Relationships between endurance exercise training-induced muscle fiber-type shifting and autophagy in slow- and fast-twitch skeletal muscles of mice
Article information
Abstract
[Purpose]
Endurance exercise induces muscle fiber-type shifting and autophagy; however, the potential role of autophagy in muscle fiber-type transformation remains unclear. This study examined the relationship between muscle fiber-type shifting and autophagy in the soleus (SOL) and extensor digitorum longus (EDL) muscles, which are metabolically discrete muscles.
[Methods]
Male C57BL/6J mice were randomly assigned to sedentary control (CON) and exercise (EXE) groups. After 1 week of acclimation to treadmill running, the mice in the EXE group ran at 12-15 m/min, 60 min/day, 5 days/week for 6 weeks. All mice were sacrificed 90 min after the last exercise session, and the targeted tissues were rapidly dissected. The right side of the tissues was used for western blot analysis, whereas the left side was subjected to immunohistochemical analysis.
[Results]
Endurance exercise resulted in muscle fiber-type shifting (from type IIa to type I) and autophagy (an increase in LC3-II) in the SOL muscle. However, muscle fiber-type transformation and autophagy were not correlated in the SOL and EDL muscles. Interestingly, in contrast to the canonical autophagy signaling pathways, our study showed that exercise-induced autophagy concurs with enhanced anabolic (increased p-AKTSer473/AKT and p-mTOR/mTORSer2448 ratios) and suppressed catabolic (reduced p-AMPKThr172/AMPK ratio) states.
[Conclusion]
Our findings demonstrate that chronic endurance exercise-induced muscle fiber-type transformation and autophagy occur in a muscle-specific manner (e.g., SOL). More importantly, our study suggests that endurance training-induced SOL muscle fiber-type transition may underlie metabolic modulations caused by the AMPK and AKT/mTOR signaling pathways rather than autophagy.
INTRODUCTION
Skeletal muscle is a heterogeneous tissue that sustains plasticity to adapt to various stimuli, such as hormones, cytokines, nutritional status, and exercise. For example, exercise training strongly influences the structural, metabolic, and functional characteristics of individual muscle fiber types [1-3], determined by myosin heavy chain (MHC) isoforms and biochemical properties. Human skeletal muscles include three MHC isoforms, namely type I, type IIa, and type IIx fibers, whereas rodent skeletal muscles, in addition to types I and IIa, contain type IIx/d and type IIb fibers [4-6]. Specifically, type I fibers are characterized as slow-twitch fibers owing to their slow contraction speed and predominantly oxidative metabolic capacity, which confers the highest endurance capacity and energy efficiency elicited by a higher mitochondrial content. In contrast, type IIx and IIb fibers are classified as fast-fatigable-twitch fibers because of their fast contraction speed and predominantly anaerobic metabolic capacity owing to their faster ATPase activity. Moreover, type IIa fibers are characterized as intermediate (hybrid)-twitch fibers that exhibit both fast and slow contractile speeds and biochemical characteristics [6-8].
The patterns of exercise-induced intracellular signaling modulations distinctively reinforce muscle fiber-type shifting and regulate the rate of muscle protein turnover, ensuring proper adaptation through the reprogramming of transcriptional and translational processes [9]. For instance, endurance exercise training, which requires prolonged repetitive cycles of muscle contraction and relaxation, induces a muscle MHC phenotype from glycolytic to oxidative metabolism [10,11], increases mitochondrial biogenesis, capillary density, fatigue resistance, and endurance performance [12-14], and improves insulin sensitivity [15-17].
Accumulating evidence has demonstrated that a series of molecular signaling interplays, including calcium-dependent calmodulin/calcineurin, nuclear factor of activated T-cells, adenosine monophosphate-activated protein kinase (AMPK), sirtuin 1, and p38γ mitogen-activated protein kinase plays a pivotal role in endurance exercise-induced muscle phenotypes [12,18-20]. However, the underlying mechanism remains unclear.
