Twelve healthy untrained young males with no exercise habits (mean ± SD: age 21.7 ± 1.6 years; height 174.6 ± 3.6 cm; body weight 65.6 ± 5.2 kg; \(}}\)O2max 41.4 ± 4.3 ml/kg/min; maximal heart rate (HRmax) 197.6 ± 5.3 beats/min) participated in this study. In this study, individuals were excluded if, in the last year, they had physiological disorders or chronic diseases, were currently taking any medications, smoked or consumed alcohol daily, or exercised more than 3 h/week. All participants were given verbal and written briefings on the study and provided written informed consent. This study was approved by the Ethics Committee of Ritsumeikan University (BKC-LSMH-2021–054) and conducted in accordance with the Declaration of Helsinki.

\(}}\)O2max was measured approximately 1 week before the main trials. The cycling exercise load for each exercise trial was calculated from the relationship between \(}}\)O2max and work rate, equivalent to 35% \(}}\)O2max, 55% \(}}\)O2max, and 75% \(}}\)O2max. Subjects were instructed not to perform excessive exercise and not to consume both caffeine and alcohol for 24 h before each exercise trial. The subjects performed each exercise trial after fasting and drinking water from 10:00 PM. Hydration status was standardized to 500 ml before each exercise trial (Li and Gleeson 2005). The subjects were instructed to drink 250 ml water before bedtime and after waking up. To eliminate the effect of the circadian rhythm on the saliva flow rate and s-IgA concentration (Hucklebridge et al. 1998; Dimitriou et al. 2002), each measurement was conducted from 9:00 AM to 12:00 PM. The participants arrived at the laboratory at 9:00 AM and remained in a sitting position for 30 min before each exercise trial (Fig. 1). Body weight was measured using a digital platform scale (Innerscan DUAL; TANITA, Tokyo, Japan), and hydration status was checked at rest before exercise; it was confirmed that there was no change between each exercise trial within the subjects (Walsh et al. 2004). This was a randomized crossover study. After each exercise trial, the subjects were allowed to rest in a sitting position for 60 min. The heart rate (HR) (WEP-5204; Nihon Kohden, Tokyo, Japan) was measured 1 min before the end of exercise. Saliva samples were collected before (baseline), immediately after (post), and 60 min after the end of exercise (post-60) according to a previous study (Allgrove et al. 2008). The participants were instructed to fast and not drink anything until 60 min after the end of the exercise. The laboratory temperature and relative humidity were 23.1 ± 1.5 °C and 71.2 ± 12.1%, respectively.

Experimental protocol in this study. Experiment 1: Trial 1 (cycling exercise for 30 min at 35% \(}}\)O2max), Trial 2 (cycling exercise for 30 min at 55% \(}}\)O2max), Trial 3 (cycling exercise for 30 min at 75% \(}}\)O2max). Experiment 2: Trial 2 (cycling exercise for 30 min at 55% \(}}\)O2max), Trial 4 (cycling exercise for 60 min at 55% \(}}\)O2max), Trial 5 (cycling exercise for 90 min at 55% \(}}\)O2max). \(}}\)O2max: maximal oxygen uptake

All subjects randomly performed five exercise trials (Trial 1: cycling exercise for 30 min at 35% \(}}\)O2max, Trial 2: cycling exercise for 30 min at 55% \(}}\)O2max, Trial 3: cycling exercise for 30 min at 75% \(}}\)O2max, Trial 4: moderate-intensity continuous cycling exercise at 55% \(}}\)O2max for 60 min, Trial 5: moderate-intensity continuous cycling exercise at 55% \(}}\)O2max for 90 min) with a washout period of at least 3 days between each trial as a reference to a previous study (Allgrove et al. 2008). To determine the effect of exercise intensity on salivary parameters (Experiment 1), continuous cycling exercise at three different exercise intensities (35% [low], 55% [moderate], and 75% \(}}\)O2max [high]) was performed for 30 min. Additionally, to determine the effect of exercise duration on salivary parameters (Experiment 2), continuous cycling exercise at 55% \(}}\)O2max as a moderate-intensity exercise at three different exercise durations (30 min [short], 60 min [medium], and 90 min [long]) was performed. In this study, considering the burden on the participant, 55% \(}}\)O2max for 30 min was performed only once and the duplicate was not implemented in either experiment. Furthermore, it was divided into 30-min intervals, and exercise trials were set for 30 min (short duration), 60 min (medium duration), and 90 min (long duration).

