Physical exercise is one of the major modulators of systemic metabolism. It increases the rate of metabolic processes and modulates the levels of different metabolites. A persistent, high exercise volume and intensity with limited recovery periods could lead to declined performance, and subsequent accelerated fatigability (Margonis et al., 2007). Various biological markers, hormonal markers and immunological markers have been used to evaluate the physical status within exhaustive exercise (Aguilo et al., 2005; Ferreira et al., 2018; Sarikaya et al., 2017; Zheng et al., 2018). However, these conventional methods, which involved measuring the concentrations of only a few biochemistry parameters or a few target metabolites in test samples, could not sensitively reflect all physiological differences and would be unable to explain the interaction between varieties of the metabolites (Taysi et al., 2008).
Metabolites are the final biological products of the system, which means that investigating the changes of these endogenous compounds can verify the response end-point of biological substances (Gao et al., 2014). Metabolomics strategy is a useful analytic platform to determine endogenous metabolites and assess the global and dynamic metabolic responses of living system (Nicholson and Lindon, 2008). Recently, metabolomics has attracted increasing interest in the field of sports medicine and shown great potential for monitoring the changes of physiological state. Huang et al. (2010) used metabolomics to determine the biochemical variations in male rats' liver with exhaustive and endurance exercises. Kume et al. (2015), using plasma metabolome analysis, identified potential biomarkers of fatigue in male rats. Miao et al. (2018) investigated the mechanism of the anti-fatigue effect of Danggui Buxue Tang on fatigue of male mice induced by forced swimming, and identified 14 metabolites providing evidence for the anti-fatigue effect. Jang et al. (2018) determined changes in urinary metabolic profiles and metabolomics markers in muscle damage following eccentric exercise in men and women, which clarify the metabolic response within eccentric exercise-induced muscle damage and gender-dependent patterns. By measurement of the endogenous metabolites, metabolomics investigations have the potential to distinguish the change of metabolic profiling and to identify the biomarkers associated with exercise performance, fatigue or exercise-induced disorders (Ra et al., 2014; Yan et al., 2009).
Although there have been several studies that examined the metabolic response to exercise, there is a lack of information on the gender-related differences in the metabolic response to exhaustive exercise. Therefore, in this study, the metabolomics study was used to distinguish the variations in metabolite profiles and determine biomarkers changed during exhaustive exercise and recovery in male and female rats.
Rat blood sampling
Sixty eight-week-old adult male and female Sprague-Dawley rats were randomly subdivided into control, exhaustive exercise and recovery groups consisting of 10 rats each. Before the commencement of the experiment, rats from exhaustive exercise and recovery group were familiarized to adaptive swimming training for 10 min/day for 3 days. Weight-loaded, forced swimming was performed according to the methodology described previously (Xu et al., 2013) with some modifications. The rats swam with a load of aluminum sheets that weighed 5% of their body weight and were attached to their tails. Rats swam individually until exhaustion once a day for 10 consecutive days; exhaustion was reached when the rat was unable to constantly keep its nose out of water and its nose remained below the water surface for 10 seconds (Ma et al., 2017). The recovery group was given a 7-day period to refresh after the exhaustive exercise period, while the rats in the control group were left in cages without swimming. The water temperature was controlled around 32-36 [degrees]C. The animals were housed at a controlled ambient temperature of 22-25 [degrees]C with 55-65% relative humidity and a 12 h light/12 h dark cycle (lights on at 7:00 AM) and were given food and water ad libitum at the Laboratory Animal Center in Shanghai University of Traditional Chinese Medicine, Shanghai, China. All animal procedures were performed in accordance with Institutional Animal Care and Use Committee guidelines of Shanghai University of Traditional Chinese Medicine. To investigate the metabolites generated in the system, several blood collections from each animal were collected immediately at day 5 and day 10 of swimming, as well as day 3 and day 7 post recovery.
Serum testosterone (T) and cortisol (CORT) were analyzed using specific assays strictly according to the manufacturer's instructions (Nanjing Jiancheng CO., China) and measured by the microplate reader (Biotek, USA). The malondialdehyde (MDA) assay kit (TBA method) was used to detect the MDA level as a marker of lipid peroxidation and the kit was bought from Nanjing Jiancheng CO., China. Blood urea nitrogen (BUN), creatine kinase (CK) and lactic dehydrogenase (LDH) were measured using an automated biochemistry analyzer (Hitachi, Japan). Plasma reduced glutathione (GSH) and oxidized glutathione (GSSG) were evaluated as markers of oxidative stress using specific assay kits from Nanjing Jiancheng CO., China and measured by the microplate reader (Biotek, USA).
