The present study was conducted according to the policy of the Iranian convention for the protection of vertebrate animals used for scientific experimental purposes and the National Institutes for Health guide for the care and use of laboratory animals (NIH publications no. 80–23). In addition, the protocol of the present study was approved by the ethics committee of the university of Mazandaran medical sciences (Code: IR.UMZ.REC.1400231).

Twenty-eight Wistar female rats (6-8 weeks old, 100-120 g body mass) were maintained in the Animal House (12-hour light-dark cycle, temperature: 22 (°C) ± 1.4 (°C) in which the floor was covered with wood chips. Animal standard compressed food was provided at regular times. Water was provided in 500 ml bottles that was attached to the animal cages.

Animals were randomly assigned into groups including Saline-Control (SC) (n=7), Saline-High Intensity Training (ST) (n=7), Ziziphus jujuba-Control (ZJC) (n=7), and Ziziphus jujuba-High Intensity Training (ZJT) (n=7). Animals performed exercise on a treadmill (0% grade, 35 m/min, 60 min/day, for 5 days/week), or administered supplements intragastrically (ZJ or saline solution equivalent to one mL/100g of body mass) for 6 weeks, depending on the group category. Supplements were administered intragastrically immediately after each exercise session or at the corresponding rest time.

A high dose and chronic consumption of ZJ can reduce sex potential and libido in males (20) so we decided to examine female rats only.

Vaginal lavage observation was used to determine the estrous cycle each morning using a microscope. Female rats showing at least two consecutive 4- or 5-day estrous cycles were used in the present study (21).

Dried Ziziphus jujuba fruit was obtained from the main market of Birjand city (South Khorasan province, Iran), cleaned, washed with tap water, re-dried for 3 days in lab oven at 40°C. The herbalist of herbarium collection of the Biology department of Mazandaran University, Babolsar, Mazandaran, Iran identified the plant material. The seeds were removed and powdered by a house blender and the extraction was prepared according to Cacige et al. (22). Briefly, 40 mg of the powdered Ziziphus jujuba extracted with 600 ml water for 70 h at 10°C, followed by filtration. The final extract was 400 ml.

These procedures were chosen to prevent any possible increase in 5-hydroxymethylfurfural (HMF) which has been recognized as an end product of carbohydrate degradation when heating for Ziziphus jujuba extraction at different and high temperatures (23). According to Cacig et al. (22) who noticed that the preparation of Ziziphus jujuba extraction by maceration method at 10°C and four 70 hours did not show any mold growing as a result of fermentation.

Rats were orally gavaged by a liquid crude Ziziphus extract at dose 1g/kg of body weight (BW) (100 mg/100 g of BW) in 10 ml/kg of BW. The saline group was treated with the same volume of the normal saline as Ziziphus treated rats.

The rats were familiarized with the treadmill training protocol for 4 days prior to the study commencing after which there was a wash out period of 12 hours. The exercise groups were trained with a fixed intensity programme of treadmill running at 35 m/min for 60 minutes, 5 d/week for 6 weeks (9).

Seventy-two hours after the last training session, rats were anesthetized with intraperitoneal administration of ketamine (30– 50 mg/kg body mass) and xylazine (3– 5 mg/kg body mass). Three hours before sacrifice, food but not water was removed. Following sacrifice blood was collected from the inferior vena cava close to the heart into EDTA test tubes. The test tubes were centrifuged for 10 min at 3000 rpm for plasma preparation. Hypothalamus tissue, heart and gastrocnemius muscles were excised, cleaned and washed with normal saline. All plasma and tissue samples were frozen in liquid nitrogen and stored at − 80 (°C) until laboratory analyses.