Macroautophagy (hereinafter referred to as autophagy) is a major catabolic system that maintains cellular homeostasis by removing misfolded, damaged, and unwanted proteins, lipids, and DNA components in a lysosome-dependent manner. Indeed, energy disturbance caused by a nutrient-deficient state promotes autophagy to ensure the steady energy state necessary for cell survival by promoting catabolism [21-23]. Recent studies have suggested that autophagy is essential for the physiological adaptation of skeletal muscle in response to regular exercise training [24,25]. Exercise-induced autophagy regulates protein modulation, such as unc-51-like autophagy activating kinase 1 (ULK1), AMPK, AKT (protein kinase B), and mammalian target of rapamycin (mTOR), which are involved in each autophagy process, including initiation, nucleation, elongation, maturation, fusion, and degradation [26]. In addition, a recent study has shown that increased phosphorylation of the B-cell lymphoma 2 (BCL2) protein is essential for endurance exercise-induced autophagy in skeletal muscle, as the mutation of phosphorylation sites of BCL2 abolishes it [27].
Over the past two decades, many researchers have focused on identifying the mechanisms of exercise-induced autophagy and its physiological and pathological functions [28-33]. However, whether autophagy contributes to muscle fiber-type transformation remains unclear. Therefore, this study aimed to examine the relationship between exercise-induced muscle fiber-type shifting and autophagy in slow-twitch (endurance training-dominant) and fast-twitch (endurance training-recessive) muscles.
METHODS
Animals
Seven-week-old male C57BL/6J mice (N=14, 8-week-old) were purchased from ENVIGO company (Tampa, FL, USA) and housed in a temperature (22˚C) and humidity (55%)-controlled room under a 12:12-hour light/dark cycle with food and water ad libitum. After 1 week of environmental acclimation, the mice were divided into sedentary control (CON, n=7) and endurance exercise training (EXE, n=7) groups. All animal handling and experimental procedures were conducted in accordance with a protocol approved by the Institute of Animal Care and Use Committee of the University of West Florida (2015002).
Endurance exercise training
Endurance exercise intervention was conducted on a modified motor-driven animal treadmill, as described in a previous study [34]. The mice in the EXE group were familiarized with the treadmill at a speed of 8-12 m/min for 30 min/d for 5 consecutive d (1st day: 8 m/min; 2nd day: 9 m/min; 3rd day: 10 m/min; 4th day: 11 m/min; 5th day: 12 m/min). Then, 6 weeks of exercise training proceeded as follows: 1st week: 12 m/min (60% of maximal oxygen consumption [VO2max]), 60 min/d, 5 d; 2nd week: 13 m/min (65% of VO2max), 60 min/d, 5 d; and 3rd – 6th weeks: 15 m/min (80% of VO2max), 60 min/d, 5 d (Fig. 1). The range of training intensity has been predicted as 60-80% of VO2max according to a previous study [35] and the efficacy of the exercise protocol for this subject was validated by other experimental studies [36-38].
Tissue collection
To observe the peak response of exercise-induced autophagy as previously demonstrated [39], 90 min after the last exercise training session, the mice were sacrificed via cervical dislocation, and the soleus (SOL) and extensor digitorum longus (EDL) muscles were rapidly dissected from both legs. We chose the SOL because it is a predominant slow-twitch muscle with a higher proportion of type I MHC and the EDL because it predominantly contains fast-twitch fibers accompanying type IIb MHC [40]. The left side of the targeted muscles was embedded with an optical cutting temperature freezing medium, frozen in isopentane pre-chilled in liquid nitrogen, and kept at -80˚C until immunohistochemical analysis. The right side of the targeted muscles was immediately frozen in liquid nitrogen in cryogenic tubes and kept at -80˚C until western blot analysis.
