Exercise Improves Metabolism and Alleviates Atherosclerosis via Muscle-Derived Extracellular Vesicles

Animal experiments

Male C57BL/6J 8-week-old mice were obtained from the Experimental Animal Center of Fourth Military Medical University. Male ApoE^-/- 8-week-old mice on a C57BL/6 background were obtained from Model Animal Research Center of Nanjing University. All animal experiments were approved by the Animal Care and Use Committee of Fourth Military Medical University, in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23). All mice were housed in specific pathogen-free conditions at 22°C-24°C and 40-60% humidity, with a 12 light/12 dark cycle. After seven days of adaptation, C57BL/6J mice were fed a normal chow diet and were subjected to a 4-week exercise protocol to obtain exercise related muscle extracellular vesicles. Obese C57BL/6 mice were induced by feeding with a high-fat diet for 12 weeks. In order to induce atherosclerosis, ApoE^-/- mice were fed with a high-fat diet for 12 weeks (D12492, study diet containing 45% fat, 20% protein, and 20% carbohydrate). At the end of the experiment, mice were anesthetized with 120 mg/kg body weight of ketamine and 24 mg/kg body weight of xylazine, and then sacrificed by CO₂ exposure.

Exercise protocol and additional treatment

The swim training protocol was modified from previous studies [19]. Briefly, an endurance swimming program was undertaken five times a week in the exercise group in a temperature-controlled water bath (35-36 °C). All training sessions were conducted from 5:00 p.m. to 6:00 p.m. Before the formal exercise program begins, mice have a 5-minute swimming session to acclimate. The endurance swimming program began with 10 minutes of swimming once a day and increased by 5 minutes every day to 30-minute sessions each day.

For extracellular vesicles biogenesis inhibition, GW4869 (S7609, Selleck, USA) dissolved in DMSO (D8418, Sigma, USA) was injected intraperitoneally at a single dose of 2.5 μg/g three times per week [20].

Blood tests

For blood glucose tolerance test, mice were intraperitoneally injected with glucose at 2 g/kg, and then blood glucose from tail vein was measured at 0, 15, 30, 60, and 120 minutes after injection utilizing an ACCU-CHEK glucometer (Roche, Germany). For other blood tests, the whole blood was collected from the eyeballs of anesthetized mice. The blood was centrifuged for 3,000 g at 4 °C after being placed at room temperature for 20 minutes, and the supernatant was collected for further analysis. Total cholesterol (TC), total triglyceride (TG), high-density lipoprotein (HDL) cholesterol, and low-density lipoprotein (LDL) cholesterol, liver function test including the measurement of ALT (Alanine Amino-transferase) and AST (Aspartate Aminotransferase) were measured using Chemray 240 and Chemray 800 Chemistry analyzers (Rayto, Shenzhen, China).

Noninvasive pulse wave velocity (PWV) measurements

Mouse PWVs were measured by experienced technicians using Vevo 2100 Imaging System (FUJIFILM, VisualSonics, Canada) armed with MS550 transducer (22-55 MHz). A chemical hair remover was used to remove mouse hairs in the chest and abdomen. After anesthesia with isoflurane (2% induction, 1.2% maintenance), the mice were placed on a heating pad with temperature control to maintain their body temperature. During the examination, the heart rate was maintained between 400-450 beats per minute. Peak velocities of the ascending and abdominal aortic arteries were measured using PWV Doppler mode. PWV was calculated by the following formula: D/[T2-T1] (m/s), in which T1 is the starting time of R wave in the ascending aortic Doppler waveform, and T2 is the starting time of R wave in the abdominal aortic Doppler waveform. D is the distance between the ascending aorta and the abdominal aorta [21]. A spectrum analyzer and real-time signal acquisition system were used to measure the parameters.

Extracellular vesicles isolation, characterization, and tail vein injection

For isolation of muscle-derived EV, muscles were collected from sacrificed mice under sterile conditions, followed by cut into small pieces less than 1mm³ and cultured in serum-free DMEM for 24 h. The culture medium was centrifuged at 300 g for 10 minutes, and 10,000 g for 30 minutes to remove the cell debris and/or dead cells. The sample was then ultracentrifuged at 100,000 g for 70 minutes, and the pellet was resuspended in sterile PBS and ultracentrifuged for another 70 minutes at 100,000 g. The pellets were resuspended in PBS at a suitable concentration and stored at -80°C till use.

