and ratings of perceived exertion were periodically assessed [16]. Thereafter, participants were asked to remain seated (reading, writing, or working on a computer) in the laboratory until 1445. The mean temperature and humidity during the experimental trials were 25.4 ± 1.1 °C and 56.3 ± 3.9% (mean ± SD), respectively.

Gastrointestinal permeability was evaluated by analysing the plasma samples using the liquid chromatography-tandem mass spectrometry method based on previously published methods [19]. Participants consumed a standardized, dual-sugar solution containing 10 g of lactulose and 5 g of mannitol in 100 mL of warm water (37 °C) and 200 mL of warm water (37 °C) 20 min after the start of running in each exercise trial in the main experiment (1005) or rest (i.e., the baseline gastrointestinal permeability evaluation) in the preliminary test. Participants were asked to ingest 200 mL of warm water (37 °C) immediately post-exercise (1045) and 1 h post-exercise (1145). Thereafter, participants were asked to ingest 500 mL of room temperature water (25 °C), and this volume was asked to be consumed until the end of each main trial (1445). The timing and volumes of ingestion mimicked the baseline (i.e., resting) gastrointestinal permeability evaluation conducted in the preliminary test. The plasma samples were deproteinized with acetonitrile after adding stable isotope labels as internal standard substances. The prepared samples were quantitated using ultra-high performance liquid chromatography with electrospray ionization tandem mass spectrometry (Nexera, Shimadzu Co., Kyoto, Japan; QTRAP 5500, AB Sciex Pte. Ltd., Tokyo, Japan). The analysis method was validated concerning the bioanalytical method validation guidelines [20]. The intra-assay coefficients of variation were 9.3% for lactulose and 10.4% for mannitol. The validity of the measured values was confirmed by the quality control sample, and the criteria were 100 ± 15%.

Plasma I-FABP concentrations were determined using EDTA plasma via enzyme-linked immunosorbent assay kits (HK406, Hycult Biotech Inc., Pennsylvania, USA). All analyses for each participant were completed within the same run for each measure. The intra-assay coefficient of variation was 4.4%. We planned to analyze absolute values. However, since there was a large inter-participant variability in plasma I-FABP in the present study as was the case in a previous study [21], we presented plasma I-FABP concentration as changes compared with fasting (0900) values.

The sample size was set to 14 based on our pilot study and a previous study [6]. The validity of this sample size was examined by calculating the statistical power. The sample size was validated using L:M as the primary study outcome, using the same sample size calculation methods used in a previous crossover study [6]. Analyses were performed using G*Power 3.1.0 [22]. Based on a pilot study with ten participants, we calculated that an estimated total sample size of 14 was needed to provide 88% power to detect between-trial differences for two trials, with an alpha level set at 0.05 and a correlation of 0.5. This sample size estimation was powered to detect an effect size of 0.91 (Cohen’s d), using a paired t-test for comparison between trials. We also calculated the required sample size based on previous data [6]. The study reported within-subject effects (trial: glutamine vs. placebo; effect size, Cohen’s d = 0.83 for the plasma lactulose to rhamnose ratio) using glutamine versus the placebo response to a 60-min treadmill run [6]. The sample size was calculated to detect an effect size of 0.83 using a paired t-test for comparison between trials. For two trials with an alpha level set at 0.05 and a correlation of 0.5, an estimated total sample size of 14 would provide 82% power to detect between-trial differences. Based on these calculations, we recruited 16 participants to allow for potential withdrawals. Data were analyzed using R (version 3.4.3; R Foundation for Statistical Computing, Vienna, Austria), while linear mixed models were analyzed using the lme4 and lmerTest libraries in R. The full analysis set was based on the intention-to-treat concept. A linear mixed model for repeated measures [trial (placebo or amino acids), time (treated as categorical with levels at 0, 30, 60, 120 and 240 min post-exercise], factors related to a crossover study design (i.e., term and sequence) as fixed factors, and participants as the random factor were used for each dependent variable. Statistical significance was accepted at a two-sided 5% level. Results are reported as mean ± SD. However, in the linear mixed models for longitudinal data, the results for the trial are reported as coefficient ± standard error (SE), with the notation “coefficient ± SE” to distinguish them from the mean ± SD.

Serum TG, NEFA, TC, LDL-C and HDL-C and plasma glucose concentrations measured on the morning of day 6 for each trial, are shown in Table 1. The fasting concentrations of all parameters did not differ between the trials.

No significant differences were observed between the placebo, and cystine and glutamine supplementation trials regarding mean heart rate (placebo: 157 ± 12 beats/min; cystine and glutamine: 158 ± 13 beats/min, linear mixed model; P = 0.151), relative exercise intensity (placebo: 78.3 ± 5.8% of maximum oxygen uptake; cystine and glutamine: 78.9 ± 5.7% of maximum oxygen uptake, linear mixed model; P = 0.168), or peak rectal temperature (placebo 38.3 ± 0.8 °C; cystine and glutamine: 38.4 ± 0.5 °C, linear mixed model; P = 0.686).

The absolute plasma L:M ratio is shown in Fig. 2A. Plasma L:M was lower in the cystine and glutamine supplementation trial than in the placebo trial (linear mixed model, coefficient ± SE; − 0.011 ± 0.004, P = 0.0090). However, increased plasma L:M in the cystine and glutamine supplementation trial was observed at 120 min and 240 min after exercise. Given that this trend was also observed in the resting state (i.e., the baseline gastrointestinal permeability evaluation), an additional analysis was performed using the delta values (i.e., the change from the value in the resting state). In this additional analysis, plasma L:M was lower in the cystine and glutamine supplementation trials compared to that in the placebo trial (linear mixed model, coefficient ± SE; -0.011 ± 0.004, P = 0.0096) (Fig. 2B).

Delta [by subtracting the value at each time point from the baseline value (i.e., 0900)] plasma I-FABP concentrations are shown in Fig. 3. Plasma I-FABP concentrations were lower in the cystine and glutamine supplementation trial than in the placebo trial (linear mixed model, coefficient ± SE [pg/mL]; − 195.3 ± 65.7, P = 0.0035).