Skeletal muscle fiber-type composition and cross-sectional area
Frozen SOL and EDL muscles were sliced into 10 µm-thick sections in the middle spot using a cryostat microtome (LEICA CM1860; Leica Biosystems, Germany), and serial transverse-sectioned samples were collected on microscope slides. The tissue samples were air-dried for 30 min at room temperature (RT) and were permeabilized with 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 30 min at RT, followed by 5 min of washing with PBS. Sectioned samples were delimited in a circle using an ImmEdge™ pen (H-4000; Vector Laboratories, USA) and blocked with 5% normal goat serum blocking solution (50062Z; ThermoFisher Scientific, Waltham, MA, USA) for 1 h at RT, followed by washing thrice with PBS for 5 min each. Then, two different Fab blockers were applied to eliminate nonspecific cross-reactivity: (i) AffiniPure Fab Fragment Goat Anti-Mouse IgG (#115-007-003; Jackson ImmunoResearch Laboratories Inc.) and (ii) AffiniPure Fab Fragment Goat Anti-Mouse IgM, µ Chain Specific (#115-007-020; Jackson ImmunoResearch Laboratories Inc.). After 1 h of incubation at RT, the tissues were washed thrice with PBS. Next, the corresponding primary antibodies were applied to each sample and incubated overnight at 4˚C: dystrophin (1:100, #PA1-21011; ThermoFisher Scientific), MHC type IIa (1:25, #SC-71; Developmental Studies Hybridoma Bank, USA), and MHC type I (1:10, #A4.840; Developmental Studies Hybridoma Bank) for SOL muscles, or MHC type IIb (1:50, #BF-F3; Developmental Studies Hybridoma Bank) for EDL muscles. The sectioned slides were then washed with PBS (3 × 5 min) and incubated with the corresponding secondary antibodies for 1 h at RT in a dark environment: Alexa Fluor 488 (1:50, #111-545-144; Jackson ImmunoResearch Laboratories Inc.), AMCA 350 (1:50, #115-155-020; Jackson ImmunoResearch Laboratories Inc.), and Cy™ 3 (1:50, #115-165-071; Jackson ImmunoResearch Laboratories Inc.). Next, the sections were washed with PBS (3 × 5 min), semi-dried, mounted with VECTASHIELD®Hard+Set™ Mounting Medium (H-1400; Vector Laboratories), covered with a cover glass, and sealed with clear nail polish. The fluorescence images were visualized and captured using an EVOS Auto fluorescence microscope (#AMAFD1000; Life Technologies, CA, USA) with three filters (4′,6-diamidino-2-phenylindole, green fluorescent protein, and red fluorescent protein).
The boundaries of each muscle fiber were demarcated using dystrophin staining (green). The percentage of the MHC isoform composition of the SOL was measured by counting the number of each muscle fiber type: type I (blue), type IIa (red), and type IIx+IIb (black). In contrast, the percentage of MHC isoforms in the EDL was measured by counting the number of each muscle fiber type: type IIa (red), type IIx (black), and type IIb (blue). The cross-sectional area (CSA, µm2) of myofibers was obtained by measuring the circumferences (green-colored dystrophin) of each fiber (60 random muscle fibers from each of the five random domains) using ImageJ software (NIH, USA).
Protein extraction and western blot analysis
The muscle tissues were homogenized 1:20 (w/v) in T-PER® buffer (#78510; ThermoFisher Scientific) containing a Halt™ Protease and Phosphatase Inhibitor Cocktail (#78446; ThermoFisher Scientific) using a Polytron™ PT 2500E Homogenizer (#08-451-320; Kinematica, Bohemia, NY, USA). Tissue homogenates were centrifuged at 14,000 × g for 15 min, and the total protein concentration was assessed using a Pierce™ Coomassie Plus reagent (#23236; ThermoFisher Scientific). For western blot assays, 40 μg proteins were separated via sodium dodecyl-sulfate poly-acrylamide gel electrophoresis using Bolt™ 4-12% and 12% Bis-Tris Plus Gels (#NW04125BOX and #NW00125BOX; Invitrogen, Carlsbad, CA, USA) for 1 h at 100 volts and then transferred onto polyvinylidene difluoride membranes for 1 h at 30 volts. Non-specific proteins were blocked for 1 h at RT in a blocking solution (5% bovine serum albumin in Tris-buffered saline containing 0.1% Tween-20 [TBS-T]), and the membranes were then incubated with the corresponding primary antibodies overnight at 4˚C. The antibodies used are as follows: BCL2 (#7382) and p-BCL2Ser87 (#377576) from Santa Cruz Biotechnology (TX, USA); microtubule-associated protein 1 light chain protein 3 (LC3) (#4108), Beclin1 (BECN1) (#3738), autophagy-related 7 (ATG7) (#2631), p62/sequestosome 1 (SQSTM1) (#5114), AMPKa (#2532), p-AMPKaThr172 (#2535), AKT (#9272), p-AKTSer473 (# 9271), mTOR (#2972), p-mTORSer2448 (#2971), ULK1 (#8054), p-ULKSer555 (#5869), and p-ULK1Ser757 (#14202) from Cell Signaling Technology (MA, USA); lysosome-associated membrane protein 2 (LAMP2) (#PA1-655) from ThermoFisher Scientific; and cathepsin L (CTSL) (#133641) from Abcam (Cambridge, UK). After overnight incubation, the membranes were washed thrice with 1X TBS-T for 10 min and then incubated with the corresponding horseradish peroxidase-conjugated secondary antibodies for 1 h at RT: goat anti-rabbit IgG (#G-21234) and goat anti-mouse IgG (#62-6520) both from ThermoFisher Scientific. Then, the membranes were washed thrice with 1X TBS-T for 10 minutes, after which immunoreactivity was detected using a SuperSignal™ West Dura Extended Duration Substrate (#34076; ThermoFisher Scientific). Band images were acquired using a ChemiDoc™ XRS + System (#1708299; Bio-Rad Laboratories, CA, USA), and band intensities were quantified using Image Lab™ software (version 6.1; #1709690; Bio-Rad Laboratories). The values of Ponceau (#P7170; Sigma-Aldrich, St. Louis, MO, USA)-stained total proteins were used to normalize the target proteins on the same membrane. The band intensity of the protein of interest was divided by that of the entire lane of the Ponceau-stained proteins of the protein of interest.
Statistical analysis
All data analyses were conducted using GraphPad Prism® software (version 10.0; San Diego, CA, USA). Data are presented as means ± SEM. An independent t-test was used to examine the significant differences between the CON and EXE groups. The relationship between LC3-II protein expression and each of the targeted muscle fiber types was examined using Pearson’s correlation coefficient (r). Statistical significance was set at p < 0.05.
RESULTS
Endurance exercise training induces muscle fiber-type transformation in a muscle-specific manner.
To verify whether moderate-to-high-intensity endurance exercise training (60-80% of VO2max) is sufficient to induce muscle fiber-type transformation, MHC isoform distribution was measured in both slow- and fast-twitch muscle fibers. Our results showed that endurance exercise significantly increased type I fibers (8.2%), decreased type IIa fibers (10.7%), and maintained type IIx/IIb fibers in SOL muscle (Fig. 2A and B). The CSA of MHC type I and type IIa fibers remained the same, whereas that of MHC type IIx+IIb fibers decreased significantly by 11.8% in the EXE group compared with the CON group (Fig. 2A and C). Next, we measured the same variables in the fast muscle EDL. Unlike in the SOL, endurance exercise training failed to alter MHC isoforms in the EDL muscle (Fig. 2D and E). Interestingly, however, endurance exercise training augmented the CSA of MHC type IIx and type IIb fibers by 15.4% and 8.7%, respectively, compared with that in the CON group (Fig. 2D and F).
Endurance exercise training-induced autophagy does not correlate with muscle fiber-type transformation.
To examine whether endurance exercise training promotes autophagy, we assessed the levels of LC3-I and LC3-II and the LC3-II/I ratio in the SOL and EDL as autophagy markers. In the SOL muscle, LC3-II protein levels in the EXE group significantly increased by 16.2% compared with those in the CON group, while LC3-I levels and the LC3-II/I ratio remained unchanged between the CON and EXE groups (Fig. 3A and B). In the EDL muscle, no significant differences were found in the LC3-II and LC3-II/I ratio between the two groups, whereas the EXE group displayed an increase in LC3-I protein levels by 6.6% compared with the CON group (Fig. 3A and C).
