Exercise improves hypobaric hypoxia-induced skeletal muscle dysfunction via Sirt1-Mediated myotube and mitochondrial remodeling

Graphical abstract

Highlights

• •Chronic hypoxia impairs muscle performance via Sirt1 downregulation.
• •Sirt1 reactivation restores muscle function via PGC-1α/FoxO3a signaling.
• •Combined preconditioning and hypoxic exercise alleviate high-altitude atrophy.
• •Mitochondrial ultrastructure restoration underlies exercise benefits.
• •Vascular-endothelial-satellite cell crosstalk enhances hypoxia adaptation.

Abstract

Introduction

Skeletal muscle function is profoundly challenged under high-altitude environments, where hypobaric hypoxia disrupts structural integrity and impairs physiological function. However, few animal studies have examined the impact of hypobaric hypoxia on skeletal muscle and molecular basis. While exercise training holds promise for alleviating hypoxia-induced muscle dysfunction, the understanding of its protective mechanisms remains limited.

Objectives

We aimed to investigate chronic hypobaric hypoxia-induced myotube atrophy and mitochondrial dysfunction in mouse models and C2C12 cells, and develop a combined exercise strategy (preconditioning and hypoxic training) to mitigate hypoxia-related muscle pathology.

Methods

A mouse chronic hypobaric hypoxia model (45-day exposure, 6,000 m equivalent) combined with in vitro C2C12 myotube hypoxia simulations was employed. Muscle atrophy, mitochondrial ultrastructure, and molecular pathways were analyzed via histology, proteomics, and functional assays. Exercise interventions included preconditioning (9-week treadmill training) followed by voluntary wheel running under hypobaric hypoxia.

Results

Chronic hypobaric hypoxia induced pronounced skeletal muscle dysfunction and mitochondrial structural disorganization. However, exercise preconditioning combined with hypoxic training attenuated these hypoxia-induced impairments. Both hypoxic skeletal muscles in vivo and C2C12 cells in vitro exhibited significant Sirt1 downregulation. Notably, overexpression of Sirt1 or treatment with exercise mimetics partially reversed hypoxia-induced myotube atrophy and mitochondrial dysfunction through the PGC-1α/FoxO3a signaling pathway-a mechanism shared with exercise interventions.

Conclusion

This study uncovers exercise as a potent inducer of hypoxia resilience through Sirt1-dependent mitochondrial repair and multicellular crosstalk (vascular-endothelial-satellite cell axis). Our “train-before-you-climb” approach could transform how we prepare for high-altitude living, offering a drug-free way to keep muscles strong where the air is thin.

Introduction

Skeletal muscle refers to muscle tissue that is attached to bones and is responsible for facilitating the movement of certain bones. Skeletal muscle makes up approximately 40–50 % of the total body mass in individuals with a healthy weight [1,2]. An adequate oxygen supply is crucial for effective muscle contraction, maintaining homeostasis, and regulating processes such as cell proliferation and differentiation. However, insufficient oxygen delivery can lead to skeletal muscle dysfunction [3]. While transient hypoxia may play regulatory roles in specific cellular processes, for instance, intermittent hypobaric hypoxia exposure enhances skeletal muscle recovery following injury, hypoxic adaptation improves systemic capacity for oxidative metabolism and oxygen delivery, features exploitable in exercise conditioning regimens [4,5]. However, chronic hypoxia typically induces maladaptive responses, characterized by functional impairment and pathological remodeling [5].

High-altitude environments are characterized by ambient hypoxia caused by decreasing atmospheric oxygen levels with increasing elevation [6]. Hypobaric hypoxia at high altitudes induces multifaceted physiological adaptations, among which reductions in body mass and protein reserves are the most consistently observed alterations [7]. As the primary reservoir for bodily proteins, skeletal muscle accounts for 50–75 % of the total body protein stores, making this tissue particularly vulnerable to hypoxia-mediated catabolic processes [8]. High-altitude exposure is a detrimental factor for the loss of muscle mass and functional decline, often leading to muscle atrophy [9]. For instance, Indian soldiers stationed at 3,750–4,100 m for 10 months exhibited a 1.38–1.74 kg reduction in lean mass, accompanied by fluid retention and reduced fat mass, with X-ray imaging confirming diminished myofiber cross-sectional areas [10]. Similarly, studies on Mount Everest climbers revealed a 4 % decrease in muscle cross-sectional area, although mitochondrial function and exercise-induced metabolite profiles remained largely unchanged [11]. Muscle biopsies further demonstrated a 30 % decline in mitochondrial density under chronic hypoxia [12]. While the ubiquitin–proteasome pathway and calpain activation have been implicated in hypobaric hypoxia-induced atrophy, the upstream signaling mechanisms remain poorly defined [13]. Existing research predominantly attributes altitude-related muscle dysfunction to redox imbalances [9]. Although the regulatory role of hypoxia-inducible factor (HIF) in satellite cell activity and myogenesis has been well documented, the molecular mechanisms linking high-altitude hypobaric hypoxia to mitochondrial dysfunction and proteostatic imbalance remain poorly understood [5,14].

Mitochondria are essential for maintaining skeletal muscle energy balance and dynamically regulating biogenesis and functional adaptation to exercise and pathological conditions [15]. Sirtuins (SIRTs) are recognized as key regulators of skeletal muscle metabolism and participate in critical processes, such as fiber-type switching, endoplasmic reticulum stress response, and atrophy [16]. Among SIRTs, Sirt1 (NAD⁺-dependent deacetylase sirtuin 1) acts as a central metabolic regulator in muscle, activating downstream signaling via peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) to ameliorate mitochondrial dysfunction and attenuate muscle atrophy [17,18]. For instance, resveratrol enhances skeletal muscle mitochondrial biogenesis by activating Sirt1 [19], whereas myricetin alleviates dexamethasone-induced atrophy through Sirt1 upregulation [20]. In addition, under hyperglycemic or oxidative stress conditions, Sirt1 stability is compromised by ubiquitination-mediated degradation [21,22]. However, it remains unclear whether hypobaric hypoxia impairs Sirt1 stability, potentially disabling this crucial metabolic sensor, subsequently exacerbates mitochondrial reactive oxygen species (ROS) accumulation and myofiber atrophy. Meanwhile, strategies to preserve Sirt1 activity under conditions of chronic hypobaric hypoxia are still underexplored.

