A total of 30 8-week-old healthy male C57Bl/6J mice (body weight, 20.0±1.0 g) were housed under 12-h light/dark cycles with regulated temperature (18–25°C) and humidity (55–65%). Mice were purchased from the Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All mice received ad libitum access to food and water, and were fed with regular chow for 2 weeks to adapt to the environment. Contents of the regular chow were 21% protein, 55% carbohydrate and 6% fat, with a total energy of 15.36 kJ/g. Following the initial 2 weeks, mice were divided randomly into the NC group (n=10), fed with regular chow throughout the study (24 weeks); the obesity group (OB; n=10), fed with high-fat diet which consisted of regular feedstuff, lard, sucrose, milk powder and fresh egg, with a final composition of 16% protein, 38% carbohydrate and 46% fat and a total energy of 20.54 kJ/g throughout the study (24 weeks) (23); and the weight cycling group (WC, n=10) with a diet-switch protocol (24), where mice were fed with the high-fat diet in the first and last 8 weeks and regular chow for the remainder of the study. All experimental procedures were approved by the Institutional Animal Care and Use Committee of the Center for Animal Experiments, Wuhan University (Wuhan, China).

Mice had free access to food and water. Body mass (BM) was determined weekly throughout the study. Mice were provided with fresh chow every morning; chow remaining from the previous day was discarded after the mass was recorded. Differences between food supplied and remaining were considered as food intake (FI). Feed efficiency (FE) was calculated as ratio between BM gain (g) and food consumed (kJ) per animal over the course of the study (every 8 weeks), and is presented as a percentage (24).

After 24 weeks, baseline blood glucose of mice was measured from blood of the tail vein followed by 8 h fasting. An intraperitoneal injection of 10% glucose solution at 0.01 ml/g body weight was administered to the mice. Time points of blood glucose measurements were 0, 30, 60, 120 and 180 min following injection. Levels of glucose were detected using a Glucose assay kit (cat. no. F006; Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's protocol. The trapezoidal method was used to calculate the area under the curve of glucose (AUCG) (23).

At week 24, mice were anesthetized with pentobarbitone (50 mg/kg) via intraperitoneal injection following overnight fasting. Both sides of femoral veins were exposed and catheters for glucose and insulin infusions were inserted. A further catheter was inserted into the femoral artery for blood sampling. After the surgery was performed, mice were placed quietly without any further operation for 30 min to eliminate any possible stress. Then the hyperinsulinemic-euglycemic clamping test was performed, which lasted 120 min. The insulin infusion rate was 1.67 mU/kg/min and the arterial blood glucose level was clamped at the basal fasting concentration by supplying glucose at variable rates. Under hyperinsulinemic conditions, the steady glucose infusion rate (GIR) required to maintain euglycemia is a standard measure of systemic insulin sensitivity (25).

Total cholesterol (TC), FFAs and TGs were detected using a total cholesterol assay kit, a nonesterified free fatty acid assay kit and a triglyceride assay kit all obtained from Nanjing Jiancheng Bioengineering Institute (cat. nos. F002-2, A042-2 and F001, respectively). Interleukin (IL)-6 (cat. no. ab100713) and TNF-α (cat. no. ab1793) were assessed using commercially available ELISA kits obtained from Abcam, (Cambridge, UK) according to the manufacturer's protocol. The concentrations of TNF-α and IL-6 in each sample were determined by interpolation from the standard curve. The intra-essay precision and inter-assay precision were <6 and <10%, respectively.

At week 24, all animals were euthanized by CO₂ inhalation. A comprehensive assessment of death was made through observation of signs of breath, heartbeat and nerve reflex. One side of the epididymal fat was extracted from sacrificed mice. Tissue was weighed, divided into four equal pieces and fixed overnight at 4°C by immersion in phosphate-buffered 4% paraformaldehyde. Samples were dehydrated and embedded in paraffin. Embedded tissue was cut into 5-µm-thick sections for hematoxylin and eosin staining. Sections were routinely de-waxed with xylene for 30 min, followed by gradient ethanol dehydration with 100, 95, 85 and 70% ethanol for 2 min. The tissue sections were then immersed in hematoxylin at room temperature for 15 min, in 1% hydrochloric alcohol for 6–8 sec and in saturated lithium carbonate solution for 1 min. Then, the tissue sections were immersed in eosin staining solution for 4 min (26). Six light microscopic images (magnification, ×400) per mouse and 5 mice per group were selected at random for the quantitative analysis with Image J (version 1.52a; National Institutes of Health, Bethesda, MD, USA) and the average adipocyte volume (µm²) was extrapolated (27).