A direct Immune method was used to determine plasma HDL concentration (HDL-C Immuno FS, Pars Azmoun, Tehran, Iran; Intraassay coefficient of variation: 1.2%, Interassay coefficient of variation sensitivity: 1.8%, sensitivity: 0.03 mmol/L). LDL-C was obtained according to the manufacturers’ protocol (an enzymatic method; kit was purchased from Pars Azma Iranian Com) (Intraassay coefficient of variation: 0.67%, Interassay coefficient of variation sensitivity: 1.45%, sensitivity: 2 mg/dL). Plasma insulin concentration was determined using a sandwich ELISA kit from ALPCO (Intraassay coefficient of variation: 3.4%, Interassay coefficient of variation sensitivity: 4.3%, sensitivity: 0.5 μIU/mL). An enzymatic (GPO, Glycerol-3-Phosphate Oxidase) colorimetric method (Pars Azmoun, Tehran, Iran; intra-assay coefficient of variation: 1.47%, Interassay coefficient of variation sensitivity: 1.06%, sensitivity: 1 mg/dL) was used to determine plasma total Triglyceride (TG). An enzymatic (CHOD-PAP, Cholesterol Oxidase-Amino Antipyrine) colorimetric method (Pars Azmoun, Tehran, Iran; intra-assay coefficient of variation: 1.9%, Interassay coefficient of variation sensitivity: 0.93%, sensitivity: 0.08 mmol/L) was used to determine plasma total cholesterol (TC). An ELISA method (Canada Inc, Ontario, Canada Estradiol ELISA, Diagnostics Biochem, intra-assay coefficient of variation: 4.6%, Interassay coefficient of variation sensitivity: 6.2%, sensitivity: 10 Pg/mL) was used to determine plasma estrogen concentration. Plasma cortisol also measured using Diagnostics Biochem ELISA kit (intra-assay coefficient of variation: 2.9%, Interassay coefficient of variation sensitivity: 3.8%, sensitivity: 0.4 μg/dL).

Tissue nesfatin-1 concentration was measured using a Rat Nesfatin-1 ELISA Kit (CUSABIO, Catalog No, CSB-E 1478r, China; sensitivity: 3.9 pg/mL, Intraassay: 7.5% and interassay <10%). A Bioluminescence method was used to determine tissue ATP concentration by special kit (BioVision Incorporated 155 S. Milpitas Boulevard, Milpitas, CA 95035 USA; quantitation range was approximately between 1 nmol and 10 fmol/assay). A Double Beam UV Spectrophotometer (Cecil Elegant Technology CE-5501 computing Cambridge England) was used to determine tissue glycogen concentration.

To measure the tissue NUCB2 gene expression, RNA was extracted by a specific kit (Cat. No. k3090, Bioneer, Daejeon, Republic of Korea) and all hypothalamic tissue and almost 100-mg of heart and gastrocnemius muscle were used. After that measure the purity of extracted RNA by spectrophotometer, cDNA was manufactured using AccuPower RT PreMix cDNA synthesis kit, Bioneer. The cDNA was kept at −20°C for use in real-time PCR. The real-time PCR was performed using Qiagen SYBR Green kit (one μl, Cat. No. 204052, Qiagen GmbH, Hilden, Germany), cDNA (1 μl) RNAse-free water (5 μl), forward and reverse primers (1 μl of each one) and using Australian Corbett’s machine. All stages were according to the Rahmati-Ahmadabad et al. article (24). The Threshold Cycle (CT) was recorded and converted to relative gene expression data using the 2^{^- ΔΔCT} method (24, 25). Unique primers for the real-time PCR stage were NUCB2 and GAPDH. The NUCB2 sense and antisense primers were 5’- TTTGAACACCTGAACCACCA-3’ and 5′- TGCAAACTTGGCTTCTTCCT-3′, respectively. The GAPDH sense and antisense primers were 5’- GTGCCAGCCTCGTCTCATAG-3’ and 5′- GACTGTGCCGTTGAACTTGC-3′, respectively.

All statistical analysis was undertaken using SPSS software, version 24 (SPSS, Inc., Chicago, IL). Descriptive statistics were used to categorize and determine the dispersion indices. The normality of data was determined using the Shapiro-Wilk test. Two-way ANOVA and the least significant difference (LSD) test was used to determine the significant changes between groups. Repeated measures two-way analysis of variance was used to evaluate the effects of time, group and time×group. If Mauchley’s test of sphericity was significant, Greenhouse–Geisser correction was used. Partial eta-squared ( η p 2 ) is reported to emphasize the size of the difference rather than confound the sample size with 0.01, 0.03, and >0.05 considered small, medium, and large effects, respectively. Moreover, non-centrality parameter (NCP) and observed power/post-hoc power (OPW) were reported. The Pearson Product Moment correlation was used to determine relationships between variables of interest. Because of the low sample size, non-parametric tests including the Freidman test and Spearman correlation were still performed, but this did not alter the interpretation of the findings so only the results of the parametric tests are presented. Statistical significance was accepted as P ≤ 0.05.