Hypoxia-induced muscle atrophy involves multisystem crosstalk (neuroendocrine and immune), necessitating interventions that target the core regulators. Exercise-induced skeletal muscle remodeling provides various systemic health benefits in both mice and humans [23]. Anabolic stimuli such as exercise or hormones promote muscle growth via protein synthesis and organelle biogenesis [2]. During exercise and ambient heat stress, AMPK phosphorylation and Sirt1 expression are enhanced [24]. Sirt1 can deacetylate key proteins that are important for cellular adaptation to exercise training [25]. Exercise can reverse muscle atrophy by regulating Sirt1 through various mechanisms, including the activation of the AMPK/Sirt1/PGC1α signaling pathway [26,27]. Conversely, exercise modulates FOXO3 to suppress atrophy-related ubiquitin ligases [28]. The molecular adaptations of skeletal muscle to exercise training primarily involve mitochondrial biogenesis, angiogenesis, and myofiber composition [29,30]. However, it remains unclear whether exercise preconditioning equips skeletal muscle to meet hypoxic challenges by enhancing energy provision, meanwhile, can exercise-induced vascular expansion and satellite cell recruitment counteract microvascular rarefaction and regenerative impairment induced by chronic hypobaric hypoxia.

In this study, we delineated that chronic hypobaric hypoxia promotes skeletal muscle dysfunction and mitochondrial impairment through suppression of Sirt1. Furthermore, exercise training partly restores muscle function via dual mechanisms: (1) stabilization of Sirt1 to coordinate mitochondrial adaptation and anti-atrophy signaling, and (2) exercise-primed remodeling of the microvascular and activation of satellite cells (SCs). Our findings establish exercise as a nonpharmacological strategy to help prevent muscle problems caused by high-altitude living.

Materials and methods

Detailed information is provided in Supplementary Data.

Experimental animals

Animals were maintained under a 12-hour light/dark cycle with free access to standard rodent feed and water. All experiments were performed with age-matched male C57BL/6J mice. Mice in the hypobaric hypoxia group were housed in a plateau simulated chamber (provided by China Guizhou Fenglei Aviation Ordnance Co., Ltd.) for 45 days. The chamber's altitude was set to 6000 m, with an oxygen concentration of 9 %.

Exercise protocol

A modified exercise protocol was implemented based on established methods [31,32]. For pre-exercise, mice underwent a three-day acclimatization phase on a 10° inclined treadmill. Initial sessions included 5-minute runs at 8 m/min (Day 1), followed by incremental increases to 10 m/min (Day 2). On Day 3, animals performed an acute 40-minute session at 12 m/min, then increased by 1 m/min every 10 min until 17 m/min. Subsequently, a nine-week regimen was applied: five consecutive days of daily 1-hour runs at 17 m/min (10° incline), interspersed with two rest days weekly. Voluntary wheel running was used as a model for exercise-induced angiogenesis [33]. Thus, voluntary wheel running was employed during 45-day hypobaric hypoxia exposure. Maximal running capacity was assessed post-acute exercise (90-minute duration) using a graded speed protocol: treadmill velocity increased by 1 m/min every 5 min until exhaustion.

Muscle grip strength and motor coordination assessment

Grip strengths were assessed using a Grip Strength Meter (Zhongshi Technology, Beijing, China). Mice gripped a metal grid with all paws or forepaws only and were gently pulled backward by the tail until release. Normalized grip strength (three trials averaged and divided by body weight) was recorded per animal. Motor coordination was evaluated via rotarod testing (Zhongshi Technology) following a two-day acclimatization period.

Muscle tetanic force measurements

Under anesthesia, the tibialis anterior (TA) muscles were surgically exposed. Nonabsorbable silk sutures secured to tendons were connected to force transducers. Stimulating electrodes were positioned at resting tension, and optimal silk length/current were calibrated for maximal force output. Tetanic contractions were induced using a bio-signal acquisition system (Zhongshi Technology) under standardized parameters: 800 HZ sampling rate; positive voltage strength incremental stimulation; initial voltage 0.2 V; increment 0.1 V.

Cell treatment

C2C12 myoblasts (American Type Culture Collection, CRL-1772) were cultured in DMEM supplemented with 10 % FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL). Differentiation into myotubes was induced at confluence using DMEM containing 2 % horse serum, with medium replenished every 48 h. Experiments utilized four-day post-differentiated myotubes. For Sirt1 overexpression, cells were transfected with 2 μg/mL plasmids (empty vector, pLV3-CMV-Sirt1, or UB-K48) via Lipofectamine 2000 (Invitrogen). At 24 h post-transfection, cultures underwent medium replacement and hypoxia exposure 72 h. Exercise-mimetic conditions were simulated using 0.25 mM AICAR (HY-13417, MCE) treatment for 12 h prior to hypoxia.

Statistical analysis

All experiments had at least three repetitions. Figure legends specify the number of replicates (n) for each animal experiment. Data analysis used GraphPad Prism (version 9). Results are presented as mean ± SEM per experimental group. Unpaired two-tailed t-tests were used to compare two groups, while one-way ANOVA was used to compare more than two groups, followed by Dunnett's multiple comparisons post hoc test. Two-way ANOVA followed by Dunnett's or Bonferroni's multiple comparisons post hoc test was used for comparisons involving experimental factors with more than two levels. Statistical significance is indicated by *p < 0.05, ^**p < 0.01, and ^***p < 0.001.