Next, we examined whether exercise training-induced autophagy correlated with fiber-type conformation in the SOL and EDL muscles. Our data showed no correlation between LC3-II protein expression (%) and each type of MHC isoform-stained myofiber (%) in the SOL (Fig. 3D-F) and EDL muscles (Fig. 3G-I).
Endurance exercise training-induced autophagy signaling is differently regulated in slow- and fast-twitch skeletal muscles.
To understand the nexus of the autophagy signaling pathway in response to endurance exercise training, we examined autophagy signaling modulators in SOL and EDL muscles. Since the dissociation of BECN1 from the BCL2-BECN1 complex upon BCL2 phosphorylation is crucial for the initiation of exercise-induced autophagy, we assessed the phosphorylation, total, and ratio of BCL2 proteins. In SOL muscles, total BCL2 levels did not change in both the CON and EXE groups; however, interestingly, despite an increase in LC3-II levels (Fig. 3A and B), the p-BCL2Ser87 levels decreased by 19% in the EXE group compared with those in the CON group, leading to an 11% reduction in the p-BCL2Ser87/BCL2 ratio (Fig. 4A and B). No significant alterations were found in total BCL2, p-BCL2Ser87, and the p-BCL2Ser87/BCL2 ratio in the EDL muscle in response to exercise (Fig. 4A and C), consistent with the finding of unchanged LC3-II levels (Fig. 3A and C).
Next, we assessed BECN1 protein levels as they play an important role in generating phagophores (the isolation membrane, an essential component of mature autophagosomes). The BECN1 levels in SOL muscles did not differ between the CON and EXE groups; interestingly, however, their levels in EDL muscles were significantly higher (10.4%) in the EXE group than in the CON group (Fig. 4A and D), even though LC3-II levels remained unchanged (Fig. 3A and C). Neither autophagy-related 7 (ATG7) nor p62/SQSTM1 levels were altered by exercise training in the SOL or EDL muscles (Fig. 4A, E, and F).
LAMP2 mediates the fusion of autophagosomes with lysosomes, which is important for autophagy; thus, we measured LAMP2 levels. The levels in the SOL muscles remained unchanged in the CON and EXE groups; in contrast, their levels in the EDL muscles were higher by 4.8% in the EXE group than in the CON group (Fig. 4A and G). Next, we measured the levels of the lysosomal protease CTSL, which plays a pivotal role in the final catabolic step of autophagy. Notably, CTSL in the SOL muscles was down-regulated by 6% in the EXE group compared with the CON group. In contrast, no differences were observed in the EDL muscle between the two groups (Fig. 4A and H).
Endurance training-induced muscle fiber-type transition in slow-twitch muscles depends on metabolic signaling rather than autophagy.
Endurance training is a potent inducer of autophagy and muscle fiber-type transformation, for which muscle fiber-type-specific metabolic signaling modulations are necessary. First, we examined AMPK, a critical kinase associated with catabolic processes, including autophagy and muscle fiber-type shifting (fast- to slow-type). No significant difference was found in the total AMPK protein levels in SOL muscles between the CON and EXE groups, whereas p-AMPKThr172 levels and the p-AMPKThr172/AMPK ratio significantly decreased by 43.8% and 44.3%, respectively, in the EXE group compared with the CON group (Fig. 5A and B). Similarly, in the EDL muscle of the EXE group, total AMPK levels remained unchanged whereas p-AMPKThr172 levels and the p-AMPKThr172/AMPK ratio significantly decreased by 46.2% and 45.2%, respectively, compared with those in the CON group (Fig. 5A and C).
Next, we examined two kinases, that are critically linked to anabolism, namely AKT and mTOR. In the SOL muscle, no significant changes in total AKT protein levels were observed between the CON and EXE groups whereas p-AKTSer473 levels and the p-AKTSer473 /AKT ratio were significantly higher in the EXE group than in the CON group (Fig. 5A and D). Similarly, in the EDL muscles, no significant changes in total AKT protein levels were observed between the CON and EXE groups whereas p-AKTSer473 levels and the p-AKTSer473/AKT ratio were significantly higher in the EXE group than in the CON group (Fig. 5A and E).
mTOR is a downstream target of AKT that potentiates protein synthesis. Our results showed that in the SOL muscles, no significant differences in the total mTOR and p-mTORSer2448 levels were observed between the CON and EXE groups. In contrast, the p-mTORSer2448/mTOR ratio was significantly higher in the EXE group than in the CON group, owing to a strong trend for increased p-mTORSer2448 levels in the EXE group (Fig. 5A and F). Unlike in the SOL muscle, in the EDL muscles, neither total mTOR nor p-mTORSer2448 protein levels were altered in the CON and EXE groups (Fig. 5A and G).