The results showed that the rats’ body weight significantly changed between groups over time (time×group effect, F _2.02,48.570 = 3.493, P=0.038, η p 2 = 0.127, NCP=7.070 and OPW=0.629, Figure 1 ). The rats body weight was significantly lower in trained rats compared to non-trained rats in began of 4^th (F_1,24 = 8.277, P=0.008, η p 2 = 0.256, NCP=8.277 and OPW=0.788), 5^th (F_1,24 = 7.910, P=0.01, η p 2 = 0.248, NCP=7.910 and OPW=0.770) and 6^th (F_1,24 = 4.105, P=0.05, η p 2 = 0.146, NCP=4.105 and OPW=0.494) week ( Figure 1 ). Exercise, supplementation and interaction had no significant effect on the mean of 7 repetitions of weight.

NUCB2 gene expression was significantly higher in the hypothalamus tissue (F_1,24 = 52.622, P=0.001, η p 2 = 0.687, NCP=52.622 and OPW=1, Figure 2A ), heart (F_1,24 = 13.534, P=0.001, η p 2 = 0.361, NCP=13.534 and OPW=0.942, Figure 2B ) and gastrocnemius (F_1,24 = 11.717, P=0.002, η p 2 = 0.328, NCP=11.717 and OPW=0.907, Figure 2C ) muscles of trained rats compared to their control. NUCB2 gene expression was significantly higher in the hypothalamus tissue (F_1,24 = 111.890, P=0.001, η p 2 = 0.823, NCP=111.890 and OPW=1, Figure 2A ), heart (F_1,24 = 33.103, P=0.001, η p 2 = 0.580, NCP=33.103 and OPW=1, Figure 2B ) and gastrocnemius (F_1,24 = 16.197, P=0.001, η p 2 = 0.403, NCP=16.197 and OPW=0.971, Figure 2C ) muscles of ZJ rats compared to their control. Interaction of training and ZJ showed a synergic effect regarding hypothalamus (F_1,24 = 4.771, P=0.039, η p 2 = 0.166, NCP=4.771 and OPW=0.554, Figure 2A ) and heart (F_1,24 = 5.014, P=0.035, η p 2 = 0.173, NCP=5.014 and OPW=0.575, Figure 2B ) NUCB2 gene expression. Gastrocnemius NUCB2 gene expression was significantly higher in ZJT group compared to ST and ZJ groups (P=0.004 and P=0.011, respectively, Figure 2C ).

Heart nesfatin-1 concentration was significantly higher in trained rats compared to their control (F_1,24 = 9.375, P=0.005 and η p 2 = 0.281, NCP=9.375 and OPW=0.836). The concentration of nesfatin-1 was also significantly higher in ZJ treated heart compared to saline treated rats (F_1,24 = 33.337, P=0.001 and η p 2 = 0.581, NCP=33.337 and OPW=1). In addition, the combination of training and supplementation showed a synergetic increase in heart nesfatin-1 concentration (F_1,24 = 12.245, P=0.002 and η p 2 = 0.338, NCP=12.245 and OPW=0.919) ( Figure 3A ). Heart and gastrocnemius muscle nesfatin-1 responded differently in that significantly higher nesfatin-1 concentrations were observed in ST (P=0.015) and ZJC rats (P=0.043), but exercise did not significantly increase or decrease nesfatin-1 concentration in ZJT ( Figure 3B ). However, in gastrocnemius muscle, nesfatin-1 concentrations were significantly higher in ST (P=0.015) and ZJC (P=0.043) when compared to SC rats and a decrease in nesfatin-1 concentration was observed in the ZJT ( Figure 3B ).

Exercise induced significant increases in heart ATP concentration (F_1,24 = 6.864, P=0.015 and η p 2 = 0.222, NCP=6.864 and OPW=0.710). Heart ATP was significantly higher in ST (P=0.004), ZJC (P=0.047), and ZJT (P=0.05) groups when compared to SC. Data also indicated that ZJT heart had significantly higher ATP concentration than SC heart (P=0.014) ( Figure 4A ). Similar changes were observed in ST, ZJC and, ZJT gastrocnemius muscle ATP concentration ( Figure 4B ). However, a higher and significant ATP concentration was only observed in the ST group (P=0.01) ( Figure 4B ).

Aerobic training had no significant effect on heart (F_1,24 = 1.148, P=0.295, η p 2 = 0.083, NCP=2.167 and OPW=0.293, Figure 5A ) and gastrocnemius (F_1,24 = 0.798, P=0.381, η p 2 = 0.034, NCP=0.798 and OPW=0.137, Figure 5B ) muscle glycogen concentration. ZJ supplementation did not significantly change heart (F_1,24 = 2.167, P=0.154, η p 2 = 0.046, NCP=1.148 and OPW=0.177 Figure 5A ) and gastrocnemius (F_1,24 = 0.416, P=0.546, η p 2 = 0.018, NCP=0.416 and OPW=0.095, Figure 5B ) glycogen concentration. There was no significant interaction between training and supplement use on heart (F_1,24 = 0.059, P=0.810, η p 2 = 0.002, NCP=0.059 and OPW=0.056, Figure 5A ) and gastrocnemius (F_1,24 = 0.471, P=0.499, η p 2 = 0.020, NCP=0.471 and OPW=0.101, Figure 5B ) glycogen concentration.

