Mechanical overload-induced muscle-derived extracellular vesicles promote adipose tissue lipolysis

Abstract

How regular physical activity is able to improve health remains poorly understood. The release of factors from skeletal muscle following exercise has been proposed as a possible mechanism mediating such systemic benefits. We describe a mechanism wherein skeletal muscle, in response to a hypertrophic stimulus induced by mechanical overload (MOV), released extracellular vesicles (EVs) containing muscle-specific miR-1 that were preferentially taken up by epidydimal white adipose tissue (eWAT). In eWAT, miR-1 promoted adrenergic signaling and lipolysis by targeting Tfap2α, a known repressor of Adrβ3 expression. Inhibiting EV release prevented the MOV-induced increase in eWAT miR-1 abundance and expression of lipolytic genes. Resistance exercise decreased skeletal muscle miR-1 expression with a concomitant increase in plasma EV miR-1 abundance, suggesting a similar mechanism may be operative in humans. Altogether, these findings demonstrate that skeletal muscle promotes metabolic adaptations in adipose tissue in response to MOV via EV-mediated delivery of miR-1.

1 |. INTRODUCTION

A sedentary lifestyle can contribute to or exacerbate numerous diseases that represent urgent public health concerns, including heart failure, coronary heart disease, type 2 diabetes, breast and colon cancer.^1,2 While there is a general consensus that regular physical activity confers health benefits by enhancing the function of non-skeletal muscle tissues, the molecular mechanisms that transduce these beneficial effects are largely unknown.³ The majority of studies investigating the systemic benefits of exercise have focused on endurance exercise, leading to a relative lack of information regarding the systemic health benefits of resistance exercise beyond improvements in muscle function. The few studies that have examined the effects of resistance exercise on non-muscle tissues suggest that endurance and resistance exercise may confer comparable metabolic benefits.^4–8

In the past two decades, skeletal muscle has emerged as a secretory organ, producing myokines that may mediate some of the beneficial effects of exercise.^9–11 More recently, extracellular vesicles (EVs), primarily exosomes, have emerged as potential mediators of intercellular communication. Pioneering work demonstrated that exosomes contain miRNAs that could be delivered to recipient cells and modulate mRNA stability.¹² We and others have shown that EVs are released from cultured rodent and human muscle cells.^13–19 In particular, we reported that EVs released from primary myoblasts deliver miR-206 to primary fibroblasts and down-regulate Rrbp1, a master regulator of collagen synthesis.¹⁷ In addition, we demonstrated that satellite cells, independent of fusion, deliver EV-containing miRNAs to myofibers thereby regulating matrix metalloproteinase 9 (Mmp9) expression and facilitating muscle growth.¹⁸ The mouse C2C12 cell line was shown to release exosomes containing muscle-specific miR-133a which, when taken up by 3T3-L1 cells, repressed target gene expression.²⁰ A recent study showed an acute bout of endurance exercise promoted the release of EVs that were preferentially taken up by the liver.²¹

The primary role of adipose tissue is to store excess energy as lipids, and it is the principal source of serum non-esterified fatty acids (NEFA) during times of energy demand (ie, fasting or exercise). As such, the metabolic function of adipose tissue is critical for determining whole body metabolic outcomes. Adipocyte lipolysis is driven by catecholamines (norepinephrine and epinephrine), which stimulate β-adrenergic receptors (Adrβ). Adrβ agonism enhances insulin sensitivity, lean mass, energy expenditure and thermogenesis.^22–26 Beta-adrenergic stimulation of lipolysis is a complex mechanism involving several proteins (eg, adipose triglyceride lipase [ATGL], comparative gene identification-58 [CGI-58], protein kinase A [PKA], hormone-sensitive lipase [HSL], monoglyceride lipase [MGL], perilipin-1 [PLIN-1]).²⁷ It has been demonstrated that Adrβ3 stimulation directly or indirectly activates these proteins leading to a consecutive hydrolysis of triglycerides and the formation of fatty acids and glycerol.²⁷ MiRNAs in adipose tissue have been shown to modulate Adrβ signaling thereby improving catecholamine sensitivity of adipose tissue.^28,29

Although relatively few studies have investigated whether adipose adrenergic signaling is affected by resistance exercise, it is known that a single bout of resistance exercise can increase circulating NEFA and resting energy expenditure and decrease respiratory quotient for up to 24 h,^30–34 indicative of increased adipocyte lipolysis and muscle fatty acid oxidation. In the present study we show that adipose transcriptional responses to mechanical overload of muscle (MOV) are EV-dependent, and that serum EVs enhance adipocyte catecholamine sensitivity and lipolysis for at least 24 h in response to MOV. To our knowledge, this is the first demonstration of a potential mechanism whereby a hypertrophic stimulus imparts metabolic adaptations in adipose tissue. Since adipose metabolic function is crucial for determining whole-body metabolic outcomes, the ability of resistance exercise-induced EVs to improve adipose metabolism has significant clinical implications.

2 |. MATERIALS and METHODS

2.1 |. Reagents or resources

3T3-L1 (ATCC, CL-173); 3-isobutyl-1-methylxanthine (Sigma, cat. no. I5879); Bioanalyzer pico-chip (Agilent Technologies, cat. no. 5067–1513); Clarity Western ECL substrate (Bio-rad, 170–5060); Collagenase Type I (Sigma-Aldrich, cat. no. 1148089); Collagenase Type II (Gibco, cat. no. 17101015); Dexamethasone (Sigma, cat. no. D4902); Dulbecco’s Modified Eagles Medium (Hyclone, cat. no. SH3028401); Detergent compatible (DC) protein assay (BIO-R AD, cat. no. 5000112); Direct-zol kit (Zimo Research, cat. no. R2072); Dispase (Gibco, cat. no. 17105041); Dulbecco’s Phosphate-Buffered Saline (ThermoFisher Scientific, cat. no. 14190–144); ExoEasy membrane affinity column (Qiagen, cat. no. 77064); Exo Glow (System Biosciences, cat. no. EXOR100A-1); ExoQuick-LP (System Biosciences, cat. no. exolp5a-1); ExoQuick-TC (System Biosciences, cat. no. exotc50a-1); ExoRNeasy kit (Qiagen, # 77044); Exosome-depleted FBS (Gibco, cat. no. A2720801); Fast SYBR Green Master Mix (ThermoFisher Scientific, cat. no. 4385612); Fetal Bovine Serum (Gibco, cat. no. 26140079); Fragment crystallizable block (Biolegend, cat. no. 101319); Fragment crystallizable block TruStain (Biolegend, cat. no. 422301); Free Glycerol Reagent (Sigma Aldrich, cat. no. F6428); GW4869 (N,N′Bis[4-(4,5-dihydro-1H-imidazol-2-yl) phenyl]-3,3′-p-phenylene-bis-acrylamide dihydrochloride; Cayman Chemical); Halt Protease Inhibitor Cocktail (Thermo Scientific, cat. no. 78430); Ham’s F-10 (ThermoFisher, cat. no. 11550043); Horse Serum (Gibco, cat. no. 26050088); Isoproterenol (Millipore Sigma, cat. no. 420355); Lipofectamine 2000 (Invitrogen, cat. no. 11668–019); Low Glucose Dulbecco’s modified Eagle’s medium (Gibco #11885–084); miRCURY RNA Isolation Kit—Biofluids (Exiqon, #300112); MyCult serum free differentiation media (Stemcell Technologies, cat. no. 05965); Nano-Glo Luciferase Assay System (Promega, cat. no. N1610); Newborn Calf Serum (Gibco, #16010–159); NIH/3T3 (ATCC, CRL-1658); Nitrocellulose membrane (BIO-RAD, cat. no. 1620115); Non-Fat dry milk (Bio-Rad, 170–6404); Opti-MEM reduced serum medium (ThermoFisher Scientific, #31985–070); Paraformaldehyde (Millipore Sigma, 158127); Penicillin/Streptomycin (Gibco, cat. no. 15140122); Recombinant Human FGF-basic (Peprotech, cat. no. 100–18B); Red Blood Cell Lysis Buffer (Biolegend, cat. no. 420301); RNAse inhibitor (ThermoFisher Scientific, cat. no. N8080119); SuperScript VILO (ThermoFisher Scientific, cat. no. 11756050); TaqMan Fast Advanced Master Mix (ThermoFisher Scientific, cat. no. 4444557); Taqman MicroRNA Reverse Transcription Kit (ThermoFisher Scientific, cat. no. 4366596); TRIzol Reagent (ThermoFisher Scientific, cat. no. 15596026); Wako HR series NEFA-HR (FujiFilm, 999–34691; 995–34791; 991–34891; 993–35191).