Modulation of catabolic and anabolic states via competitive reciprocal inhibitory interactions between AMPK and mTOR determines the overall metabolic state. For example, AMPK-induced mTOR inhibition or ULK1 phosphorylation at serine555 facilitates autophagy activation, whereas mTOR-induced ULK1 phosphorylation at serine757 antagonizes autophagy-induced catabolism to promote anabolism. To examine these metabolic interactions, we investigated ULK1 activity (p-ULK1Ser555) regulated by AMPK. No significant differences in total ULK1 protein, p-ULK1Ser555 levels, and the p-ULK1Ser555/ULK1 ratio in the SOL and EDL muscles were observed between the CON and EXE groups (Fig. 5A, H, and I). In terms of ULK1 activity (p-ULK1Ser757) regulated by mTOR, no significant changes in total ULK1 protein, p-ULK1Ser757 levels, and the p-ULK1Ser757/ULK1 ratio were observed in the SOL muscle (Fig. 5A and J). In contrast, in the EDL muscles, the p-ULK1Ser757/ULK1 ratio in the EXE group increased owing to a trend toward a slight increase in p-ULK1Ser757, although no significant changes were observed in total ULK1 protein and p-ULK1Ser757 levels (Fig. 5A and K).
DISCUSSION
The underlying molecular mechanisms of skeletal muscle adaptation caused by chronic endurance exercise training are relatively unclear compared to those of acute endurance exercise training. Moreover, while regular endurance exercise is known to be a prime mode of exercise for inducing muscle fiber transformation and has emerged as a potent inducer of autophagy, whether exercise-induced autophagy contributes to muscle fiber-type transformation in the predominantly slow SOL and fast EDL muscles remains unknown. Thus, using a mouse model, the present study, aimed to identify whether skeletal muscle remodeling induced by regular endurance exercise training (treadmill running) is associated with autophagy and to examine molecular crosstalk, focusing on metabolic signaling patterns.
Our results demonstrated that 6 weeks of endurance treadmill running is sufficient to induce muscle fiber type transformations in a muscle-specific manner; exercise training increased type I fibers (oxidative slow-twitch), decreased type IIa fibers (oxidative fast-twitch), and reduced the CSA of type IIx+IIb fibers (glycolytic fast-twitch) in the SOL muscle. In contrast, in the EDL muscle, exercise training failed to induce muscle fiber-type transformation from fast- to slow-twitch but instead promoted an increase in the CSA of type IIx and IIb fibers. These findings suggest that chronic endurance exercise training preferentially targets the oxidative SOL muscles and promotes endurance. However, the lack of muscle fiber-type shifting in glycolytic EDL muscles in our study may be because the total duration of endurance exercise may not have been long enough to promote muscle fiber-type transformation. For instance, although the exercise mode was different, a previous study reported that long-term exercise training (5 months of voluntary wheel running) alters muscle fiber-type compositions by converting fast (type IIx) to intermediate (type IIa) fibers but not slow (type I) fibers in the plantaris muscle [41].