Results

Hypobaric hypoxia-induces skeletal muscle dysfunction: Characterized by decreased muscle mass and impaired physical performance

To model chronic physiological adaptations of populations subjected to prolonged high-altitude environments, we employed a chronic hypobaric hypoxia protocol in this study. Mice were exposed to hypobaric hypoxia for 45 days at an altitude of 6000 m, during which the ambient partial pressure of oxygen (pO₂) was reduced to 9 % (Fig. 1A). Following this exposure, the mice exhibited reduced limb strength (Fig. 1B), which negatively affected exercise endurance (Fig. 1C, Fig. S1A) and motor coordination (Fig. 1D). In the TA muscle, there were reductions in isometric twitch and tetanic forces (Fig. 1E). Additionally, we found a reduction in limb muscle weight in the hypobaric hypoxia group (Fig. 1F, Fig. S1B). Examination of myofiber cross-sectional area (CSA) by histochemical staining revealed a marked decrease in myofiber size compared with the normoxia control group (Fig. 1G). The gastrocnemius (GAS) muscle, known for its mixed fiber composition, exhibited a significant decrease in myofiber cross-sectional area (Fig. S1C). Additionally, Masson's trichrome staining of TA muscle revealed fibrosis development under hypoxic conditions (Fig. S1D), accompanied by a substantial increase in fibrotic markers (Fig. S1E).

Fig. 1.

Hypobaric hypoxia-induces skeletal muscle dysfunction: characterized by decreased muscle mass and impaired physical performance. (A) The schematic of the experimental setup for the 45-day chronic hypobaric hypoxia murine model. (B) Following 45-day hypobaric hypoxia exposure, mice exhibited reduced four-limb grip strength compared to normoxia controls. n = 5–7/group. (C) Bars represent mean running time on a motorized treadmill. n = 8/group. (D) Bars represent mean rotarod test time on a rotarod latency. n = 10/group. (E) Measurements of muscle tetanic contraction show reduced force in the hypobaric hypoxia group. n = 13–15/group. (F) Representative TA muscles from indicated mice. TA muscle weight was normalized to body mass after hypobaric hypoxia exposure. n = 7/group. (G) Representative haematoxylin and eosin (H&E) staining of TA muscle sections and quantification of CSA. Scale bar: 50 μm. n = 3/group. (H) Representative images of MyHC immunofluorescence of C2C12 cells under control and hypoxia. Measurements of myotube diameter. Scale bar: 50 μm. n = 100 myotubes per condition. (I) Immunoblot analysis and quantification of protein expression in hypoxic C2C12 cells using the indicated antibodies. n = 6/group. *p < 0.05; **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels.

To investigate the effect of hypoxia on skeletal muscle function, C2C12 myotubes were cultured under 5 % oxygen for 72 h [34]. Hypoxic exposure led to significant myotube atrophy, characterized by a marked reduction in myotube diameter (Fig. 1H), downregulation of the myogenic regulatory factor Myogenin, and increased expression of the skeletal muscle negative regulator Myostatin (Fig. 1I). Collectively, hypobaric hypoxia-induced skeletal muscle dysfunction characterized by decreased muscle mass and impaired physical performance.

Chronic hypobaric hypoxia exposure impairs the mitochondrial function of skeletal muscle

Mitochondria in skeletal muscle are vital for muscle function and integrity. We performed proteomic analysis of TA muscles under normoxia and hypobaric hypoxia. Gene Ontology (GO) analysis, using |Log₂FC|>1 and adjusted p-value < 0.05 as the cut-off criteria, revealed that the observed changes were primarily localized to the mitochondrial inner membrane (Fig. 2A). To identify genes associated with oxidative phosphorylation (OXPHOS) complexes in the mitochondria, we presented the key change genes on a heatmap (Fig. 2B). Additionally, changes in protein expression were validated using western blot analysis, CIV-Mtco1 and CII-Sdhb exhibited the most significant downregulation in the hypobaric hypoxia group (Fig. 2C). A partial decrease in mitochondrial-associated proteins was also observed in the GAS muscle (Fig. S2A). Concurrently, the isolation of intact mitochondria from muscles similarly demonstrated that hypoxia affected mitochondrial fission and fusion in muscle tissue (Fig. S2B). The mitochondrial DNA content in the muscles were also significantly reduced (Fig. 2D, Fig. S2C). Moreover, hypoxia induced a decrease in mitochondrial membrane potential and downregulation of OXPHOS complexes expression in C2C12, indicative of mitochondrial dysfunction (Fig. 2E and F). These results clearly show that chronic hypobaric hypoxia exposure impairs the mitochondrial function of skeletal muscle.

Fig. 2.

Chronic hypobaric hypoxia exposure impairs the mitochondrial function of skeletal muscle. (A) GO analysis of TA muscle proteomes from normoxia and hypobaric hypoxia groups, profiled by LC-MS/MS. n = 6/group. (B) Heatmap analysis of mitochondrial-related proteins identified by proteomics. (C) Immunoblot analysis and quantification of OXPHOS complexes in TA muscle lysates using the Total OXPHOS Rodent Antibody Cocktail. n = 6/group. (D) Mitochondrial DNA content in TA muscle assessed by qPCR (primers: mitochondria-encoded mtDNA vs. Nuclear-encoded Rbm15). n = 3/group. (E) Mitochondrial membrane potential in C2C12 cells under control and hypoxia using MitoTracker® Deep Red FM. Quantification of the average optical density of MitoTracker. Scale bar: 50 μm/100 μm. (F) Immunoblot analysis and quantification of OXPHOS complexes in C2C12 cells. n = 4/group. **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

E3 ubiquitin ligase facilitates degradation of the NAD⁺-dependent deacetylase Sirt1 under hypoxic conditions

The GO enrichment and pathway analysis revealed that the proteasome-mediated ubiquitin mediated proteolysis was enriched in functional categories (Fig. 3A). Under hypoxic conditions, we identified Trim63 (Murf1) upregulation as a critical driver of muscle atrophy, alongside other atrogenic ubiquitin ligases like Mafbx, Fbxo31, and Musa1 (Fig. 3B and C, Fig. S3A and B). Additionally, we detected elevated expression of Myostatin in the muscles (Fig. 3B, Fig. S3A). Consistently, the hypoxia-stimulated C2C12 upregulated atrogenes (Fig. 3D and E).