2.2 |. Animals

All animal procedures were conducted in accordance with institutional guidelines for the care and use of laboratory animals as approved by the Institutional Animal Care and Use Committee of the University of Kentucky. Adult (4 months of age), male C57BL/6J mice were housed in a temperature-and humidity-controlled room and maintained on a 14:10-h light-dark cycle with food and water ad libitum.

2.3 |. Synergist ablation

The synergist ablation surgery is an established rodent model of resistance exercise that induces robust hypertrophy of the plantaris muscle as a result of MOV. Four-month-old male C57BL/6J mice underwent either a sham or synergist ablation surgical procedure as previously described by us.³⁵ Briefly, following a small incision on the dorsal aspect of the lower hind limb of a fully anesthetized mouse (2% isoflurane at 0.5 L/min), a small portion of gastrocnemius muscle (proximal to the Achilles tendon) was excised without disturbing the blood supply or innervation to the plantaris muscle; the sham surgery involved all aspect of the synergist ablation surgery without excising a portion of the gastrocnemius muscle. The plantaris muscles were collected at 1, 2, and 3 days after the surgery depending on the experiment.

2.4 |. Isolation of EVs from mouse serum or culture media

EVs were isolated from fresh serum or culture media using ExoQuick-LP or ExoQuick-TC, respectively, according to the manufacturer’s directions. Briefly, serum (500 μL) or culture media (12 mL) was centrifuged at 3000 g for 20 min twice. The resulting supernatant was then filtered through a 0.22 μm syringe-driven filter to remove large contaminating vesicles. The serum was incubated with magnetic beads and pre-clearing reagents to deplete lipoproteins. Following a 3 h incubation at 4°C, the mixture was then placed on a magnetic separator with the resulting lipoprotein-depleted serum incubated overnight with ExoQuick reagent at 4°C. Following overnight incubation, the serum was centrifuged at 16 000 g for 10 min with the final pellet resuspended in 25 μL of Dulbecco’s phosphate-buffered saline (DPBS). The filtered media was incubated overnight with ExoQuick reagent at 4°C and then centrifuged at 1500 g for 30 min with the resulting pellet resuspended in 25 μL of DPBS. Due to the limitations of ExoQuick, we also isolated EVs in a separate set of mice utilizing ExoEasy (described below) to confirm EV isolation.

2.5 |. Isolation of EVs from human plasma

The initial centrifugation and filtration of human plasma followed the same procedure as that described using mouse serum. Following the filtration, EVs from the filtered plasma (or mouse serum) were isolated using the ExoEasy membrane affinity column according to the manufacturer’s directions. Briefly, filtered plasma (or serum) was mixed 1:1 with 2× binding buffer and then added to the membrane affinity column. After centrifugation, the column was washed and EVs eluted in 400 μL of elution buffer.

2.6 |. Treatment of mice with GW4869

GW4869 is an inhibitor of neutral sphingomyelinase which has been shown to block the release of exosomes.³⁶ GW4869 was first suspended in DMSO at a stock concentration of 8 mg/mL, which was then diluted to 2 mg/mL with sterile 0.9% saline immediately prior to administration. Mice were administered GW4869 by intraperitoneal injection at a final concentration of 0.3 mg/mL as describe by Dinkins and colleagues.³⁶ Vehicle-treated mice were administered an equal volume of 3.75% DMSO diluted in sterile saline. Mice were treated with either GW4860 or vehicle for eight consecutive days with mice undergoing the sham or synergist ablation surgery on day 7 and blood and muscle collected 24 h later.

2.7 |. Nanoparticle tracking analysis

EV size and concentration were measured in scatter mode by nanoparticle tracking analysis using NanoSight (NanoSight) or ZetaView (Particle Metrix) instruments. Prior to each analysis, camera sensitivity and focus were adjusted to ensure the sensitivity remained constant throughout the analysis. EVs were diluted in DPBS and measured at 23–24°C. For ZetaView, the capture and analyze settings were: sensitivity 65, shutter 100, minimum trace length 10, minimum brightness 30, maximum brightness 255, minimum area 10, and maximum area 1000. For NanoSight, ideal measurement concentrations were found by pre-testing the ideal particle per frame value (20–100 particles/frame). For each measurement, five 1-min videos were captured under the following conditions: cell temperature: 25°C; Syringe speed: 40 μL/s.

2.8 |. Transmission electron microscopy

For transmission electron microscopy (TEM), freshly isolated EVs were used. A small aliquot of EVs was allowed to adsorb on to a 300 mesh Formvar/carbon support film grid for 30 minutes with excess wicked off with a filter paper. The grid was then negatively stained with a small droplet of 1% phosphotungstic acid for 1 min with excess stain removed with filter paper. TEM visualizations were performed using a Zeiss EM10 transmission electron microscope at 60 kV.

2.9 |. EV labeling and tracking

EVs were isolated from the serum of mice that had undergone one day of either sham or MOV. EV RNA was labeled using ExoGlow according to manufacturer’s instruction. Labeled EVs were resuspended in DPBS and then intraperitoneally injected into exercise-naïve C57BL/6J mice with tissue collected 6 h later. Upon collection, each tissue was immediately lysed in 50 mM Tris-HCL, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 2 mM EDTA, 50 mM sodium fluoride supplemented with RNAse inhibitor. Tissue lysate fluorescence was measured using a plate reader with excitation/emission wavelength of 460/650 nm.

For live visualization filtered serum was incubated with 1 mM fluorescent lipophilic tracer DiR (1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyanine iodide) (D12731, Invitrogen, Life Technologies) at room temperature for 15 min prior to EV isolation by ExoQuick method as described above. After tail-vein injection, labeled EVs were visualized in both mice and tissue by IVIS (in vivo imaging system) Spectrum (PerklinElmer, Waltham, MA, USA), using excitation/emission wavelength of 710/760 nm.

