All animal treatment complied with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). All animal procedures were approved in accordance with the institutional guidelines established by the Committee of Ethics on Animal Experiments at the Chinese Academy of Sciences. Wild type (WT) mice were provided by the Institute of Laboratory Animal Science of Peking Union Medical College. AMPKα2-knockout (AMPKα2^−/−) mice were kindly provided by Dr. Benoit Viollet (Department of Endocrinology, Metabolism and Cancer, Institute Cochin, University Paris Descartes, Paris, France) and bred in a specific pathogen-free environment. Male AMPKα2^−/− and WT mice were both with C57BL/6J genetic background. All mice were 2 months old with a mean body weight of 18 ± 2 g at the start of the experiment.

After allowing acclimatization to their housing and the treadmill, WT mice (n = 20), and AMPKα2^−/− mice (n = 20) were randomly divided into two groups: the control group and the training group, with 10 mice in each group. Mice in the training group ran on the treadmill for 90 min/day at 9.0 meters/min (0% grade), 5 days/week for 6 weeks (Fernando et al., 1993). Body weight, heart rate and systolic/diastolic blood pressure were assessed in all animals. After 12 h of the last training, mice were anesthetized of pentobarbital (5 mg/100 g) with an intraperitoneal injection and sacrificed.

The thoracic aortas were dissected out and immersed in liquid nitrogen immediately. Then the frozen tissues were lysed in RIPA (Radio Immunoprecipitation Assay) buffer containing 150 mM NaCl, 50 mM Tris (pH 7.4), 1% sodium deoxycholate, 1% Triton X-10, 0.1% SDS, protease inhibitor (sodium fluoride, sodium orthovanadate, leupeptin, EDTA) (Beyotime, Haimen, China). After sonication on ice for 30 min and centrifugation at 12 000 rpm for 20 min at 4°C, the supernatant was collected for Western blotting as previously described (Li et al., 2012). The primary antibodies were as follows: anti-MnSOD (ABclonal, MA, USA), anti-AMPKα2 (Abcam, Cambridge, England), anti-phospho-AMPKα1/α2 (Thr¹⁷²), anti-BNIP3L (BH3 domain-containing BCL2 family members BNIP3-like) (Bioworld, St. Louis, Park, USA), anti-eNOS, anti-phospho-eNOS (Ser¹¹⁷⁷) (BD Biosciences, NJ, UK), anti-Complex I, anti-PGC-1α (peroxisome proliferator-activated receptor gamma coactivator 1 alpha), anti-Drp1(dynamin related protein 1), anti-Mfn1 (mitofusin 1), anti-LC3, anti-catalase, anti-GAPDH (Santa Cruz, CA, USA), anti-mTOR (mammalian target of rapamycin), anti-phospho-mTOR (Ser²⁴⁴⁸) (Sigma-Aldrich, St. Louis, MO, USA). Immunoreactive bands were highlighted by electrochemiluminescence (ECL) technology and quantified by densitometry using imaging software (Image Jversion 1.46, NIH, Maryland, USA). The individual values were originally expressed as a percentage of a target protein and an internal protein standard (GAPDH) (target protein content/GAPDH content) and then expressed as a fold change of the normal WT control group (target protein content/GAPDH content) value.

The paraffin sections were deparaffinized by dimethylbenzene and rehydrated by graded alcohol. Antigen retrieval was processed by citric acid buffer (pH 6.0) for 5 min at 100°C. Then the slides were incubated in hydrogen peroxide for 10 min and were blocked in TBST (tris-buffered saline and tween) containing 5% Bovine Serum Albumin at room temperature for 30 min. Some sections were subsequently incubated with 300 nM MitoTracker Green (Invitrogen, CA, USA) at room temperature for 30 min. Other sections were incubated at 4°C overnight with antibodies against AMPKα2 (1:100, Abcam, Cambridge, England), fluorescent anti-rabbit secondary antibody at a 1:400 dilution for 30 min, and then nucleus dye 4,6-diamidino-2-phenylindole (DAPI) for 3 min. All images were taken by using a Zeiss Pascal LSM 710 confocal microscope (Germany). Fluorescence intensity was analyzed with Image Pro Plus in three independent samples.