Fig. 3.

E3 ubiquitin ligase facilitates degradation of the NAD⁺-dependent deacetylase Sirt1 under hypoxic conditions. (A) The GO enrichment and pathway analysis of TA muscle proteomes. (B) Immunoblot analysis and quantification of muscle-inhibitory factor expression in TA muscle. n = 6/group. (C) The atrogenes expression level in TA muscle assessed by qPCR. n = 3/group. (D) Immunoblot analysis and quantification of Trim63 in C2C12 cells. n = 6/group. (E) The atrogenes expression level in C2C12 cells assessed by qPCR. n = 3/group. (F) Immunoblot analysis and quantification of Sirt1 expression in TA muscle and C2C12 cells. n = 6/group. (G) Immunoprecipitation analysis of Sirt1 ubiquitination upon co-transfection with Ub-K48 plasmids in C2C12 cells. After 48 h hypoxia exposure, cells were treated with 10 μM MG132 for 24 h. (H) Immunoblot analysis of the expression of Nedd4 and Mdm2 in the whole cell lysates (WCL) of C2C12 cells. (I) Immunoprecipitation analysis of the interaction between E3 ubiquitin ligases and Sirt1 after hypoxic stimulation. **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels.

In hypoxia-induced muscles and C2C12, the expression levels of the Sirt1 protein were significantly reduced (Fig. 3F, Fig. S3A). To investigate whether Sirt1 degradation may be mediated by the ubiquitin–proteasome system (UPS) under normoxic or hypoxic conditions, we transfected Ub K48-encoding plasmids into C2C12 and conducted an immunoprecipitation assay with Sirt1 in the presence of the proteasome inhibitor MG132. A ladder of ubiquitinated Sirt1 products were detected in hypoxic cells, confirming that Sirt1 was targeted for degradation under hypoxic conditions (Fig. 3G). Sirt1 depletion may occur through both transcriptional downregulation (Fig. S3C) and prominent ubiquitin-proteasomal regulation, as evidenced by reduced gene expression and UPS-targeting assays. To investigate the negative regulatory mechanism of Sirt1, we utilized UbiBrowser software (https://ubibrowser.ncpsb.org/ubibrowser/) to predict the potential E3 ubiquitin ligases of Sirt1, identifying several candidates (Fig. S3D). There was no significant increase in the expression of ubiquitin ligases Nedd4 and Mdm2 in the whole cell lysates of C2C12 cells under hypoxic conditions (Fig. 3H). Immunoprecipitation analysis revealed that Mdm2 enhanced interaction with Sirt1 following hypoxic stimulation (Fig. 3I). Our findings suggest that the ubiquitin–proteasome system may contribute to hypoxia-induced Sirt1 downregulation.

Overexpression of Sirt1 attenuates hypoxia-induced myotube atrophy and mitochondrial dysfunction

The above results show that hypoxia decreases Sirt1 protein levels. Therefore, we used an overexpression plasmid (OE-Sirt1) to reverse this process (Fig. 4A, Fig. S4A). Overexpression of Sirt1 improved hypoxia-induced myotube wasting to some extent, with a partial improvement in myotube diameter (Fig. 4B, Fig. S4B). Next, we explored the role of Sirt1 in myotube atrophy. We analyzed the expression levels of Trim63 and other atrogenes in myotubes overexpressing Sirt1. Our results indicated that these atrogenes were upregulated at both the transcriptional and protein levels in Sirt1-overexpressing myotubes, accompanied by a significant improvement in Myogenin (Fig. 4C, Fig. S4C). Giemsa staining confirmed that Sirt1 overexpression in C2C12 undifferentiated myoblasts improved hypoxia-induced defects in fusion for myotube formation (Fig. S4D).

Fig. 4.

Overexpression of Sirt1 attenuates hypoxia-induced myotube atrophy and mitochondrial dysfunction. (A) Immunoblot analysis and quantification of Sirt1 in C2C12 cells transfected with pLV3-CMV-Sirt1 (OE-Sirt1) or empty vector. Myotubes were cultured under control or hypoxia conditions for 72h. n = 5/group. (B) Representative images of MyHC immunofluorescence in C2C12 cells infected with pLV3-CMV-Sirt1 or empty vector under control or hypoxia conditions. Scale bar: 100 μm. Measurements of myotube diameter. n = 100 myotubes per condition. (C) Immunoblot analysis and quantification of Trim63 and Myogenin in C2C12 cells transfected with the indicated constructs. n = 5–6/group. (D) Representative images of mitochondrial membrane potential in myotubes transfected with the indicated constructs using JC-10. Scale bar: 100 μm. Quantification of the ratio of JC-10 aggregates to monomers, i.e., the ratio of red to green fluorescence. n = 3/group. (E) Representative images and quantification of ROS levels in myotubes transfected with the indicated constructs using the DHE probe. Scale bar: 100 μm. n = 3/group. (F) Mitochondrial DNA content and ATP production in C2C12 cells transfected with the indicated constructs. n = 3/group. (G) Graphical representation of oxygen consumption rates (OCR) and ATP production in treated cells is shown. n = 3/group. (H) Immunoblot analysis and quantification of mitochondrial-related protein in C2C12 cells transfected with the indicated constructs. n = 4/group. (I) Immunoblot analysis and quantification of PGC-1α expression in TA muscle and C2C12 cells. n = 5–6/group. (J) Immunoblot analysis and quantification of PGC-1α in C2C12 cells transfected with the indicated constructs. n = 5/group. (K) Immunoprecipitation using PGC-1α antibody and immunoblot analysis of the indicated proteins. *p < 0.05; **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. One-way ANOVA was used for comparisons involving more than two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Sirt1 is involved in regulating several biological processes in mitochondria. Given its role in maintaining cellular energy − metabolic capacity, we studied the molecular mechanisms of Sirt1-mediated mitochondrial quality control under hypoxia. Sirt1 overexpression reduced the hypoxia-induced increase in JC-10 monomer ratio (Fig. 4D). To measure reactive oxygen species (ROS) levels in vitro, we used the DHE probe in differentiating C2C12 cells. As shown in Fig. 4E, ROS levels were elevated after hypoxia, but Sirt1 overexpression significantly lowered ROS levels. Consistent with the reduced mitochondrial membrane potential and ROS levels, the decrease in mitochondrial DNA content and ATP production were partially restored by Sirt1 overexpression (Fig. 4F). Based on these findings, we performed Seahorse XF analysis to evaluate mitochondrial function. Sirt1 overexpression partially attenuated hypoxia-induced impaired oxidative phosphorylation, as evidenced by an increase in mitochondrial ATP production rate (Fig. 4G). In addition, modulating the expression of Sirt1 partially increased the protein levels of mitochondrial-associated proteins, including CIV-Mtco1, CI-Ndufb8, and CII-Sdhb in the OXPHOS complexes, thereby alleviating the hypoxia-induced reduction in protein expression (Fig. 4H). We found that Sirt1 overexpression restores the mRNA levels of key mitochondrial dynamics factors (Drp1, Opa1, Mfn2) under hypoxic conditions (Fig. S4E).