2.10 |. Western blot analysis

Mouse serum EVs or 3T3-L1 adipocyte cells were lysed in 50 mM Tris-HCL, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 2 mM EDTA, 50 mM sodium fluoride supplemented with Halt Protease Inhibitor Cocktail. Protein concentration was measured using DC protein assay. Equivalent amounts of protein (20 μg) were resolved on 4%−12% polyacrylamide gels and then transferred to nitrocellulose membrane. Membranes were then blocked (5% milk in Tris-buffered saline, 0.1% Tween 20) and incubated with primary antibodies GAPDH (1:1000; Cell Signaling, 5174), TSG101 (1:500; BD Bioscience, Clone 51), HSP70 (1:1000), CD63 (1:1000), CD81 (1:1000) (System Biosciences), and horse radish peroxidase (HRP) conjugated secondary antibodies (1:10 000), Goat anti-Mouse IgG (H + L) Cross-Adsorbed secondary Antibody, HRP (ThermoFisher Scientific, RRID:AB_2536527) and Goat anti-Rabbit IgG (H + L) Cross-Adsorbed secondary Antibody, HRP (ThermoFisher Scientific, RRID:AB_2536530). Immunoblots were then exposed to Clarity enhanced chemiluminescence reagent (BioRad, Hercules, CA, USA), imaged (ChemiDoc MP, BioRad), and analysed for signal density (ImageLab 5.2.1, BioRad).

2.11 |. RNA isolation and cDNA synthesis

Total RNA was isolated from mouse plantaris muscle, human vastus lateralis, EVs, human primary muscle cells, epidydimal adipose tissue or 3T3-L1 adipocyte cells previously flash-frozen in liquid nitrogen. Samples were homogenized in a tissue homogenizer (Bullet Blender, Next Advance Inc) using TRIzol Reagent and Direct-zol kit according to the manufacturer’s instructions. RNA concentration and purity were assessed using Nanodrop 2000 (ThermoFisher Scientific). cDNA was synthesized from 500 ng of total RNA using the SuperScript VILO according to the manufacturer’s instructions. For EVs, total RNA was isolated using either miRCURY RNA Isolation Kit—Biofluids or ExoRNeasy kit following the manufacturer’s directions. RNA concentration was measured using the Agilent Bioanalyzer 2100 pico-chip. Reverse transcription (RT) reactions for miR-1, cel-miR-39, and U6 small nuclear RNA (U6) were performed with 1 ng of total RNA using Taqman MicroRNA Reverse Transcription Kit according to the manufacturer’s directions. A 5′ phosphorylated cel-miR-39 oligonucleotide (Taqman, cat. no. 478293_mir) was spiked-in to each EV RNA sample for normalization of EV miRNA abundance.

2.12 |. Gene and miRNA expression

Tfap2α (Mm00495574_m1), Adrβ3 (Mm02601819_g1) and Rpl38 (Mm03015864_g1) gene expression were analyzed using TaqMan Fast Advanced Master Mix. To detect all Cebpα splice variants, primers (forward: TGCGCAAGAGCCGAGATAAA and reverse: TCACTGGT CAACTCCAGCAC) were designed that amplified exon 2. Primers (forward: CTCCCTCCAAAGGCGTTCTT and reverse: TGGCCTCAGATTCCCCAAAC) were also designed to detect VCP (valosin containing protein) using SYBR Green Master Mix. A 10-fold dilution of cDNA into a 10 μL qPCR final reaction. TaqMan Fast Advanced Master Mix was used to measure miR-1 (#002222), cel-miR-39 (#000200), and U6 (#001973) expression according to manufacturer’s directions. qPCR was performed using the ABI 7500 qPCR system. qPCR efficiency was calculated by linear regression during the exponential phase using LinRegPCR software v11.1.³⁷ For tissues and cells, the comparison of mRNAs or miRNAs expression between groups was determined following normalization with Rpl38 (Taqman), VCP (SYBR Green) or U6, respectively. For EVs, spiked in cel-miR-39 was used for normalization of miR-1 abundance.

2.13 |. Ex vivo adipocyte lipolysis

Adipocyte lipolysis was performed as previously described.^38,39 Epidydimal white adipose tissue (eWAT) was excised from each mouse and minced into 1 mm pieces in digest medium (Dulbecco’s Modified Eagles Medium [DMEM], 1 g/L glucose, 4% fatty acid-free bovine serum albumin [BSA]). Tissue was digested in 0.3% collagenase I for 1 h at 37°C with gentle agitation. Cells were strained using a fine mesh to remove connective tissue and then centrifuged at 700 g for 5 min to pellet stromal/vascular cells. Adipocytes (floating cells) were removed and washed twice in lipolysis medium (50% phosphate buffered saline, 50% DMEM, 0.5 g/L glucose, 2% fatty acid-free BSA), and then resuspended to a final 10% (vol/vol) cell concentration. Cell aliquots were treated with adenosine deaminase for 5 min and then treated with 100 μM isoproterenol or lipolysis medium only. The cells were incubated for 1 h at 37°C with gentle agitation. Finally, cells were centrifuged at 700 g for 3 min at 4°C with the infranatant collected and used to measure free fatty acids (Wako HR Series NEFA-HR kit [FUJIFILM Wako Diagnostics]) and glycerol.

2.14 |. Indirect calorimetry

Indirect calorimetry was performed according to the best practices and standard operating procedure for use of TSE LabMaster- Indirect Calorimetry Research Platform Version-1 (TSE Systems, Chesterfield, MO). Individually housed mice were allowed to acclimate to LabMaster chambers with standard bedding and nest material for one week prior to data acquisition. Baseline (pre-surgery) calorimetry data were obtained for three days, and then mice underwent sham or synergist ablation surgery and returned to LabMaster chambers for one additional day of data acquisition. The TSE LabMaster system provides accurate measurement of food and water intake and indirect calorimetry (oxygen consumption and carbon dioxide production).

2.15 |. Isolation of human and mouse myogenic progenitor cells (MPCs)

Human MPCs were isolated from discarded muscle tissue from the gracilis following informed consent as described, with some modifications.⁴⁰ Participants were young-healthy donors (age 28–37 y old). Briefly, muscles were excised and dissociated using Collagenase Type II (800 U/mL) and Dispase (2.4 U/mL) in wash media (Ham’s F-10 supplemented with 10% horse serum (HS) and 100 units/mL of Penicillin and 100 μg/mL of Streptomycin) for 1 h at 37°C with gentle agitation. The cell suspension was then passed through three cell strainers (100, 70, and 40 μm, Miltenyi Biotec). Following centrifugation at 300 g for 5 min at 4°C, the cell pellet was resuspended in red blood cell (RBC) lysis buffer and incubated at room temperature for 10 min. RBC lysis buffer was neutralized with wash media. Cells were pelleted by centrifugation at 300 g for 5 min at 4°C and resuspended in 500 μL of wash media. Fragment crystallizable (FC) block TruStain (1 μg/mL) was added prior to antibody incubations for 15 min on ice. Cells were sorted using fluorescence-activated cell sorting (FACS) for surface markers anti-human CD31-FITC (1:100, Biolegend), CD34-FITC (1:100, Biolegend), CD45-FITC (1:100, Biolegend), CD56-APC (1:20, Biolegend). MPCs were identified as CD31−/CD34−/CD45−/CD56+. Differentiation of human MPCs using MyoCult differentiation media was accomplished on Cytoo fibronectin coated myogenesis chips by day five.