Genomic DNA of the thoracic aorta tissue was extracted by using UniversalGen DNA Kit(Cwbiotech, Beijing, China). The mitochondrial (mt) copy number was analyzed by real-time PCR (ABI 7900 Real Time PCR System; Foster City, CA) as previously described (Ray Hamidie et al., 2015), through the relative value of mitochondrial and nuclear DNA (mt:nuclear DNA) which reflects the amounts of mitochondria per cell. The mitochondrial DNA (mtDNA) forward primer was CCTAGGGATAACAGCGCAAT (5′-3′) and the reverse primer was ATCGTTGAACAAACGAACCA. The nuclear DNA (nDNA) forward primer was AGAGCTCTGCGGGTACATCT and the reverse primer was CATCAGTGACGGTGCCTTAC. Q-PCR were performed in a real time PCR system: the PCR began with 95°C denaturation for 30 s followed by 40 cycles of 95°C denaturation for 5 s, and annealing and elongation for 34 s at 60°C. Samples were assayed in triplicate. Cycle threshold (CT) was used for data analysis, and CT (nDNA)—CT (mtDNA) or ΔCT was used to reflect the difference in CT values. Results were expressed as the copy number of mtDNA per cell, 2 × 2^−ΔCT.

Mice thoracic aortas were separated, cleared of fat and connective tissues, cut into 2–3 mm rings, and fixed on isometric force transducers (Danish Myo Technology Model 610 M, Denmark) in a 5 ml organ bath, and aerated with 95% O₂ and 5% CO₂ under an initial resting tension of 2.5 mN (Zhou et al., 2014). Force was recorded in a PowerLab/8sp data acquisition system (A.D. Instruments, Castle Hill, Australia). After 1 h of incubation in oxygenated Krebs' medium (containing: KCl 4.7 mmol/L, NaCl 118 mmol/L, CaCl₂ 2.5 mmol/L, KH₂PO₄ 1.2 mmol/L, MgSO₄ 1.2 mmol/L, glucose 11 mmol/L and NaHCO₃ 25 mmol/L) at pH 7.4 and 37°C, rings contractility was tested 3 times by high K⁺ mediums (60 mM KCI) to stabilize the contraction. Cumulative response curve of phenylephrine (10⁻⁸ to 10⁻⁴ mol/L) was performed to assess the vasoconstriction response and cumulative concentration-response curves of acetylcholine (10⁻⁸ to 10⁻⁴ mol/L) were constructed with a phenylephrine pre-contraction (3 μmol/L).

All values are reported as means ± SD. Comparison of groups involved Student's unpaired two tailed t-test or two-way ANOVA with the Bonferroni test for post-hoc analysis (SigmaPlot Software, San Jose, CA, USA). P < 0.05 was considered statistically significant.

Firstly, there was no significant difference of body weight and systolic blood pressure in WT and AMPKα2^−/− mice pre and post exercise as shown in Table S1. Table S1 also showed that heart rate was decreased in the exercise group by 17.7%, which was comparable in WT and AMPKα2^−/− mice.

Next, we evaluated whether chronic exercise would have any effect on AMPKα2 expression. As shown in Figure 1A, AMPKα2 expression was dramatically increased in the aorta after chronic exercise training in immunofluorescence staining, although the overall vascular architecture had no significant difference in the four groups of mice in H&E staining (Figure S1). This was consistent with western blot results showing that exercise training induced a significant increase in aortic AMPKα2 protein expression by 31% in WT mice (Figure 1B). Furthermore, it was found that phosphor-AMPKα (p-AMPKα) (T172) was also significantly increased in the aorta of exercised mice by using phosphor-specific antibody against both α1 and α2 isoforms of AMPK. In AMPKα2^−/− mice, the protein of AMPKα2 was not detectable due to gene knockout, and p-AMPKα (T172) had similar basal levels to WT mice but showed no increase in response to exercise. These results indicated that exercise induced a significant increase in the expression and phosphorylaton of AMPKα2 in the aorta.