To isolate plasma EVs, mice were anesthetized and then the whole blood was collected as described above and stored in anticoagulated tubes containing EDTA. EV isolation was performed using commercial kit ExoQuick™ (ExoQ5TM-1/TMEXO-1, SBI, USA) as instructed.

EV size distribution and concentration were determined by NanoSight NS300 (Malvern, Egham, UK). Transmission electron microscopy (JEOL, Tokyo, Japan) was used to observe the morphology of extracellular vesicles. Exosome inclusive markers CD9 and TSG101, and exclusive marker GM130 were used to confirm the identity of isolated EVs. The specific procedures were the same as previously described [22]. In the study, primary antibodies included anti-CD9 (ab29726, Abcam), anti-TSG101 (ab125011, Abcam), anti-GM130 (ab30637, Abcam), and anti-GAPDH (60004-1-Ig, Proteintech). Horseradish peroxidase-conjugated goat anti-rabbit (7074, CST) and goat anti-mouse (7076, CST) secondary antibodies were used in the study.

For in vivo and ex vivo tracking of EV, mice were injected with 100 μg DiR/DiI labeled EVs via tail vein and sacrificed 4 hours later for IVIS or fluorescence microscopic analysis. For EV treatment, mice were injected with 100 μg indicated EVs via tail vein twice a week for 12 weeks, followed by systemic analysis.

Histology

ApoE^-/- mice were intracardially perfused with ice-cold phosphate-buffered saline (PBS) and 4% para-formaldehyde sequentially after deep anesthesia. Then, the aortas were exposed, and the surrounding adipose tissues were removed, with the aortic arch macroscopic view taken by a digital camera. The en-face aorta was stained with Oil Red O, and the percentage lesion area was calculated using Image J. Hematoxylin and eosin (H&E) staining, Oil Red O staining, and Masson's trichome staining were performed in the aortic root cross section to identify atherosclerotic areas, lipid deposition, and collagen content. In addition, H&E and Oil Red O staining were also performed on liver sections. eWAT sections were also included for H&E staining, adipocyte areas and diameters were calculated.

Extracellular vesicle labeling and in vivo tracing

For tracking the distribution of EVs in vivo, purified EVs were incubated with 8 μM DiR or DiI (Invitrogen, Waltham, USA) at 37 °C for 30 min, and free DiR or DiI was removed by another round of EV isolation. About 100 μg DiR-labeled EVs were injected into the mice via tail vein and IVIS Lumina II system (PerkinElmer, Thermo Fisher, USA) was used to measure EV distribution among the main organs 4 hours after injection. To further profile cellular distribution of EV in the liver, heart, spleen, and other tissues, mice were injected with DiI-labeled EVs and the organs were removed four hours after injection, followed by embedded in optimum cutting temperature (OCT) compound (Sakura, Finetek, USA) and sectioned. DAPI (Invitrogen, Waltham, USA) was used to counterstain nuclei. Fluorescence signal was detected and imaged by confocal laser scanning microscope (Nikon A1R, Tokyo, Japan).

Mass Spectrometry

EVs isolated from muscles of control or exercised mice were resuspended in PBS. RIPA lysis buffer was used to extract the proteins from the EVs and total proteins from EVs were separated by electrophorese. Then, the gel was subjected to mass spectrometry. Mass spectrometry and bioinformatics analysis were performed by Bangfei Bioscience (China). An Orbitrap Fusion Lumos (Thermo Scientific, USA) was used to analyze trypsinized samples separated by high performance liquid chromatography (HPLC). GO enrichment and KEGG pathway enrichment were used to identify differentially expressed proteins.

Statistical analysis

Data are presented as mean ± SEM. The Shapiro-Wilk test was used to determine the normality of the data distribution. If distributions were normal, student t-test was used for the comparison of two groups. If distributions were nonnormal or N was too small (N<6), a Mann-Whitney U-test was used for the comparison of two groups, and Kruskal-Wallis test with Dunn’s post hoc test for multiple-group comparisons. P<0.05 was considered statistically significant for comparisons between groups using post hoc analysis (GraphPad Prism 8.0).