Mouse MPC isolation was performed as described.⁴¹ Following FC block (TruStain (1 μg/mL) for 15 min on ice, cells were sorted using FACS for surface markers anti-mouse CD31-FITC (1:100, Biolegend), CD45-FITC (1:100, Biolegend), Sca1-FITC (1:100, Biolegend) and Vcam1-biotin (1:100, Biolegend) for 45 min followed by streptavidin-APC (1:100, Biolegend) for 30 min. MPCs were identified as CD31−/CD34−/CD45−/VCAM+. Isolated MPCs were cultured in Ham’s F-10 supplemented with 20% fetal bovine serum (FBS), 100 units/mL of Penicillin, and 100 μg/mL of Streptomycin, and 10 ng/mL basic fibroblast growth factor (FGF-basic) on coated plates (1:100 in DMEM). Differentiation of mouse MPCs was performed by replacing growth medium with DMEM supplemented with 5% horse serum and 100 units/mL of Penicillin and 100 μg/mL of Streptomycin for 3 days. For EV isolation from myotube conditioned media, exosome-depleted FBS was replaced 24 h prior to collection.

2.16 |. Electrical pulse stimulation (EPS) on human myotubes

Human MPCs were differentiated for 5 days with fresh media added prior to electrical pulse stimulation (EPS, Ionoptix C-Pace EP) for 8 h at 12 V, 1 Hz, 2 ms, followed by a 16 h rest period. Myotube conditioned media was collected 24 h post-EPS initiation from both EPS and non-EPS cells. Cells were then fixed in 4% paraformaldehyde and stained using a pan-myosin antibody (A4.1025 and DSHB) and DAPI (ThermoFisher Scientific). Cytoo chips were then mounted on glass slides using Vectashield mounting medium. Approximately 400 myotubes were quantified from each slide for myotube diameter. Myotubes were segmented from background by automatic determination of the myosin immunofluorescence intensity threshold. Myotube diameter was defined as the short edge length of the smallest rectangle that enclosed the entire myotube.

2.17 |. 3T3-L1 adipocyte lipolysis, EV treatment and transfection

3T3-L1 adipocyte cells were cultured as previous described.⁴² Briefly, cells were cultured to confluence in low glucose DMEM supplemented with 10% newborn calf serum and 100 units/mL of Penicillin and 100 μg/mL of Streptomycin. Two days post-confluence cells were induced to differentiate with DMEM supplemented with 10% FBS, 100 units/mL of Penicillin, and 100 μg/mL of Streptomycin, dexamethasone (1 μM), 3-isobutyl-1-methylxanthine (IBMX, 0.5 mM) and 0.32 μg/mL insulin (Novolin R) for 4 days. Cells were maintained in 10% FBS, 100 units/mL of Penicillin and 100 μg/mL of Streptomycin and 0.32 μg/mL insulin until day 16 of differentiation.

For assessment of lipolysis in differentiated 3T3-L1 adipocytes, cells were incubated overnight with serum-free media. Cells were then incubated with or without mouse myotube-derived EVs (3.11E+08) for a period of 6 h and then stimulated with isoproterenol (0.1 or 10 μM) for 2 h. Glycerol in the media was measured using Free Glycerol Reagent according to the manufacturer’s instructions. For EV treatment, myotube-derived EVs were added to the 3T3-L1 cells and incubated for a period of 24 h. Cells were harvested for RNA isolation.

2.18 |. Validation of Tfap2α as target gene of miR-1

Tfap2α 3′-UTR sequence was downloaded from UCSC Genome Browser and used for prediction of miR-1 binding site. The potential binding site was identified using RNAhybrid software.⁴³ To validate the predicted miR-1 binding site within the 3′-UTR of Tfap2α, a Tfap2α 3′-UTR luciferase reporter gene was constructed. Briefly, the mouse Tfap2α 3′-UTR was amplified from a skeletal muscle cDNA library and cloned into the XbaI site downstream of the luciferase gene in the pGL3-control vector (Promega). The miR-1 binding site within the cloned mouse Tfap2α 3′-UTR was mutated by altering four nucleotides within the predicted miR-1 binding site. Orientation and sequence of the wild-type and mutated Tfap2α 3′-UTR were confirmed by sequencing. NIH-3T3 cells (5 × 10⁴) were plated in a 12-well plate and transfected with 50 ng/well of Tfap2α 3′-UTR, with either scramble or miR-1 mimic (10 nM final concentration, Ambion, # 4464058 # 4464066). For normalization, Nanoluc luciferase vector (0.5 ng/well) was used. Transfection into NIH-3T3 cells was performed using Lipofectamine 2000 according to the manufacturer’s directions. Twenty-four-hour post transfection, cells were lysed, and luciferase activities measured using the Nano-Glo Luciferase Assay System according the manufacturer’s directions. For the positive control, the first 300 nucleotides of the Coro1c 3′-UTR (contains two 8-mer miR-1 binding sites) was cloned into Xbal site of the pGL3-control vector and transfected as described above. NIH-3T3 cells were cultured in Dulbecco’s modified Eagle’s medium with 10% FBS, 100 units/mL of Penicillin, and 100 μg/mL of Streptomycin.

2.19 |. miRNA and gene microarray analysis

The microarray hybridization and processing were performed at the University of Kentucky Microarray Core Facility according to the manufacturer’s protocol (Affymetrix, Santa Clara, CA). miRNA and gene expression were measured using Affymetrix miRNA 4.0 and Clariom S mouse chips, respectively. Briefly, 150 ng of total RNA was derived from a pooled sample of plantaris muscles or eWAT (vehicle or GW4869 treatment with one day of either sham or MOV) and then used for the hybridization. The criteria for a miRNA or mRNA to be considered differentially expressed was ≥1.5 increase or a ≥50% decrease in expression. Although a statistical analysis was not possible given the pooling approach, this strategy allowed us to identify and compare the most abundant miRNAs and mRNAs. To identify potential pathways regulated by miR-1, we queried for genes whose expression was higher in eWAT of vehicle-treated MOV mice compared to sham but showed little to no change in gene expression in eWAT of GW4869-treated MOV mice.