Sirt1, a member of the sirtuin family of histone deacetylases, modulates skeletal muscle metabolism by deacetylating proteins such as PGC-1α. Hypoxia also downregulated the expression of PGC-1α, a key regulator of mitochondrial biogenesis (Fig. 4I). It is known that reduced Sirt1 activity leads to an increase in the acetylation level of PGC-1α, which in turn results in decreased expression of mitochondrial respiratory chain genes and dysfunction of mitochondrial oxidative phosphorylation [35,36]. Notably, Sirt1 overexpression restores PGC-1α expression under hypoxic conditions (Fig. 4J). Immunoprecipitation results demonstrated that Sirt1 overexpression suppressed the hypoxia-induced increase in acetylated PGC-1α levels (Fig. 4K).

In addition, we found that hypoxia increases FoxO3a expression in C2C12 (Fig. S4F). As FoxO3a is involved in the transcriptional regulation of atrophy-related factors and Sirt1 directly deacetylates it, we carried out an immunoprecipitation assay for Sirt1. As shown in Fig. S4G, oxygen deprivation increased the acetylation level of FoxO3a, whereas Sirt1 overexpression promoted its deacetylation.

Pre-exercise alleviates skeletal muscle dysfunction upon hypobaric hypoxia exposure

It is well known that exercise promotes morphological and functional adaptations in muscles after contraction. To study whether exercise primes skeletal muscle adaptation to hypobaric hypoxia, we used pre-exercise combined with voluntary wheel running (EXE-Hypobaric Hypoxia-VWR) as a model (Fig. 5A). Following a 9-week pre-training period and 45-day voluntary wheel running under high-altitude hypoxic conditions, mice subjected to pre-training and hypoxic exercise demonstrated significant improvement in grip strength (Fig. 5B). Running endurance tests revealed that the hypoxia-exercised group exhibited markedly prolonged running durations compared to the hypobaric hypoxia-sedentary group (Fig. 5C). However, no significant differences were observed in the latency to fall in the suspension test between these groups (Fig. S5A). Rotarod performance analysis indicated that exercised mice displayed enhanced capacity to dynamically adjust skeletal muscle contraction patterns, thereby maintaining postural stability under accelerating rotational challenges (Fig. 5D). Muscle force measurements further demonstrated that the threshold voltage required to elicit tetanic forces in the hypobaric hypoxia-exercised group was reduced by over 50 % relative to the hypobaric hypoxia-sedentary group (Fig. 5E). Notably, within the stimulation range of 0.2–5 V, muscles from the hypobaric hypoxia-exercised group generated significantly greater contractile forces compared to those from sedentary controls (Fig. 5F, Fig. S5B). Thus, we propose that exercise ameliorates skeletal muscle adaptation to hypobaric hypoxia.

Fig. 5.

Pre-exercise alleviates skeletal muscle dysfunction upon hypobaric hypoxia exposure. (A) The schematic diagram of exercise interventions, including 9 weeks of treadmill preconditioning followed by voluntary wheel running under hypobaric hypoxia. (B) Mice in the Exercise-Hypobaric Hypoxia-Voluntary Wheel group (EXE-Hypobaric Hypoxia-VWR) exhibited increased muscle grip strength. n = 13–14/group. (C) Exercise tolerance was assessed based on running time in a treadmill test. n = 7–9/group. (D) Motor coordination was evaluated based on rotarod time and distance in a rotarod latency test. n = 8/group. (E) The voltage required to elicit the initial twitch of electrical-triggered force. Parameters were as follows: 800 HZ sampling rate; positive voltage strength incremental stimulation; initial voltage 0.2 V; increment 0.1 V. n = 6-9/group. (F) Maximum muscle contraction force within the stimulation range of 0.2–5 V. The unit is g. n = 6–8/group. (G) Representative TA muscles from sedentary and exercised mice under normoxia or hypobaric hypoxia conditions (BW: 27.5 g; 27.4 g; 28.3 g). TA muscle weight was normalized to body mass. n = 9–12/group. (H) Representative H&E staining of TA muscle sections. Scale bar: 50 μm. Quantification of myofiber CSA in TA muscles. n = 100 myofibers per condition. (I) Representative images of immunostaining for eMyhc (red), laminin (green), and DAPI (blue) in TA muscle cross sections. Scale bars: 50 μm. Quantification of the percentage of eMyhC-positive myofibers within laminin staining. n = 3/group. (J) Immunoblot analysis and quantification of Trim63, Mafbx and Myostatin in TA muscles. n = 6/group. (K) Expression levels of atrogenes Musa1, Smart and Mafbx in muscles determined by qPCR. n = 3/group. *p < 0.05; **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. One-way ANOVA was used for comparisons involving more than two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels. SED indicates sedentary, EXE indicates exercise and VWR indicates voluntary wheel running. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Histological and morphometric analyses of harvested muscle tissues revealed that the hypobaric hypoxia-exercised group exhibited partial improvements in muscle mass and myofiber CSA, with more pronounced restoration of muscle loss in the TA and GAS compared to sedentary controls (Fig. 5G and H, Fig. S5C). Furthermore, while no significant differences in absolute muscle weight were detected, the soleus muscle (SOL), predominantly composed of slow-twitch fibers, demonstrated a marked increase in fiber CSA following hypoxic exercise, suggesting fiber type-specific structural remodeling (Fig. S5D). Immunofluorescence analysis of SOL revealed elevated expression of the slow myosin heavy chain isoform (MHC-1) (Fig. S5E). These findings suggest that the SOL adapts to hypobaric hypoxia by upregulating MHC-1, a mechanism that may balance hypoxia-induced oxidative stress with metabolic energy demands. Next, we investigated the adaptive alterations in TA muscles following exercise under hypobaric hypoxia. Notably, these adaptations include fiber-type transitions and dynamic regulation of embryonic myosin heavy chain (eMyhc) expression. Fluorescence staining of two fiber types in the TA revealed a progressive shift from MHC 2b-dominant to MHC 2a-enriched fiber composition in the exercised group (Fig. S5F). Skeletal muscle exhibits remarkable regenerative capacity in response to injury or external stimuli, and hypoxic exercise further amplified the formation of newly regenerated myofibers (Fig. 5I). Furthermore, we analyzed negative regulators of muscle function in the TA at both protein and transcriptional levels. Remarkably, the hypobaric hypoxia-exercised group exhibited partially reduced expression of atrophy-inducing factors, including Trim63, Mafbx, Musa1 and Smart, along with decreased secretion of Myostatin, compared to the hypobaric hypoxia-sedentary group (Fig. 5J and K). Taken together, these results confirm pre-exercise alleviates skeletal muscle dysfunction upon hypobaric hypoxia exposure.