Fresh epididymal adipose tissue was homogenized with RLT lysis buffer (Rneasy Mini kit; Qiagen, Inc., Valencia, CA, USA) followed by chloroform delipidation steps. Total RNA was extracted from the aqueous phase using a silicon spin column. cDNA was reverse-transcribed with SuperScript III (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) using 1 µg total RNA, according to the manufacturer's instructions. PCR primers (Shanghai Saibaisheng Gene Technology Co. Ltd., Shanghai, China) used in the present study were as follows: CTRP3, forward 5′-GCCCCCGTATCAGGTGTGTATTT-3′ and reverse 5′-TGAAGACTGTGTTGCCGTTGTGC-3′; IL-6, forward 5′-AGTTGCCTTCTTGGGACTGA-3′ and reverse 5′-CAGAATTGCCATTGCACAAC-3′; TNF-α, forwards 5′-ACGGCATGGATCTCAAAGAC-3′ and reverse 5′-CGGCAGAGACCACCTTGAACT-3′; glucose transporter (GLUT)4, forward 5′-CCCCGCTGGAATGAGGTTTTTGAGGTGAT-3′ and reverse 5′-CAGACAGGGGCCGAAGATTGGGAGACAGT-3′; and β-actin, forward 5′-ACACCCGCCACCAGTTCGC-3′ and reverse 5′-TCTCCCCCTCATCACCCACAT-3′. β-actin mRNA was quantified as internal control. Analysis was performed using the 2⁻ΔΔ^Cq method (28). Cycle conditions were 95°C for 2 min followed by 40 cycles at 95°C for 15 sec and 65°C for 34 sec.

Proteins of epididymal adipose tissue homogenates were prepared by adding 30 µl/ml protease inhibitors (Roche Applied Science, Penzberg, Germany) and 10 µl/ml phosphatase inhibitors (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) in tissue extraction reagent I (cat. no. FNN0071; Thermo Fisher Scientific, Inc.). The protein concentration was determined using a BCA protein assay kit (Pierce; Thermo Fisher Scientific, Inc.). Samples were heated to 100°C for 5 min in SDS buffer (cat. no. orb154330; Biorbyt Ltd., Cambridge, UK) and equal volumes of protein extracts (25 µg) were separated on 12% SDS-PAGE gels and transferred to polyvinylidene fluoride membranes. Membranes were blocked with 5% skimmed milk in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) for 1 h at 20°C, followed by incubation overnight at 4°C with the following affinity-purified goat polyclonal primary antibodies obtained from Abcam: CTRP3 (cat. no. ab135301; 1:1,000), phosphatidylinositide 3-kinases (PI3K; cat. no. ab86714; 1:1,000), phosphorylated protein kinase B (p-PKB)-Ser473 (cat. no. ab18206; 1:1,000), PKB (cat. no. ab8805; 1:1,000) and β-actin: (cat. no. ab8226; 1:1,000). Horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (1:5,000; cat. no. CL7802AP; Cedarlane Laboratories, Burlington, Canada) conjugated to horseradish peroxidase was incubated with membranes for 1 h at room temperature. Following five washes with TBS-T, membranes were processed for autoradiography by enhanced chemiluminescence (Pierce; Thermo Fisher Scientific, Inc.). Densitometric analysis was performed using ImageQuant (version LAS-4000; GE Healthcare Life Sciences, Little Chalfont, UK). Experiments were performed in triplicates.

Data are presented as mean ± standard error and were evaluated statistically by one-way analysis of variance (ANOVA) with SPSS (19.0; IBM Corp., Armonk, NY, USA). Comparisons among three groups were performed by one-way ANOVA followed by Bonferroni's post-hoc tests. P<0.05 was considered to indicate a statistically significant difference.

Changes in BM of mice throughout the experiment are presented in Fig. 1. At the beginning of the study, no significant differences in BM among the groups were observed. During the first 8 weeks, the BM of mice in OB and WC groups, who were provided with high-fat diet, increased faster compared with the NC group, which was fed regular chow. Following the switch to regular chow in the WC group, BM decreased over the course of 4 weeks to match that of the NC group. Following the reintroduction of a high-fat diet at week 16, mice of the WC group regained weight and matched with the OB group in the final 3 weeks of the study. At the end of the study, BM of the OB and WC groups were significantly increased compared with the NC group (P<0.01). No significant difference in BM between OB and WC group was observed.