2.20 |. Resistance exercise (RE) bout in humans

Ten healthy male and female subjects (age range 26–50) including six males and four females (for subject characteristics, please see Table S1). All subjects were recreationally active (non-elite). Inclusion criteria were 18–50 years of age, and exclusion criteria were cardiovascular disease, neuromuscular disease or severe knee problems. The study was approved by the Regional Ethical Committee in Linköping, Sweden (Dnr 2017/183–31), was conducted according to the Declaration of Helsinki and applied to the laws of Sweden. At least five days prior to the exercise bout, subjects were familiarized with the experimental set-up. All subjects performed a submaximal test to estimate peak oxygen uptake,^44,45 and seven repetitions maximum (7RM) for knee extension and leg press was determined. These data were used to determine the load for the acute RE bout. Subjects were instructed not to perform any strenuous leg exercise three days prior to the exercise bout. Liquid formula (1.05 g carbohydrates/kg body weight (bw), 0.28 g protein/kg bw, and 0.25 g fat/kg bw) was provided as breakfast 1 h prior to collection of the pre-exercise sample. The liquid formulas contained 5.6 g protein, 21 g carbohydrates, and 5.0 g fat per 100 mL (Resource Komplett Näring, Nestlé Health Science, Stockholm, Sweden). A peripheral vein catheter was inserted in the cubital vein 30 min before sampling. Blood samples were drawn via the catheter and collected in EDTA containing vacutainer tubes (BD, Franklin Lakes, NJ) before the intervention and at 30 min post-exercise. Plasma was prepared accordingly to the manufacturer’s recommendations, aliquoted and stored at −80°C until further analysis. Following blood collection, muscle biopsies were obtained from the vastus lateralis muscle. The biopsies were obtained percutaneously after injection of local anesthetic (carbocain 10 mg/mL), by using a biopsy needle with a diameter of 5 mm (Stille AB, Torshälla, Sweden). Muscle biopsies were frozen in liquid nitrogen and then stored at −80°C until further analysis. The acute RE bout consisted of two lower limb exercises: leg press and knee extension. After a short warm-up on submaximal loads, the subjects performed 4 sets per exercise at 7 RM load with 2 min rest between sets and 5 min between exercises. Loads were decided during the familiarization.

2.21 |. Statistical analysis

All data were analyzed by the Sigma Plot 14 statistical software. All data are presented as the mean ± SE. In order to control for lean mass, indirect calorimetry data were analyzed using analysis of covariance (ANCOVA), in which independent variables included surgery group and lean mass with the dependent variable as energy expenditure. For the time-course experiments one-way ANOVA was performed and when a significant overall effect was detected, differences among individual means were assessed with Bonferroni’s post-hoc test. For GW4869 experiments, two-way ANOVA was performed. If a significant interaction or overall effect was detected, differences among individual means were assessed with Dunnett’s post-hoc test. In the cell culture experiments two-sided Student’s t test was performed. Finally, for changes in human muscle and plasma miR-1, a paired t test was performed. Statistical significance was set at P < .05.

2.22 |. Resource availability

The raw data was deposited in the NCBI Gene Expression Omnibus database (GSE15 0162).

3 |. RESULTS

3.1 |. MOV induces down-regulation of muscle-specific miR-1 associated with greater serum EV miR-1 abundance

To identify miRNAs that may have a role in regulating skeletal muscle hypertrophy, we performed a microarray analysis of RNA isolated from plantaris muscle subjected to either one or three days of MOV induced by synergist ablation, a well-established rodent model that has been used to identify the molecular and cellular mechanisms regulating muscle growth in response to resistance exercise. Under resting conditions, the muscle-specific miRNA, miR-1, was 2.1-fold more abundant than any other miRNAs and showed the greatest change in expression (69% lower) after three days of MOV (Figure 1A). qPCR confirmed plantaris muscle miR-1 expression was significantly lower by 40% after a single day of MOV and remained lower for the next two days (Figure 1B). There was no change in the expression of primary miRNA-1A and −1B gene expression indicating the lower miR-1 expression after a single day of MOV was not caused by a change in transcription of either miR-1 gene (Figure 1C). Based on earlier work of Hudson and colleagues showing the exosomal export of miR-23a during muscle atrophy,⁴⁶ we hypothesized miR-1 was being exported from the myofiber by EVs in response to the hypertrophic stimulus induced by MOV. To test our hypothesis, we isolated EVs from serum of mice subjected to 1–3 days of MOV and quantified miR-1 abundance. Isolation of serum EVs was confirmed by nanoparticle tracking size determination (~70–150 nm), detection of EV-enriched membrane proteins (HSP70, TSG101, CD81 and CD63) by western blot and transmission electron microscopy (Figure S1A–D). There was no change in serum EV concentration or size after one day of MOV compared to sham-control (Figure S1A,B); however, serum EV concentration was higher after 3 days of MOV with smaller EV size at day 2 and 3 of MOV (Figure S1A,B). Serum EV miR-1 abundance was substantially higher after one and two days of MOV compared to sham, returning to baseline by day 3 of MOV (Figure 1D). There was no change in the expression of miR-1 or the miR-1 gene in the gastrocnemius indicating the injury caused to the muscle by the synergist ablation surgery was not contributing to the higher level of EV miR-1 abundance in response to MOV (Figure S1E,F). Using an immune-capture based method to isolate EVs, we confirmed the abundance of serum EV miR-1 was higher following 1 day of MOV compared to sham (Figure S1G).

FIGURE 1.

Export of muscle-specific miR-1 via extracellular vesicles (EVs) in response to a hypertrophic stimulus. A, miRNA microarray analysis of the plantaris muscle after 1 and 3 days of mechanical overload (MOV) compared to sham surgery with the 10 most abundant miRNAs shown. B, miR-1 expression in response to 1–3 days of MOV by qPCR. C, Primary miRNA-1A or 1B (pri-miRNA-1A or 1B) expression in response to MOV. D, Abundance of miR-1 in serum EVs after MOV. Data is presented as mean ± SE (n = 10 for A-B; n = 8 for D-F) with significance denoted by an asterisk compared to sham or a cross compared to 3-day MOV

3.2 |. Hypertrophic stimulus-induced EV export of miR-1 is conserved in human skeletal muscle

We wanted to determine if a bout of high-intensity resistance exercise in humans could induce a similar response in miR-1 expression as observed in the mouse following MOV. Male and female subjects performed four sets of leg extension and leg press exercises at 7-RM (repetition maximum) load with two-minute rest between sets and five-minute rest between exercises. Muscle biopsy (vastus lateralis) and blood were collected prior to and 30 minutes post-exercise. In response to resistance exercise, skeletal muscle miR-1 expression was significantly decreased while plasma EV miR-1 abundance was significantly increased (Figure 2A,B). We next wanted to directly assess if a hypertrophic stimulus could specifically induce the down-regulation of muscle cell miR-1 expression and higher EV miR-1 abundance. Using an in vitro hypertrophy model, human primary myoblasts were differentiated into myotubes and then subjected to electric pulse stimulation (EPS) for 8 h with a 16 h rest period to induce myotube growth. In response to EPS, myotubes were 13% larger with lower miR-1 expression and ~12-fold higher media EV miR-1 abundance following 24 h (Figure 2C–G). Collectively, these findings demonstrate that both mouse and human skeletal muscle have a conserved response to a hypertrophic stimulus that involves the decrease of miR-1 abundance along with an increase in miR-1 in EVs.

FIGURE 2.