4-Chlorophenyl-alanine, heptadecanoic acid, methoxyamine and N,O-bis(trimethylsilyl)-trifluoroacetamide (BSTFA) were purchased from Sigma-Aldrich (Sigma-Aldrich, USA). Methanol (HPLC Grade) and chloroform were obtained from Thermo Fisher (Thermo Fisher, USA). Pyridine was purchased from China National Pharmaceutical Group Corporation (China National Pharmaceutical Group Corporation, China). Pure water was produced by a Milli-Q purification system (Millipore, USA).
Sample preparation for GC-MS analysis
Plasma samples were treated with chemical derivatization following our previously published procedure (Liao et al., 2012). Each plasma sample (100 [micro]L) was added with two internal standard solutions (10 [micro]L of 4-chlorophenyl-alanine in water, 0.3 mg/mL; 10 [micro]L of heptadecanoic acid in methanol, 1 mg/mL) and with 300 [micro]L methanol/chloro-form (3:1) to extract the metabolites. After vortex mixing for 30 s, the mixtures were incubated at -20 [degrees]C for 10 min and centrifuged at 12,000 rpm for 10 min. A 300 [micro]L supernatant aliquot was transferred into a GC vial and evaporated to dryness under [N.sub.2] at 30 [degrees]C. Methoxyamine (80 [micro]L) in pyridine (15 mg/mL) was added to the dried residue and vortex mixed for 1 min. The methoximation reaction was carried out for 90 min while rocking in an air-shaker at 30 [degrees]C, followed by trimethylsilyl for 60 min by adding 80 [micro]L BSTFA at 70 [degrees]C. At last, the solution was vortex mixed 30 s and cooled into room temperature for GC-MS analysis.
Metabolite analysis by GC-MS
Each 1 [micro]L derivatized sample was injected onto a HP-5MS capillary column (30 m x 250 [micro]m inner diameter, 0.25 [micro]m film thickness, Agilent J&W Scientific, USA) on an Agilent 7890A GC/5975C MSD (Agilent J&W Scientific, USA). Helium was used as the carrier gas through the column with a constant flow rate of 1.0 mL/min. The sample was injected at 270 [degrees]C in splitless mode. The optimized GC-MS gradient temperature programming was selected following our previous experiment (Liao et al., 2012): the GC oven was started at 80 [degrees]C for 2 min, then the temperature was increased step-wise, starting at 10 [degrees]C /min to 140 [degrees]C, 4 [degrees]C/min to 210 [degrees]C, 10 [degrees]C /min to 240 [degrees]C, 25 [degrees]C/min to 290 [degrees]C and then maintained at 290 [degrees]C for 3 min. The ion source temperature and the quadrupole temperature were set at 230 [degrees]C and 150 [degrees]C, respectively. The mass data were acquired in scan (m/z 30-600) mode at a rate of 20 spectra/s with electron impact ionization (70 eV). The solvent delay time was set to 5.0 min.
The data from the GC-MS were converted into CDF formats, and the data were processed by the XCMS toolbox (http://metlin.scripps.edu/download/) to carry out baseline correction, peak deconvolution and alignment using XCMS's default settings. The result (CSV file) was exported into Microsoft Excel (Microsoft Inc., USA) where normalization was performed. The resulting data were analyzed in the SIMCA-P 11.0 Software (Umetrics, Umea, Sweden) for multivariate statistical analysis. The supervised orthogonal partial least-squares (OPLS) were employed to process the acquired data and to identify the general separation and cluster. Then the differential variables were selected based on a threshold of variable importance in the projection (VIP) value (VIP>1.0) from OPLS model. Subsequently, those variables were validated at a univariate level using the nonparametric Wilcoxon-Mann-Whitney test by SPSS 18.0 (SPSS, Chicago, IL, USA) with the p-value set at 0.05 (Liao et al., 2012). The corresponding fold change showed how these selected metabolites varied between groups in male and female rats. Additionally, compounds were identified by searching in NIST 2011 database. The Kyoto Encyclopedia of Genes and Genomes (KEGG) and Human Metabolome Database (HMDB) were used to give the biochemical interpretation of changed metabolites.
The biochemical parameters data were reported as mean [+ or -] Standard Deviations (SD). Statistical analysis was performed using SPSS 18.0 (SPSS Inc. Chicago, IL, USA). A two-way ANOVA was performed to examine the main effects of gender and exhaustive exercise and interaction on the measured variables. When a significant main effect or interaction was detected, data were subsequently analyzed by use of a post-hoc Bonferroni test. The...