Mice fed regular chow and high-fat diet had a similar FI (g/animal/day) throughout the study. The FE in the WC and OB groups was higher compared with the NC group (P<0.01). In the first 8 weeks, there were no significant differences in FE between the OB and WC group (1.01±0.04 g/kJ vs. 0.99±0.04 g/kJ); however, both were significantly higher than the NC group (0.55±0.03 g/kJ, P<0.01). In the second phase of the study, the FE of the WC group was significantly lower than the OB group (0.84±0.03 g/kJ vs. 0.93±0.03 g/kJ; P<0.05), but all remained significantly higher than the NC group (0.49±0.03 g/kJ; P<0.01). In the last 8 weeks, FE in the WC group was significantly increased compared with the OB group (1.05±0.05 g/kJ vs. 0.87±0.03 g/kJ; P<0.01), indicating that the FE of WC mice was increased in the phase of body weight regaining (Table I).

At the end of the experiment, intraperitoneal glucose tolerance tests were performed to indicate levels of systemic glucose metabolism. High-fat dieting in the OB and WC group significantly increased blood glucose concentrations compared with regular chow in the NC group (P<0.01). Weight cycling resulted in an additional significant increase in blood glucose levels compared with the OB group (P<0.01; Fig. 2). The AUCG from 0–180 min (AUCG_0–180) determined for the OB and WC groups were significantly increased compared with the NC group (P<0.01). AUCG_0–180 of the WC group was significantly increased compared with the OB group (P<0.01; Table I), indicating that weight cycling deteriorated metabolic dysfunction compared with continuous high-fat dieting.

GIRs were used to evaluate insulin sensitivity. The GIR in the OB group was 6.72±0.83 mg/kg/min, which was significantly lower compared with the NC group at 10.3±0.98 mg/kg/min (P<0.01). These results indicated that a high-fat diet may induce systemic IR. Weight cycling resulted in an additional significant decrease in GIR compared with the OB group (5.30±0.72 mg/kg/min vs. 6.72±0.83 mg/kg/min; P<0.01; Table I).

Levels of TG, FFA and TC in the OB and WC groups were significantly increased compared with the NC group (P<0.01). No significant differences in TG and FFA were observed between the OB and WC groups, whereas TC levels in the WC group were significantly increased compared with the OB group (P<0.05). Compared with the NC group, IL-6 and TNF-α blood levels in the OB and WC group were significantly increased (P<0.01). IL-6 and TNF-α in blood levels in the WC group were significantly increased compared with the OB group (P<0.01; Table I).

Relative CTRP3 mRNA expression in adipose tissue of the OB and WC groups was significantly decreased compared with the NC group (P<0.01). CTRP3 mRNA expression in the WC group was significantly decreased by 20.3% compared with the OB group (P<0.01; Fig. 3). Compared with the NC group, relative CTRP3 protein expression in adipose tissue of the OB and WC groups was significantly decreased by 31.0 and 47.0%, respectively (P<0.01; Fig. 4). Significant differences in CTRP3 protein expression in adipose tissue were further observed between the OB and the WC group (P<0.01).

Adipocyte sizes in the OB and WC groups were significantly increased compared with the NC group (P<0.01). Compared with the OB group, the adipocyte size of the WC group was significantly increased by 48.42% (P<0.01; Fig. 5). Relative IL-6 and TNF-α mRNA expression in the OB and WC groups were significantly elevated compared with the NC group (P<0.01). Compared with the OB group, IL-6 and TNF-α mRNA expression in the WC group was significantly increased by 20.8 and 19.7%, respectively (P<0.01; Fig. 3).

Compared with the NC group, relative PI3K expression and PKB (Ser473) phosphorylation in adipose tissue of the OB and WC groups were significantly decreased (P<0.01; Fig. 5). PI3K expression and PKB (Ser473) phosphorylation in adipose tissue of the WC group were significantly decreased by 9.7 and 24.4%, respectively, compared with the OB group (P<0.01). Relative GLUT4 mRNA expression in the OB and WC group was decreased by 24.0 and 31.0%, respectively, compared with the NC group (P<0.01; Fig. 3). Differences in relative GLUT4 mRNA expression between the WC and OB groups were significant (P<0.01).