\(}}\)O2max was determined during an incremental cycling exercise test using a cycle ergometer (828E; Monark, Stockholm, Sweden) by monitoring breath-by-breath oxygen consumption and carbon dioxide production (AE-310SRD; Minato, Osaka, Japan). The subjects were instructed to maintain a minimum pedaling rate of 60 rpm according to the protocol of a previous study (Hasegawa et al. 2016). Following a 5-min warm-up at 60 W, they began cycling at 60 W ± 30 W in increments of 15 W each minute until exhaustion (Hasegawa et al. 2016). During the incremental cycling exercise test, the heart rate (HR) (WEP-5204; Nihon Kohden, Tokyo, Japan) and rating of perceived exertion (RPE) were continuously measured every minute during the cycling exercise. The test was considered valid if at least three of the four following criteria were met: [1] plateau in \(}}\)O2 with an increase in maximal effort, [2] HRmax of the age-predicted maximum (220-age ± 5 beats/min), [3] maximal respiratory exchange ratio of ≧1.1, and [4] an RPE of ≧18 (Borg 1982; American College of Sports Medicine 2000).

Saliva samples were collected between 9:00 AM and 1:00 PM. According to a previous study (Usui et al. 2011), the subjects rinsed their mouths with distilled water (30 s × 3 times). Saliva production was stimulated by chewing a piece of paraffin wax (B.S.A paraffin wax; B.S.A, Aichi, Japan) for 1 min at a frequency of 1 chew/sec (Libicz et al. 2006; Balsalobre-Fernández et al. 2014; Gomar-Vercher et al. 2018). The collected saliva was separated from the paraffin wax by centrifugation at 4 °C, 1500 g, for 5 min. The saliva flow rate (ml/min) was measured by weighing and the saliva density was estimated to be 1.0 g/ml (Usui et al. 2011). Saliva supernatants were stored at − 20 °C until analysis.

Salivary IgA and cortisol concentrations were measured using an enzyme-linked immunosorbent assay (Salimetrics, State College, PA, USA) according to previous studies (Mc Naughton et al. 2006; Peñailillo et al. 2015). The absorbance was measured at 450 nm using a microplate reader and an xMark microplate spectrophotometer (Bio-Rad Laboratories, Hercules, CA, USA). The salivary IgA secretion rate (μg/min) was calculated from the saliva flow rate per minute (ml/min) and salivary IgA concentration (μg/ml), according to a previous study (Koch et al. 2007). In this study, salivary cortisol concentrations were measured based on the assumption that the concentration of cortisol in the saliva reflects the concentration secreted into the blood (Chicharro et al. 1998; Hackney and Walz 2013). The average coefficients of variation for IgA and cortisol assays were 3.3% and 2.6%, respectively.

All statistical analyses were performed using the StatView software (5.0; SAS Institute, Tokyo, Japan). All values are expressed as mean ± SD. In the Experiment 1, the saliva flow rate, s-IgA concentration, salivary cortisol concentration, and s-IgA secretion rate at different exercise intensities (35%, 55%, and 75% \(}}\)O2max) were compared using three trials × three time points (baseline, post, and post-60) of a two-way repeated-measures ANOVA. Fisher’s post-hoc test was used to correct for multiple comparisons when the ANOVA revealed significant differences. The relationship between the change in saliva flow rate and the change in salivary cortisol concentration was determined using Pearson’s correlation coefficient. In the Experiment 2, the comparison of saliva flow rate, s-IgA concentration, salivary cortisol concentration, and s-IgA secretion rate at different exercise durations (30, 60, and 90 min) was performed using three trials × three time points (baseline, post, and post-60) in a two-way repeated-measures ANOVA. Fisher’s post-hoc test was used to correct for multiple comparisons when the ANOVA revealed significant differences. Statistical significance was defined as P < 0.05. We calculated the effect size of 0.25 (large) for the two-way repeated-measures ANOVA using G*Power (version 3.1), which determined the sample size needed for this study. To detect the effect size at alpha levels of 0.05 and 80% power, the sample size was set to 12 participants.