Exercise activates Sirt1 signaling and facilitates mitochondrial remodeling under hypoxic stress

Exercise training serves as a critical modulator of skeletal muscle function by activating AMPK through mitochondrial remodelling [37]. Building on this, we hypothesized that exercise-induced adaptive responses in hypoxic skeletal muscle may be associated with mitochondrial restructuring. Electron microscopy (EM) analysis of longitudinal TA sections revealed ultrastructural alterations in mitochondrial architecture. We observed that subsarcolemmal (SS) compartments in hypoxic muscles exhibited disorganized and fragmented cristae with membrane rupture. In contrast, while the exercised group showed reduced matrix electron density, their cristae maintained intact morphology (Fig. 6A left). Similarly, intermyofibrillar (IMF) compartments in the hypoxia-sedentary group displayed marked dilation, vesiculation, and extensive void areas, whereas these pathological features were attenuated in the exercised group (Fig. 6A right). Concurrently, marked improvements in mitochondrial DNA content and ATP production were observed in exercised muscles under hypobaric hypoxia (Fig. 6B and C, Fig. S6A). We next assessed Sirt1 levels in skeletal muscle from sedentary and exercised mice under hypobaric hypoxia. Notably, the hypoxia-exercised group exhibited elevated Sirt1 expression compared to sedentary controls (Fig. 6D). These results were further validated through immunofluorescence analysis of TA muscle sections (Fig. 6E). Sustained exercise training preserves muscle functional integrity under hypoxic stress through satellite cell-mediated repair and regeneration [38]. Notably, the hypoxia-exercised group exhibited a pronounced increase in activated Pax7⁺ satellite cells (Fig. 6F, Fig. S6B).

Fig. 6.

Exercise alleviates skeletal muscle dysfunction induced by hypobaric hypoxia through mitochondrial remodeling and the activation of Sirt1 signaling. (A) Representative electron micrographs of TA muscle showing subsarcolemmal (SS) and intermyofibrillar (IMF) mitochondria in a longitudinal section from sedentary and exercised mice under normoxia or hypobaric hypoxia conditions. (B) ATP production in TA muscles. n = 3/group. (C) Mitochondrial DNA content in TA muscles from the indicated mice assessed by qPCR. n = 3/group. (D) Immunoblot analysis and quantification of Sirt1 in TA muscles from sedentary and exercised mice under normoxia or hypobaric hypoxia conditions. n = 6/group. (E) Representative images of immunostaining for Sirt1 (red), laminin (green), and DAPI (blue) in TA muscle cross sections. Scale bars: 50 μm. Quantification of the percentage of Sirt1-positive myofibers within laminin staining. n = 3/group. (F) Representative images of immunostaining for Pax7 (red), Laminin (green), and DAPI (blue) in muscle cross sections. Scale bars: 20 μm. Quantification of Pax7-positive SCs per section. n = 3/group. (G) Representative images of immunostaining for CD31 (red), laminin (green), and DAPI (blue) in TA muscle cross sections. Scale bar: 20 μm. Capillary density was measured as the total number of capillaries per total number of muscle fibers. n = 3/group. *p < 0.05; **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. One-way ANOVA was used for comparisons involving more than two groups. SED indicates sedentary, EXE indicates exercise and VWR indicates voluntary wheel running. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

As enhanced muscle vascularization boosts adaptive resources [39], we explored the association between angiogenesis and the functional resilience of exercised muscles under hypoxic stress. Immunofluorescence staining for endothelial markers CD31/PECAM-1 revealed a partially elevated capillary density in the hypoxia-exercised group, forming an intricate vascular network (Fig. 6G, Fig. S6C). Additionally, we investigated potential alterations in endothelial cell proliferation. The dual localization of CD31 and proliferation markers Ki67 indicated that, beyond the increase in vascular density, the number of endothelial cells might also be altered (Fig. S6D). These results support the fact that exercise facilitates mitochondrial remodeling and activates Sirt1 signaling. It also enhances vascularization and stimulates satellite cells. These combined effects may collectively ameliorate skeletal muscle dysfunction induced by hypobaric hypoxia.