Autophagy is an intricately controlled catabolic system, a proper extent of which is necessary to preserve muscle mass and myofiber integrity [42,43]. Nutrient deficiency and exercise are known to be potent stimuli for autophagy. However, the cellular and molecular mechanisms underlying skeletal muscle phenotypes differ [21,31,32,44]. For example, starvation-induced autophagy is associated with preferential atrophy in fast-twitch plantaris muscles compared to slow-twitch SOL muscles [45], whereas endurance training-induced autophagy is positively correlated with oxidative SOL muscles compared to fast-twitch plantaris and white vastus lateralis muscles [46]. Consistent with this, our data showed that autophagy (LC3-II) occurred only in the SOL muscle and not in the EDL muscle. The mechanisms by which exercise-induced autophagy is specifically observed in the SOL muscle compared to that in fast muscles are still being investigated. However, a potential mechanism seems to converge with the fact that the high oxidative capacity of slow muscles containing more mitochondria may confer higher sensitivity to autophagy, given that mitochondrial homeostasis, regulated by frequent mitochondrial turnover, requires autophagy. Nevertheless, genetically modified animal models, such as ATG5 or ATG7 knockout (KO) mice, which allow the modulation of autophagy, will be essential in identifying the precise mechanism.
Currently, there is no clear evidence providing the indispensability of muscle fiber-type shifts and morphological remodeling. The results of our study along with the current literature suggest that autophagy may selectively occur in oxidative muscle fibers (type I and IIa) and is necessary for exercise-induced muscle fiber-type transformation. Our correlation analysis results showed no significant correlation between muscle fiber-specificity and autophagy in the SOL but showed a tendency. Our findings did not clarify whether exercise-induced autophagy critically contributes to muscle fiber-type shifting; thus, further studies using an autophagy KO model are warranted.
Although an increase in LC3-II levels has been accepted as a primary autophagy marker in general, its modulation can be triggered by dysregulated autophagy flux (e.g., failure of autophagosome removal or outrun formation of autophagosomes over its removal rate), leading to misinterpretation of autophagy results. Therefore, the overall measurement of autophagy processes, including the initiation (e.g., BCL2 phosphorylation and BECN1) and lysosomal degradation steps (e.g., p62, LAMP, and CTSL), is necessary to confirm our results on exercise-induced autophagy in SOL muscles. A previous study showed that BCL2 phosphorylation at serine 87 (p-BLC2Ser87) is necessary for autophagy in hepatocarcinoma cells [47]; in contrast, exercise-trained SOL muscles exhibited reduced p-BLC2Ser87 levels along with increased LC3-II levels, indicating that exercise-induced autophagy is inversely related to p-BLC2Ser87. However, the importance of BCL2 phosphorylation in exercise-induced autophagy should not be disregarded because other BCL2 phosphorylation sites (e.g., Thr69, Ser70, and Ser84) have been linked to exercise-induced autophagy [27] but were not examined in the present study.
Reduced p62/SQSTM1 levels along with increased LC3-II levels have been considered critical markers of enhanced autophagy as internalized p62/SQSTM1 in autophagosomes degrades with other cargos in response to acute exhaustive endurance exercise in a mouse model [27]. Interestingly, contrary to this study, human acute endurance cycling exercise showed that p62/SQSTM1 levels remain unchanged, suggesting that LC3-II levels and p62/SQSTM1 may not always match [48]. Consistent with this study, our data showed that chronic endurance training did not alter p62/SQSTM1 levels or LAMP2. We postulate that the cause of the discrepant observations between the study by He et al. [27], and ours is that while acute exhaustive exercise may intensively activate autophagy to remove the potentially harmful debris in the cell, long-term endurance exercise may induce autophagy reconditioning processes in SOL muscles to properly regulate autophagy, which may downregulate autophagy to prevent an unnecessary catabolic state despite the same exercise stimuli.
Alterations in the metabolic signaling nexus, including AMPK, AKT, mTOR, and ULK1, in response to exercise training have been widely reported. For example, AMPK activation is necessary for 6 weeks of voluntary wheel running-induced muscle fiber-type shifting (type IIb to type IIa/x) in the triceps brachii muscles [49]. Contrary to this, we found that chronic treadmill running exercise-induced muscle fiber-type shifting (type IIa to type I) occurred with repressed AMPK activity (e.g., reduced p-AMPKThr172 levels) in the SOL muscles. Although the reason for this conflicting observation cannot be precisely explained here, we propose that the time of exercise-induced muscle fiber-type transformation may alter the metabolic adaptive responses. For example, once muscle fibers reach an optimal muscle type conversion from fast- to slow-twitch fibers and its termination is necessary to sustain homeostasis, the turndown of AMPK influence may occur. Supporting this notion, our subsequent results showing the concurrent upregulation of anabolic signaling activities (p-AKTSer473/AKT ratio and p-mTORSer2448/mTOR ratio) suggest that restoration of the metabolic state from catabolism to anabolism is indispensable. Furthermore, variations in the mode of exercise (e.g., voluntary wheel running vs. treadmill running) and the time of sacrifice (60 min vs. 3-4 hr. post-exercise) may lead to different results in AMPK activity.