Plasma insulin concentration was higher in ST than CS (P=0.02). ZJ use led to no significant changes in plasma insulin concentration (F_1,24 = 0.733, P=0.40, η p 2 = 0.030, NCP=0.733 and OPW=0.130). The interaction between training and supplement use showed a significant increase in plasma insulin concentration (F_1,24 = 5.949, P=0.022, η p 2 = 0.199, NCP=5.949 and OPW=0.648) ( Figure 6A ).

Aerobic training showed a significant decrease in plasma cortisol concentration (F_1,24 = 11.852, P=0.002, ηp² = 0.331, NCP=11.852 and OPW=0.911). ZJ use showed a significant decrease in plasma cortisol concentration (F_1,24 = 9.439, P=0.005 and η p 2 = 0.282, NCP=1.264 and OPW=0.190). Plasma cortisol concentration was lower in ZJT than ST (P=0.007) ( Figure 6B ).

Aerobic training led to a significant increase in plasma HDL (F_1,24 = 19.990, P=0.001, η p 2 = 0.454, NCP=19.990 and OPW=0.990, Figure 6C ) and significant decrease in plasma TG (F_1,24 = 4.117, P=0.05, η p 2 = 0.161, NCP=4.117 and OPW=0.495, Figure 6D ) concentration. ZJ supplementation led to a significant increase in plasma HDL concentration (F_1,24 = 16.903, P=0.001 and η p 2 = 0.413, NCP=19.903 and OPW=0.976). Plasma HDL concentration was higher in ZJT than ST (P=0.001) ( Figure 6C ).

Aerobic training had no significant effect on plasma LDL (F_1,24 = 1.475, P=0.236, η p 2 = 0.058, NCP=1.475 and OPW=0.215, Figure 6E ), TC (F_1,24 = 0.726, P=0.403, η p 2 = 0.029, NCP=0.726 and OPW=0.130, Figure 6F ) and estradiol (F_1,24 = 2.394, P=0.135, η p 2 = 0.091, NCP=2.394 and OPW=0.318, Figure 6G ) concentrations. ZJ supplementation had no significant effect on plasma TG (F_1,24 = 1.572, P=0.222, η p 2 = 0.061, NCP=1.572 and OPW=0.226, Figure 6D ), LDL (F_1,24 = 0.001, P=0.997, η p 2 = 0.001, NCP=0.001 and OPW=0.050, Figure 6E ), TC (F_1,24 = 0.297, P=0.591, η p 2 = 0.012, NCP=0.297 and OPW=0.082, Figure 6F ) and estradiol (F_1,24 = 2.385, P=0.136, η p 2 = 0.090, NCP=2.385 and OPW=0.317, Figure 6G ) concentrations. There was no significant interaction between training and supplement use on plasma TG (F_1,24 = 0.001, P=0.981, η p 2 = 0.001, NCP=0.001 and OPW=0.050, Figure 6D ), LDL (F_1,24 = 0.002, P=0.964, η p 2 = 0.001, NCP=0.002 and OPW=0.050, Figure 6E ), TC (F_1,24 = 0.072, P=0.790, and η p 2 = 0.003, NCP=0.072 and OPW=0.058, Figure 6F ) and estradiol (F_1,24 = 0.001, P=0.969 and η p 2 = 0.001, NCP=0.001 and OPW=0.050, Figure 6G ) concentrations.

There were positive and significant correlations between hypothalamus gene expression with heart, nesfatin-1gene/concentration, gastrocnemius nesfatin-1 gene expression, plasma HDL and estradiol concentration, while it had negative and significant correlations with plasma cortisol and TG concentrations. Nesfatin-1 gene expression had positive and significant correlations with heart nesfatin-1 and plasma HDL concentration, while it had negative and significant correlations with heart glycogen and plasma cortisol concentration. Gastrocnemius nesfatin-1 gene expression had a positive and significant correlation with plasma HDL concentration and had negative and significant correlations with plasma cortisol and TG concentration. All correlations between key variables are shown in Table 1 .