Human skeletal muscle down-regulate miR-1 associated with release of EV containing miR-1 in response to a hypertrophic stimulus. A, Decreased in miR-1 expression 30 min after a bout of resistance exercise compared to pre-exercise. B, Increased miR-1 abundance from EVs isolated from plasma 30 min post-exercise compared to pre-exercise. C, Representative images of human myotubes with or without electrical pulse stimulation (EPS). Myotubes were stained with a pan myosin antibody to accurately measure myotube diameter. D, Increased human myotube diameter 24 h post-EPS. E, Decreased miR-1 expression in primary human myotubes 24 h post-EPS and in (F) Increased EVs miR-1 abundance from culture media of EPS treated myotubes. G, No change in EV concentration isolated from culture media of myotubes with or without EPS. Data are expressed as mean ± SE (n = 10 for A-B; n = 3 for D-F; n = 7 for G) with significance (P < .05) denoted by an asterisk compared to respective sham

3.3 |. Preferential uptake of MOV-induced serum EVs by eWAT

We next performed fluorescence tomography to determine if MOV-induced serum EVs were being taken up by other tissues. DiR-labeled serum EVs, isolated from mice that had undergone either sham or one day of MOV, were tail vein injected into naive mice and scanned 24 h later. To achieve better resolution, we excised and scanned eWAT, heart, liver and kidney. Compared to injection with sham serum EVs, fluorescence was ~5-fold higher in eWAT and ~3-fold higher in the kidney following injection of MOV-induced EVs (Figure 3A). To confirm the fluorescence tomography results, we used ExoGlow to fluorescently label EV RNA from mice subjected to either sham or MOV. Labeled serum EVs were intraperitoneally injected into naive mice with tissues collected 6 h later and used to generate whole-cell lysates for fluorescence measurements. In comparison to sham mice, only eWAT showed a higher fluorescence signal following injection of MOV-induced serum EVs (Figure 3B). In addition, for whole-body visualization, we tail vein injected DiR-labelled EVs and observed higher fluorescence in the abdominal region only in mice which were injected with DiR-labeled EVs from mice that had undergone MOV and not sham-derived EVs or PBS-DiR control (Figure S2). Given the preferential uptake of MOV-induced serum EVs by eWAT, we speculated miR-1 levels would be higher in adipose tissue. As shown in Figure 3C, miR-1 abundance was 1.8-fold higher in eWAT following one day of MOV compared to sham as determined by qPCR. qPCR analysis showed no difference in miR-1 abundance in brain, liver and heart between sham and MOV groups (Figure S3). We also observed higher miR-1 abundance in subcutaneous adipose tissue demonstrating uptake of miR-1 from this fat depot in addition to eWAT (Figure S3). The higher level of miR-1 in eWAT following MOV was not associated with a change in pri-miRNA-1A or −1B gene expression as there was no difference between sham and MOV groups (Figure 3D). Collectively, these findings provide evidence that the higher miR-1 abundance in eWAT following MOV was the result of serum EV delivery of skeletal muscle-derived miR-1 to eWAT.

FIGURE 3.

Preferential uptake of MOV-induced serum EVs by eWAT. A, Representative IVIS image of individual organs (epidydimal white adipose tissue [eWAT], heart, liver and kidney) 24 h post-injection of DiR-labeled serum EVs isolated from mice subjected to sham or MOV. B, Fluorescence of whole-cell lysates of various tissues 6 h post-injection of ExoGlow labeled serum EVs isolated from mice subjected to sham or MOV. C, Increased in eWAT miR-1 abundance in response to MOV. D, No change in eWAT primary miRNA-1A and 1B expression in response to MOV. Data are expressed as mean ± SE (n = 2 for A; n = 3–5 for B; n = 8 for C-D) with significance (P < .05) denoted by an asterisk compared to respective sham or a cross compared to 3-day MOV

3.4 |. eWAT miR-1 abundance is dependent on the release of muscle-derived EVs in response to MOV

While the delivery of miR-1 from skeletal muscle fibers to eWAT was likely mediated by MOV-induced serum EVs, an alternative mechanism might involve the passive release of miR-1 as the result of MOV-induced muscle damage with subsequent transport to eWAT by lipoproteins⁴⁷ or Ago2.⁴⁸ To determine if the MOV-induced down-regulation of skeletal muscle miR-1 expression required EV release, mice were treated with either vehicle (DMSO) or GW4869 for six days prior to sham or MOV surgery and then until one day after surgery (Figure 4A). GW4869 is an inhibitor of neutral sphingomyelinase which has been shown to effectively block the release of EVs.³⁶ Nanoparticle tracking analysis confirmed GW4869 treatment reduced serum EV concentration by over 50% in both sham and MOV groups compared to the vehicle-treated groups (Figure 4B). As expected, vehicle-treated mice showed lower skeletal muscle miR-1 expression following a single day of MOV compared to sham mice which was prevented by GW4869 treatment (Figure 4C). The higher abundance of miR-1 in EVs and eWAT in response to MOV was also prevented by GW4869 treatment (Figure 4D,E). Together, these results provide compelling evidence the lower skeletal muscle miR-1 expression following MOV was mediated by EV export with the higher miR-1 abundance in eWAT a result of MOV-induced, skeletal muscle-derived EV delivery of miR-1.

FIGURE 4.

Inhibiting EV release prevents higher eWAT miR-1 abundance in response to MOV. A, Diagram of the treatment schedule for administering vehicle or GW4869 (GW) of mice by i.p. injection. B, GW treatment significantly reduced serum EV concentration as determined by nanoparticle tracking analysis. In response to MOV, GW treatment prevented (C) the down-regulation of miR-1 expression in skeletal muscle, (D) higher EV miR-1 abundance and (E) eWAT miR-1 abundance. Data are expressed as mean ± SE (n = 3 for B; n = 8 for C-E) with significance (P < .05) denoted by an asterisk compared to respective sham

3.5 |. MOV alters expression of genes involved in adrenergic signaling in eWAT

To determine if the increased abundance of miR-1 was driving changes in eWAT gene expression, we performed microarray analysis in eWAT following one day of MOV. To identify potential pathways regulated by miR-1, we queried the dataset for genes that were differentially affected in eWAT of vehicle-treated MOV mice but showed little to no change in expression in eWAT of GW4869-treated MOV mice. Using this criterion, we identified genes involved in adrenergic signaling and lipolysis (Figure 5A). Based on these differentially expressed genes, we developed a gene regulatory network in which the delivery of MOV-induced, skeletal muscle-derived EVs activates eWAT lipolysis (Atgl, adipose triglyceride lipase; Lipe, hormone-sensitive lipase; Plin1, perilipin) through enhanced catecholamine sensitivity via CCAAT/enhancer binding protein alpha (Cebpα) activation of adrenergic receptor beta 3 (Adrβ3) expression as a result of miR-1 repression of transcription factor AP-2 alpha (Tfap2α), a known repressor of Cebpα gene transcription (Figure 5B).⁴⁹ There was no difference in food intake or body weight in response to MOV with or without GW4869 treatment, suggesting the elevated expression of genes involved in lipolysis of eWAT vehicle-treated mice following MOV was not caused by lower food intake (Figure S4A,B).

FIGURE 5.