Exercise-mimetic stimulation activates myotube function improvement via Sirt1- PGC-1α/FoxO3a in C2C12 cells

To verify the impact of exercise on myotube function in vitro, we used the AMPK activator AICAR, a master regulator of energy metabolism in skeletal muscle, to mimic exercise conditions in cell culture systems. We found that exercise-mimicking conditions activated the phosphorylation of AMPKα, which reversed the downregulation of Sirt1 expression caused by hypoxia (Fig. 7A). Pretreatment of differentiated myotubes with AICAR partially reversed the hypoxia-induced atrophy of myotubes, with a noticeable improvement in myotube diameter (Fig. 7B). At the protein level, this treatment partially counteracted the upregulation of the atrophy factor Trim63 (Fig. 7C). AICAR also partially reversed the hypoxia-induced increase in atrogenes at the transcriptional level (Fig. S7A).

Fig. 7.

Exercise-mimetic stimulation activates myotube function improvement via Sirt1-PGC-1α/FoxO3a in C2C12 cells. (A) C2C12 myotubes were treated with the AMPK activator AICAR (0.25 mM) for 12 h, then exposed to hypoxia for 72 h. Immunoblot analysis and quantification of Sirt1 and p-AMPK in C2C12 cells were performed. n = 6/group. (B) Representative images of MyHC immunofluorescence in C2C12 cells treated with AICAR under control or hypoxia conditions and measurements of myotube diameter. Scale bar: 100 μm. n = 100 myotubes per condition. (C) Immunoblot analysis and quantification of Trim63 in C2C12 cells treated with AICAR under control or hypoxia conditions. n = 6/group. (D) Representative images of mitochondrial membrane potential in myotubes treated with AICAR under control or hypoxia using JC-10. Scale bar: 100 μm. Quantification of the ratio of JC-10 aggregates to monomers. n = 3/group. (E) Representative images of ROS levels and quantification of the average optical density of DCFH-DA in myotubes treated with AICAR. Scale bar: 100 μm. n = 3/group. (F) Mitochondrial DNA content (n = 6/group) and ATP production in C2C12 cells treated with AICAR(n = 3/group). (G) The redox status of myotubes treated with AICAR was characterized by the ratio of NAD⁺/NADH. n = 3/group. (H) Graphical representation of OCR and ATP production in treated cells is shown. n = 3/group. (I) Immunoblot analysis and quantification of PGC-1α in C2C12 cells treated with AICAR. n = 6/group. (J) Immunoprecipitation using PGC-1α antibody and immunoblot analysis of the indicated proteins were performed. (K) Immunoblot analysis and quantification of FoxO3a in C2C12 cells treated with AICAR. n = 6/group. (L) Immunoprecipitation using FoxO3a antibody and immunoblot analysis of the indicated proteins were performed. *p < 0.05; **p < 0.01; ^***p < 0.001. Unpaired two-tailed t-test compared two groups. One-way ANOVA was used for comparisons involving more than two groups. Two-way ANOVA was employed for comparisons involving experimental factors with more than two levels.

The drug activates AMPK, mimicking exercise and driving mitochondrial oxidative metabolism in skeletal muscle. We explored the effects of AICAR on skeletal muscle mitochondria under hypoxic conditions. Using the JC-10 fluorescent probe, we observed a significant improvement in the mitochondrial membrane potential after exercise mimicking, along with a marked reduction in monomer levels induced by hypoxia (Fig. 7D). Moreover, the hypoxia-induced increase in ROS levels was substantially reversed (Fig. 7E). Under hypoxic conditions, the AMPK agonist specifically regulated mitochondrial biogenesis, increased mitochondrial DNA content, and enhanced ATP production in the myotubes (Fig. 7F). Our results showed that AICAR maintained the redox status of hypoxic C2C12 cells at a relatively stable level, ensuring normal cellular function (Fig. 7G). As expected, assessment of cellular oxygen consumption rate demonstrated that AICAR enhanced oxidative phosphorylation in C2C12 cells compared to the hypoxic group (Fig. 7H). Hypoxia forces muscles to shift from oxidative phosphorylation to glycolysis, leading to the accumulation of lactate that exacerbates intracellular acidosis and further inhibits muscle contraction function. The inclusion of AICAR reduced the accumulation of lactate in the cell supernatant under hypoxic conditions (Fig. S7B and C), further suggesting that AICAR pretreatment can improve the metabolic alterations in myotubes. Notably, we observed upregulation of PGC-1α expression in the hypoxic AICAR group (Fig. 7I), suggesting that AICAR, upon activation of AMPK, may influence its function by regulating post-translational modifications of PGC-1α. Immunoprecipitation assays with PGC-1α showed that AICAR partially reversed the hypoxia-induced increase in acetylation levels (Fig. 7J), maintaining its deacetylated active form, thereby ensuring its regulatory role in mitochondrial function. Additionally, we demonstrated that AICAR treatment effectively reduced the hypoxia-induced upregulation of FoxO3a in myotubes (Fig. 7K). Through immunoprecipitation assays with FoxO3a, we found that AICAR decreased the acetylation levels of FoxO3a, which was attributed to the activation of Sirt1 by AMPK (Fig. 7L). Collectively, exercise-mimetic stimulation activates myotube function improvement partly via Sirt1-PGC-1α/FoxO3a in C2C12 cells.

Discussion

Current research on high-altitude hypoxia largely remains descriptive; this study focuses on elucidating the physiological mechanisms underpinning skeletal muscle adaptation to enable targeted interventions. Our study reveals that chronic hypobaric hypoxia promotes skeletal muscle dysfunction via mitochondrial impairment and Sirt1 downregulation. However, exercise training partially counteracts these adverse outcomes through coordinated multi-system adaptations.