Swim Training improves the metabolism in obese mice via EVs

It was shown that exercise has induced robust changes of EVs [16]. To explore whether EVs are involved in exercise-mediated metabolic remodeling, we first used GW4869 (a widely used blocker of exosome biogenesis and secretion [23-25]) to block the biogenesis and secretion of EV. As expected, injection with GW4869 in WT mice resulted in a significant decrease in plasma EV concentration (Supplementary Fig. 1A-C). In the following experiments, HFD fed WT mice were subjected to a 12-week swimming regimen and additionally injected with control or GW4869 intraperitoneally three times a week 4 h before exercise (Fig. 1A). HFD induced body weight increase was significantly inhibited in swimming-trained mice (exe), while this benefit was not affected by GW4869 (Fig. 1B). In contrast, exe-mice increased glucose tolerance as evidenced by the lower area under the curve (AUC) of glucose tolerance test (GTT). Additional treatment of GW4869 suppressed the effects (Fig. 1C, D). Exercise also reduced TG, TC, and LDL cholesterol, and GW4869 treatment blunted the protective effects of exercise on TG and TC. GW4869 treatment had no significant effects on LDL and HDL (Fig. 1E-H). Interestingly, GW4869 treatment on sedentary control mice had no significant effects on the metabolic parameters examined (Supplementary Fig. 2A-H). Overall, the above results demonstrate that exercise ameliorates metabolic disorders in obese mice and GW4869 at least partially attenuates exercise-mediated metabolic benefits by inhibiting EV biogenesis.

Swim Training alleviates atherosclerosis in ApoE^-/- mice via EVs

Given the causal role of systemic metabolic disorders in atherosclerosis, we next explored whether swim training alleviated atherosclerosis in an EV dependent manner in ApoE^-/- mice (Fig. 2A). GW4869 treatment also decreased plasma EV concentration in ApoE^-/- mice (Supplementary Fig. 3A-C). Similarly in obese WT mice, swimming training improved the metabolic profile of ApoE^-/- mice (Fig. 2B-H), and the protective effects on glucose tolerance, TG, and TC were at least partially blocked by GW4869 (Fig. 2C-F). In addition, exercise also reduced lipid accumulation in hepatocytes of ApoE^-/- mice, while additional treatment of GW4869 partially suppressed the effects (Fig. 2I, J). Similarly, exercise reduced the size of visceral adipocytes, which could be slightly inhibited by GW4869 (Supplementary Fig. 4A-C). To evaluate hepatic function, we also tested serum glutamic aminotransferase (ALT) and glutamic oxalacetic aminotransferase (AST), and the results showed that exercise significantly decreased AST and ALT in ApoE^-/- mice, while additional treatment of GW4869 attenuated the protective effects (Supplementary Fig. 4D, E). Notably, GW4869 treatment on sedentary ApoE^-/- mice had no significant effects on the metabolic parameters examined (Supplementary Fig. 5A-O). In conclusion, EVs at least partially contribute to the exercise-mediated metabolic remodeling.

Consistent with the beneficial effects on metabolism, exercise also delayed progression of atherosclerosis in EV dependent manner. We first assessed arterial stiffness by PWV, a classical indicator of vascular stiffness [26]. Compared to the control group, the aortic PWV of exercised ApoE^-/- mice was much lower. Additional treatment of GW4869 partially blocked the PWV decrease (Fig. 3A, B). Accordingly, exercise reduced the plaque size and number in aorta of ApoE^-/- mice. Notably, Masson staining showed that the collagen content in the aortic roots did not change in exercise group (Fig. 3G, I). GW4869 treatment reduced the protective effects of exercise (Fig. 3C-H), while GW4869 alone had minimal effects on atherogenesis (Supplementary Fig. 6). Together, these data revealed that exercise alleviates atherosclerosis in ApoE^-/- mice at least partially via EVs.

Exe-EVs remodels systemic metabolism similar as exercise in obese WT mice

Given the data described above, we hypothesized that muscle-derived EVs might be involved in the process. EVs from muscle tissue of control (ctrl-EV) and exercised mice (exe-EV) were thus isolated (Supplementary Fig. 7A). Nanoparticle tracking analysis (NTA) showed no significant difference in particle size distribution, with the concentration slightly increased in exercised group (Supplementary Fig. 7B, C). Transmission electron microscopy (TEM) analysis showed that both ctrl-EV and exe-EV were typically round in shape and ranged 80 ~ 100 nm in diameter (Supplementary Fig. 7D). Western blot analysis clarified the presence of exosome marker proteins (TSG101, CD9) and absence of exclusive marker protein GM130 [27], confirming that ctrl-EV and exe-EV are mainly exosomes (Supplementary Fig. 7E, with the original western blot bands showing in Supplementary Fig. 8).