MOV of skeletal muscle induced eWAT adrenergic signaling. A, Microarray analysis of eWAT collected from vehicle or GW4869 treated mice subjected to sham or one day of MOV identified differentially expressed genes (DEGs) involved with adrenergic signaling and lipolysis. B, Based on DEGs, a proposed gene regulatory network in which EV-delivery of miR-1 induces Cebpα activation of Adrβ3 expression by repressing Tfap2α expression, a known repressor of Cebpα gene transcription that promotes eWAT lipolysis as a result of enhanced catecholamine sensitivity. C, RNAhybrid program identified a non-conventional miR-1 binding site within the 3′-UTR of the Tfap2α mRNA with mis-matches within the seed sequence but extensive homology in the 3′ region. D, Compared to scrambled oligonucleotide, transfection of miR-1 mimic into 3T3-L1 cells reduced luciferase activity of Coro1C (positive control) and Tfap2α_WT 3′-UTR reporter genes which was prevented when the predicted miR-1 binding site was mutated (Tfap2α_Mut). qPCR confirmed microarray analysis of eWAT gene expression in response to skeletal muscle MOV showed the change in expression of, (E) Tfap2α, (F) Cebpα and (G) Adrβ3 was prevented by GW4869-t reatment (GW) compared to vehicle-treated (Vehicle) mice. Data are expressed as mean ± SE (n = 5–8) with significance (P < .05) denoted by an asterisk compared to respective sham or as otherwise shown

To validate the first component of the proposed regulatory network, we used the RNAhybrid algorithm to identify potential miR-1 binding site(s) within the 3′-UTR of Tfap2α mRNA. A single non-canonical miR-1 binding site was identified within the Tfap2α 3′-UTR which contained mis-matches within the seed sequence but extensive homology within the 3′ region (Figure 5C).^50–52 To determine if Tfap2α is a miR-1 target gene, we generated 3′-UTR luciferase reporter genes harboring either the Tfap2α 3′-UTR with or without the predicted miR-1 sequence mutated (Tfap2α _WT and Tfap2α _Mut, respectively). As a positive control, we generated a 3′-UTR luciferase reporter gene harboring the 3′-UTR of Coro1c mRNA (Coro1C), the top predicted target gene of miR-1 according to TargetScan, to ensure the assay reaction conditions were optimal for detecting miR-1 regulation of luciferase expression. Luciferase activity of the Coro1c and Tfap2α 3′-UTR reporter genes were substantially lower in NIH-3 T3 cells transfected with a miR-1 mimic in comparison to a scrambled control oligonucleotide (Figure 5D). In contrast, there was no difference in luciferase activity following transfection of the miR-1 mimic compared to scrambled control oligonucleotide when the predicted miR-1 binding site within the Tfap2α 3′-UTR was mutated (Figure 5D). These results provide the first evidence demonstrating that miR-1 was capable of repressing Tfap2α expression through interaction within the 3′-UTR.

We next used qPCR to confirm the change in eWAT gene expression revealed by microarray analysis. Tfap2α transcript abundance was significantly lower by 90% with both Cebpα and Adrβ3 mRNA levels significantly higher by 1.6-fold compared to the respective sham mice (Figure 5E–G). In contrast, we detected no difference in the levels of Tfap2α, Cebpα and Adrβ3 transcripts in eWAT between sham and MOV of GW4869-treated mice (Figure 5E–G).

3.6 |. Enhanced catecholamine sensitivity in eWAT of mice undergoing MOV of skeletal muscle

We next wanted to determine if the upregulation of genes involved in adrenergic signaling and lipolysis of eWAT in response to skeletal muscle MOV resulted in enhanced catecholamine sensitivity and lipolysis. Using eWAT primary adipocyte culture, we measured the release of non-esterified fatty acids (NEFA) and glycerol in response to vehicle or 100 μM isoproterenol. Adipocytes isolated from mice following one day of MOV released considerably more NEFA and glycerol into culture media in response to isoproterenol compared to adipocytes isolated from sham mice (Figure 6A,B).

FIGURE 6.

Enhanced catecholamine sensitivity of eWAT in response to MOV of skeletal muscle. Ex vivo isoproterenol (100 μM) treatment stimulated greater release of (A) non-esterified fatty acids (NEFA) and (B) glycerol from eWAT adipocytes of mice subjected to skeletal muscle MOV compared to sham mice. C, Resting respiratory exchange ratio (RER) was lower in mice undergoing MOV compared to sham. D, Higher release of glycerol after isoproterenol (0.1 and 10 μM) treatment of 3T3L1 adipocytes pretreated with EVs isolated from mouse primary myotubes compared to untreated cells. E, miR-1 abundance after incubating NIH-3T3L1 adipocytes with EVs (3.11E+08) isolated from mouse primary myotubes for 24 h. F, Gene expression of Tfap2α, Cebpα and Adrβ3 after myotube-derived EV treatment. Data are expressed as mean ± SE (n = 7 for A-B; n = 11 for C; n = 4 for D-E; n = 5–6 for F) with significance (P < .05) denoted by an asterisk compared to cells not incubated with myotube EVs

Since we observed increased lipolysis ex vivo following MOV, we next wanted to determine if MOV would induce higher fatty acid oxidation in vivo. We used indirect calorimetry to measure respiratory exchange ratio (RER) following sham or MOV surgery. As expected, there was no difference in RER between sham and MOV groups prior to surgery (Figure S5A). As shown as Figure 6C, RER (as a function of lean mass) was significantly lower in mice subjected to MOV compared to sham mice the day after surgery despite there being no difference in food intake or body composition between the groups (Figure S5B,C). This finding indicates a higher level of fatty acid oxidation occurred in mice following MOV compared to sham, likely due to enhanced catecholamine sensitivity. Higher glycerol release by adipocytes incubated with myotube EVs (Figure 6D) was associated with higher miR-1 abundance (Figure 6E), and elevated Cebpα and Adrβ3 expression; however, we were unable to detect lower Tfap2α transcript abundance following myotube EV incubation which was likely due to the fact that Tfap2α expression is already significantly down-regulated as a result of 3T3-L1 adipocyte differentiation.⁵³ In this scenario, myotube-derived EV miR-1 is unable to effectively interact with the rare Tfap2α transcript in 3T3-L1 adipocytes thereby preventing a further down-regulation of Tfap2α expression.

4 |. DISCUSSION

The novel finding of this study is that skeletal muscle, in response to a hypertrophic stimulus, promoted eWAT lipolysis as a result of enhanced catecholamine sensitivity. During MOV, skeletal muscle released miR-1 containing EVs which were preferentially taken up by eWAT. We show that the increase in eWAT miR-1 abundance following EV uptake is associated with higher Adrβ3 expression and repression of Tfap2α, a known negative regulator of both Cebpα and Adrβ3 expression. We demonstrate that Tfap2α is a target of miR-1, and treatment of adipocytes with myotube-derived EVs increases Adrβ3 expression and lipolysis in vitro. Although the stimulation of adipocyte lipolysis via adrenergic signaling has been well-described, this is the first demonstration showing adrenergic signaling can be enhanced via intercellular communication with skeletal muscle.^54–56 We also found the release of miR-1 containing EVs from skeletal muscle in response to a hypertrophic stimulus was conserved in humans suggesting a similar skeletal muscle-adipose axis may be operative in humans following resistance exercise.