We then analyzed whether exercise would have any beneficial effect on vasodilation. As shown in Figure 2A, the vascular relaxation to acetylcholine was decreased in aorta rings from AMPKα2^−/− mice compared with age-matched wild type mice, indicating that AMPKα2 was involved in NO-dependent vasodilation. Importantly, the improvement of vasodilation was significantly lower in AMPKα2^−/− mice compared with WT mice, although exercise increased the vasorelaxation ability of the aorta in both WT and AMPKα2^−/− mice (Figure 2A). These results indicated that AMPKα2 played an important role in exercise-related vasorelaxation. In contrast, the vasoconstriction of aortas responding to phenylephrine was similar among the four groups (Figure 2B), suggesting that exercise might have no effect on vasocontraction.

We then further investigated whether the difference in vasodilation was due to changes in eNOS/p-eNOS and the possible involvement of AMPKα2. As expected, WT mice with exercise exhibited increased eNOS protein expression and phosphorylation in aorta compared with WT mice without exercise (Figure 2C). In contrast, AMPKα2^−/− mice with exercise did not show any increase in total eNOS level or p-eNOS level in aorta compared with AMPKα2^−/− mice without exercise. These results indicated that the improved vasodilation of aortas during exercise training in mice might be through an AMPKα2-dependent mechanism.

Accumulating studies indicated that mitochondrial content plays a critical role in maintaining vascular function. We thus evaluated whether exercise would have any effect on aortic mitochondrial content and the possible involvement of AMPKα2 by assessing mitochondrial fluorescence intensity, mtDNA copy number and Complex I protein expression.

As shown in Figure 3A, MitoTracker Green fluorescence intensity increased significantly in WT mice with exercise compared with WT mice without exercise. Consistently, the mtDNA copy number was upregulated by 34% in WT mice during exercise (Figure 3B). Meanwhile, Complex I protein expression also showed an increase of 2-folds in WT exercise mice compared to control mice (Figure 3C). In contrast, AMPKα2^−/− mice with exercise did not show any increase in mitochondrial fluorescence intensity, mtDNA copy number and Complex I protein compared with AMPKα2^−/− mice without exercise (Figure 3). These data suggested that exercise increased aortic mitochondrial content, and this effect was dependent on the presence of AMPKα2.

It has been reported that mitochondrial quantity and quality were controlled by biogenesis, fusion-fission, and autophagy. We therefore examined the effect of exercise on the expression of these relative proteins and the possible role of AMPKα2^−/−. We found that the expression of LC3B, an indicator of autophagy, and BNIP3L, a mitochondria-associated protein, were decreased in WT mice with exercise compared with WT mice without exercise (Figure 4B). Then we detected the protein expression of mTOR, the major autophagy negative regulator, and its phosphorylation at Ser2448. It was shown that WT exercise mice showed increased mTOR protein content compared to WT mice without exercise, but no significant alteration in phosphorylation activity (Figure 4B). In contrast, there was no significant difference in either LC3B/BNIP3L or mTOR protein levels in AMPKα2^−/− mice with exercise compared with these knockout mice without exercise. These results indicated that decreased autophagy may be responsible for exercise-related increase of mitochondrial content, and this effect was dependent on AMPKα2.

In contrast, there was no significant difference of PGC-1α (peroxisome proliferator-activated receptor gamma coactivator 1 alpha) protein expression, the main regulator of mitochondrial biogenesis, in the WT and AMPKα2^−/− mice with exercise compared with matched strain without exercise (Figure S2). Similarly, the protein levels of Drp1 (dynamin related protein 1) and Mfn1 (mitofusin 1), fission and fusion markers, also remained unchanged in the WT and AMPKα2^−/− mice with exercise compared with matched stains without exercise. (Figure S2). These results indicated that mitochondrial biogenesis, fission and fusion might not be involved in exercise-related mitochondria content increase in the aorta.

Finally, we evaluated the effect of exercise on mitochondrial antioxidant capacity and the possible involvement of AMPKα2. MnSOD and catalase are both critical to mitochondrial specific antioxidant defense. Figure 5A shows that the expression of catalase protein was significantly increased in the WT mice by 58% following exercise exposure, but not in AMPKα2^−/− mice. MnSOD protein content was significantly increased in WT mice after exercise intervention. Conversely, a marked decrease of MnSOD was observed in AMPKα2^−/− mice with exercise compared with those without exercise (Figure 5A). These results indicated that exercise might increase mitochondrial antioxidant response in the aorta, and this effect was dependent on AMPKα2.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.