Reinforced mTOR activity concomitant with increased CSA of skeletal muscles has been a commonly accepted concept in many studies, especially on fast-twitch fibers. Interestingly, unlike in the SOL, exercise-trained EDL sustained the same mTOR phosphorylation levels, despite increases in the CSA of type IIx and IIb fibers. Although the possible explanation for this observation is unclear, this discrepancy may be due to muscle-specific adaptive responses to repetitive exercise stimulation, with slow-twitch muscle fibers being more sensitive to mTOR activation than fast-twitch muscle fibers, or to time-dependent anabolic responses. For instance, since the mice were sacrificed immediately after 90 min post-exercise, we presume that mTOR activity may respond rapidly in the SOL muscle, but is delayed in the EDL to later time intervals. This concept warrants further investigation in prospective studies regarding the anabolic response timing post-exercise in different muscle types.
AMPK activation is a pivotal step in initiating autophagy induction and is, thus, a critical step in canonical autophagy. Previous studies have reported that AMPK-dependent p-ULK1Ser555 and mTOR-dependent p-ULK1Ser757 critically affect autophagy [50]. Two recent studies have shown that AMPK-induced p-ULK1Ser555 is necessary for autophagy induced by acute endurance cycling exercise in the human vastus lateralis muscle [48] and acute voluntary wheel running exercise in the flexor digitorum brevis muscle in mice [51]. Interestingly, in our study, endurance exercise promoted autophagy in the absence of AMPK phosphorylation. Indeed, AMPK activity remained significantly repressed, and mTOR did not influence ULK1 activity in chronically exercise-trained SOL muscles. That is, although the p-AMPKThr172/AMPK ratio decreased and the p-mTORSer2448/mTOR ratio increased, neither p-ULK1Ser555 nor p-ULK1Ser757 levels were affected in the SOL muscles, exhibiting an increase in autophagy (LC3-II). The discrepancies in autophagy signaling among the studies may be because the AMPK-ULK1 mutual interaction is fundamental for acute autophagy, but dispensable for muscles chronically exposed to endurance exercise. In addition, given our study design, in which muscle tissues were excised and autophagy was examined 90 min post-exercise to detect maximum autophagy activation, as demonstrated by He et al. [27], we cannot exclude the possibility that the AMPK-mTOR-ULK1 signaling nexus may be necessary for basal autophagy modulation.
Interestingly, although this single snapshot of signaling changes cannot demonstrate a true autophagy phenomenon in the EDL muscles, an increase in the p-ULK1Ser757/ULK1 ratio, independent of mTOR, appears to be associated with the failure of exercise-induced autophagy. Our study shows that 6 weeks of moderate- to high-intensity treadmill running exercise induces muscle fiber-type transformation from type IIa to type I fibers and autophagy in oxidative SOL muscles, but not in glycolytic EDL muscles. However, our answer to the question of whether autophagy is crucial for muscle fiber-type shifts is not conclusive because there is a trend for an association between muscle fiber-type transitions and LC3-II content in the SOL muscles. Nonetheless, our study provides prime evidence that chronic exercise-induced autophagy in slow muscles can coincide with a reinforced anabolic state, contrary to the nexus of canonical autophagy signaling mediated by nutrient deficiency. In addition, our results suggest that exercise-induced anabolic response intervals may vary between slow- and fast-twitch muscle fiber types.
Acknowledgements
This study was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2018S1A5B5A02037951), the KU Research Professor Program of Konkuk University, and the fellowship grant awarded to Youngil Lee by the Reubin O’D. Askew Institute for Multidisciplinary Studies. We would like to thank Editage (www.editage.co.kr) for English language editing.