The molecular details of a potential human skeletal muscle-adipose axis await further investigation but will need to reconcile the differences in β-adrenergic receptor isoform expression between humans and mice. Although all three isoforms (Adrβ1, Adrβ2, and Adrβ3) are expressed in adipocytes of both species, Adrβ3 is considered the major regulator of adipose lipolysis due to its high-level of expression in the mouse,⁵⁷ whereas in human adipose tissue, ADRβ1 and ADRβ2 are the most abundant isoforms.⁵⁸ Despite the modest expression of ADRβ3 in human adipocytes, ADRβ3-specific agonists (mirabegron and CL-316,243) have been shown to improve glucose homeostasis, insulin action, and fat oxidation.^59–62 Together, these studies suggest an increase expression of ADRβ3, as a result of the proposed skeletal muscle-adipose axis, has the ability to impact adipocyte metabolism. Alternatively, there are predicted TFAP2α binding sites in the promoter of both ADRβ1 and ADRβ2 in humans such that EV delivery of miR-1 could directly induce ADRβ1 and/or ADRβ2 expression by repressing TFAP2α expression. Future studies will seek to determine if the observed intercellular communication between skeletal muscle and eWAT is functional in humans and which β-adrenergic receptor isoform is involved in promoting lipolysis.

The rodent synergist ablation model used to study skeletal muscle hypertrophy is often criticized for not fully mimicking resistance exercise, the most commonly used modality to induce muscle hypertrophy in humans. Like resistance exercise, however, synergist ablation places a MOV on the muscle but, unlike resistance exercise, the MOV is continuous, and not intermittent, and as a result induces a robust and rapid hypertrophy much greater than what is observed following resistance exercise. Despite these differences, almost all of the molecular and cellular factors shown to be involved in regulating hypertrophic growth in humans in response to resistance exercise were first described using synergist ablation. For instance, increased protein synthesis, mTORC1 regulation of protein synthesis, satellite cell activation, myonuclear accretion and ribosome biogenesis were originally reported using synergist ablation and then subsequently found to occur in human skeletal muscle in response to resistance exercise.^63–68 Similarly, we found both synergist ablation-induced MOV and high-intensity resistance exercise cause the down-regulation of skeletal muscle miR-1 which, at least in the mouse, is dependent on EV export. One distinction between the two models that raises a question is the significant difference in the amount of active muscle undergoing mechanical loading—mouse plantaris muscles (~28 mg of 25 g mouse) versus human quadriceps, hamstring and gluteus muscles (~14 kg of 70 kg person). While it seems remarkable the relatively small muscle mass of the plantaris muscles is capable of releasing sufficient miR-1 containing EVs to alter eWAT gene expression, the continuous nature of the overload with synergist ablation appears sufficient to significantly increase miR-1 abundance in eWAT that is reduced by GW4869 treatment, suggesting a role for EVs in this process. In agreement with the finding, skeletal muscle was shown to produce and release one-hundred times more EVs compared to WAT.⁶⁹

While exercise, mainly aerobic exercise, has been shown to stimulate lipolysis in adipose tissue, exactly how skeletal muscle contractions promote changes in adipose metabolism is not completely understood. There is convincing evidence in humans and mice of crosstalk between skeletal muscle and adipose tissue through the release of myokines in response to exercise.^70–72 For example, aerobic exercise has been shown to stimulate the release of irisin from skeletal muscle which subsequently affects adipose metabolism through increased thermogenesis.⁷⁰ Likewise, IL-6, one of the most studied myokine thus far, was found to stimulate lipolysis of adipose tissue after a bout of aerobic exercise.⁷² More recently, it was found that growth and differentiation factor 15 (Gdf15) was secreted by contracting skeletal muscle and promoted lipolysis in human adipocytes.⁷³

The findings of the current study add a new dimension to how skeletal muscle communicates with other tissues, eWAT in particular, by employing an EV-dependent mechanism through which resistance exercise regulates adipose tissue metabolism. The intercellular communication we identified between skeletal muscle and adipose tissue is based on the preferential uptake of skeletal muscle-derived EVs to eWAT in response to a hypertrophic stimulus. The targeting of skeletal muscle-derived EVs to eWAT required the imposition of MOV as we did not observe any difference in EV uptake across many different tissues prior to MOV. This finding indicates that in response to MOV, skeletal muscle produced unique EVs which are preferentially taken up by eWAT. An alternative explanation, though not mutually exclusive, is eWAT underwent remodeling in response to MOV which made adipocytes more receptive to skeletal muscle EVs uptake. Following a high-intensity bout of cycling, exercise-induced EVs were shown to be preferentially taken up by the liver.²¹ Based on these findings, it is reasonable to speculate that different forms of exercise (endurance vs resistance) will result in the production of sub-types of skeletal muscle-derived EVs with differing cargo that are targeted to different cell-types based on the composition of EVs and recipient cell surface proteins. The development of methods to label cell-specific EVs for subsequent isolation and characterization (cargo and membrane protein composition) will be critical for understanding how EVs are targeted to a particular cell-type and how the delivery of EV cargo affects recipient cell metabolism.

Adipose tissue is localized to almost all compartments of the body, consisting of white and brown fat.⁷⁴ Adipose tissue has been historically considered merely a storage depot for lipids; however it is now recognized as an important endocrine organ in the regulation of the metabolic homeostasis⁷⁵ with aberrant adipose tissue remodeling leading to metabolic stress and disorders in metabolic organs.^76,77 Thus, understanding the regulation of adipose tissue metabolism by resistance training may identify new therapeutic targets for obesity and its related diseases.⁷⁸ Despite the fact that resistance exercise has been shown to impact adipose tissue,⁷⁹ the majority of the studies have been focused on the benefits provided by aerobic exercise. Usually these benefits are attributed to an increase in adipose tissue lipolysis and whole-body fat oxidation for improvement of energy expenditure in response to exercise.³⁴ It has been shown that resistance exercise reduces intramuscular triglycerides stores with the likely possibility such triglyceride stores were used for fuel during the exercise bout.⁸⁰ This evidence together with our findings suggest that the EV-mediated crosstalk between skeletal muscle and adipose tissue provides a novel mechanism for coordinating the supply and utilization of energy during resistance exercise. In conclusion, we have shown that in response to MOV, skeletal muscle releases EV-containing miR-1 which are preferentially taken up by eWAT that leads to greater lipolysis as the result of enhanced catecholamine sensitivity via elevated Adrβ3 expression (Figure 7).

FIGURE 7.

Proposed mechanism of how mechanical loading of skeletal muscle promotes adipose tissue lipolysis. Mechanical loading, via synergist ablation or resistance exercise, promotes a down-regulation of skeletal muscle miR-1 levels through extracellular vesicle (EV) release. EVs containing miR-1 are then preferentially taken up by white adipose tissue where miR-1 targets Tfap2α, a known repressor of Adrβ3 expression; subsequently, the increase in Adrβ3 expression promotes lipolysis as the result of enhanced catecholamine sensitivity.

Abbreviations:

Adrβ3

adrenergic receptor beta 3

Atgl

adipose triglyceride lipase

Cebpα

CCAAT/enhancer binding protein alpha

Coro1C

coronin-1C

EPS

electric pulse stimulation

EVs

extracellular vesicles

eWAT

epidydimal white adipose tissue

Lipe

hormone-sensitive lipase

Mmp9

metalloproteinase 9

MOV

mechanical overload

NEFA

non-esterified fatty acids

Plin1

perilipin

Rrbp1

ribosome-binding protein 1

RER

respiratory exchange rate

Tfap2α

transcription factor AP-2 alpha