To explore whether exe-EV has protective effects similar to exercise, we first tracked DiR/DiI-labeled muscle EVs (Fig. 4A). Both DiR-labeled ctrl-EV and exe-EV were found mainly transported to the liver, spleen, and lung in WT mice. Besides, ctrl-EV and exe-EV were also sparsely present in some other organs, such as the visceral WAT (Fig. 4B-D). The distribution profile of EVs was further confirmed by laser confocal microscopy analysis of DiI-labeled ctrl-EV and exe-EV (Supplementary Fig. 9A, B).

We next injected WT mice fed with high-fat diet with 100 µg ctrl-EV or exe-EV by tail vein twice a week for 12 weeks (Fig. 5A). No differences in body weight were observed among the different groups (Fig. 5B). However, exe-EV treatment improved glucose tolerance (Fig. 5C, D), and reduced TG (Fig. 5E) and TC levels (Fig. 5F), while had no significant effects on LDL and HDL levels (Fig. 5G, H). Overall, exe-EV could improve the metabolism profile in obese WT mice.

Exe-EVs alleviates atherosclerosis in ApoE^-/- mice similar as exercise

In view of the above data, we next aimed to investigate the possibility of exe-EVs to alleviate atherosclerosis. The same exe-EVs treatment protocol was conducted in ApoE^-/- mice (Fig. 6A). As expected, improved glucose tolerance and decreased TG and TC were found in the exe-EVs treated group (Fig. 6C-F). In contrast, body weight, LDL, and HDL did not change significantly in mice with exe-EVs treatment (Fig. 6B, G, H).

Exe-EV had a similar distribution profile in ApoE^-/- mice as in WT mice, with liver and spleen as the dominant organs (Supplementary Fig. 10, S11), suggesting that liver might be one of the key effector organs in EV mediated cardiometabolic protection. To further explore the therapeutic effects of EVs, 100 µg of ctrl-EV and exe-EV were injected into ApoE^-/- mice twice a week for 12 weeks via tail vein. Similar to that observed in exercised ApoE^-/- mice, exe-EVs treatment reduced the lipid deposition in the liver (Fig. 6I, J). In addition, exe-EV treatment significantly reduced the AST and ALT (Fig. 6K, L). Moreover, there was also a significant reduction of adipocyte size in visceral adipose tissue in mice with exe-EV treatment (Fig. 6M-O). In contrast, no such protective effects were observed in ApoE^-/- mice receiving ctrl-EV treatment.

We next examined whether exe-EV treatment alleviated atherosclerotic lesions. In comparison with the PBS control and ctrl-EV treated groups, exe-EV treatment significantly reduced aortic PWV (Fig. 7A, B), decreased atherosclerotic plaque size and number (Fig. 7C, D). The plaque burden, in particular the lipid core, as examined by H&E staining and oil red O staining, was significantly lowered by treatment of exe-EVs (Fig. 7E-H). However, there were no differences in collagen content in the plaques among the groups (Fig. 7F, I). The above data clearly demonstrated that exe-EV treatment attenuated atherosclerosis in ApoE^-/- mice.

Profiling of the protein components of exe-EVs

In order to identify the bioactive molecules in EVs that might be responsible for exercise-induced metabolic effects, we thus analyzed the protein contents of ctrl-EV and exe-EV by mass spectrometry. A total of 1297 proteins in ctrl-EV and 1311 proteins in exe-EV were identified (Supplementary Fig. 12A). A threshold of fold change ≥2, P<0.05 was used for screening of differential proteins. Compared with ctrl-EV, 192 proteins were upregulated, while 171 proteins were downregulated in exe-EV (Supplementary Fig. 12B-C). Gene Ontology enrichment analysis revealed that upregulated proteins in exe-EV were enriched in fatty acid metabolism and mitochondria biogenesis (Supplementary Fig. 12D). KEGG pathway analysis also revealed that upregulated proteins in exe-EV are mainly involved in carbon metabolism, tricarboxylic acid cycle propanoate metabolism, fatty acid metabolism, and pyruvate metabolism pathways (Supplementary Table 1), which may at least partially explain the metabolic benefits of exe-EV. Future study to reveal the essential component(s) for the beneficial effects might shed light on exe-EV mimetic engineering. It is also important to mention that there might be dozens of proteins together exerting the beneficial function, rather than one or two specific molecule(s).