Consistent with prior studies in high-altitude populations [40], we used a 45-day hypobaric hypoxia animal model to delineate hypoxia-induced skeletal muscle dysfunction. Our results demonstrated that sustained hypobaric hypoxia induces skeletal muscle impairment, characterized by reduced contractile performance, histological alterations, and mitochondrial deficits in myotubes. This aligns with observations that mountaineers exposed to altitudes ≥ 5250 m for 75 days exhibited a 15 % average reduction in quadriceps and biceps fiber size [41]. Chronic hypobaric hypoxia exposure decreased mitochondrial density by 21 % and PGC-1α levels by 35 %, potentially suppressing mitochondrial biogenesis and oxidative phosphorylation [42]. These adaptations suggest that prolonged high-altitude exposure may impair mitochondrial function and energy metabolism, thereby compromising exercise performance and muscle energy supply [42]. Protein synthesis is a key determinant of muscle mass and phenotype [43], with proteomic analyses further suggest that accelerated degradation of metabolic and structural proteins under hypoxia. While Sirt1 destabilization under hyperglycemia/oxidative stress is documented, hypoxia-specific suppression of Sirt1 has not been previously reported [21,22]. Notably, we observed significant downregulation of Sirt1 expression in both hypoxic muscles and C2C12 cells. As a central metabolic regulator, Sirt1 governs nuclear deacetylation of histones and transcription factors, affecting epigenetic expression and enzyme activity [26]. A prime example is PGC-1α, a master coordinator of mitochondrial biogenesis and energy metabolism [44]. PGC-1α requires this post-translational modification to attain its biologically active conformation, enabling transcriptional coactivation of oxidative metabolic programs. Additionally, Sirt1 modulates myotube atrophy by regulating acetylation levels of FoxO3 transcription factors. FoxO3 activation is both necessary and sufficient to drive muscle atrophy [45]. Our data suggest that hypoxia-induced Sirt1 depletion disrupts this regulatory balance, promoting FoxO3 and PGC-1α hyperacetylation and subsequent expression of atrophy-related factors.

Previous studies have shown that both exercise training and hypoxia induce a series of adaptive responses, including angiogenesis, mitochondrial biogenesis, and skeletal muscle fiber type transition. Chronic hypoxia elicits restricted morphological and enzymatic adaptive changes in skeletal muscle, yet hypoxic training may enhance specific functional outcomes more effectively than normoxic training [46]. Training-induced changes in protein content or enzyme activity can alter the metabolic response to exercise at the substrate level, thereby improving exercise performance [28]. In our experimental paradigm, mice underwent preconditioning with daily forced running, followed by voluntary wheel running under hypobaric hypoxia. Results showed that besides improved skeletal muscle function, Sirt1, a critical adaptation protein in endurance training, had increased expression in the exercised skeletal muscle. Elevated Sirt1 levels augment myonuclear content, nitric oxide bioavailability, and subsequent satellite cell proliferation, establishing it as a positive regulator of skeletal muscle repair and hypertrophy [25]. We simulated exercise through AICAR to activate AMPK and drive mitochondrial oxidative metabolism in skeletal muscle [47]. Similarly, exercise training activates AMPK and restores Sirt1 function improving skeletal muscle dysfunction through mitochondrial remodeling. AMPK and Sirt1 exhibit convergent biological effects on energy metabolism. AMPK enhances Sirt1 activity by elevating cellular NAD⁺ levels, thereby modulating the deacetylation and activity of downstream targets such as PGC-1α, FOXO1, and FoxO3a [48]. Studies had shown that Gouqi-derived nanovesicles could activate the AMPK/Sirt1/PGC-1α signaling pathway to improve muscle atrophy in mice, highlighting the importance of AMPK/Sirt1 in skeletal muscle mass and function [49].

Exercise improves skeletal muscle regeneration, partly due to enhanced self-renewal of SCs, because SCs activation promotes an increase in myonuclei, and that SCs self-renewal supports the maintenance of SCs and muscle regeneration [38]. Comparative analysis of Pax7 positivity rates revealed that exercise-induced increases in satellite cell abundance may protect skeletal muscle against hypobaric hypoxia-induced damage. The muscle tissue is diverse, consisting of both slow- and fast-twitch fibers, which vary in their contractile protein composition, oxidative capacity, and ATP production substrate preference. Therefore, the proportion of slow- and fast-twitch fibers in skeletal muscle plays a significant role in determining overall energy consumption in the body [1,50]. Our hypobaric hypoxia model demonstrated that exercise not only increased eMyhC-positive fibers but also promoted the fiber transition. This adaptive remodeling of myofiber type composition is critical for maintaining energy homeostasis under hypobaric hypoxia.

However, the impact of environmental hypoxia on skeletal muscle vascular adaptation and capillarization remains debated. Notably, environmental hypoxia alone exerts minimal impact on skeletal muscle capillarization [6]. Exercise training represents a potent stimulus for skeletal muscle capillarization, enhancing both capillary density and capillary-to-fiber ratio [6,33]. Notably, voluntary running (low-intensity, high-volume) elevates capillary density within 5–7 days, whereas forced running (high-intensity, low-volume) requires over six weeks to achieve similar effects [51]. In our experimental paradigm combining preconditioning exercise with voluntary hypoxic training, muscles displayed angiogenesis improvements, likely attributable to synergistic interactions between exercise regimens and environmental hypoxia. This combinatorial approach may amplify vascular adaptation in skeletal muscle. In oxidative-dominant SOL muscles, alongside angiogenesis enhancement, Ki67 staining identified differential endothelial cell proliferation patterns. Collectively, these findings underscore exercise-mediated improvements in skeletal muscle microvascular functionality.

In our study, we rescued hypoxia-induced myotube atrophy and mitochondrial dysfunction by intervening with Sirt1, which is similar to research demonstrating that resveratrol-activated Sirt1 improves skeletal muscle function [52,53]. However, unlike pharmaceutical Sirt1 activators [20], exercise simultaneously targets angiogenesis, satellite cells, and metabolic pathways, which aligns with the concept that skeletal muscle in an exercise state can integrate metabolic signals from the body. Research on sarcopenia suggests that the protective effects of exercise, resveratrol, and their combination on the GAS muscle in aged rats are associated with the AMPK/Sirt1 anti-apoptotic pathway [54,55]. In addition, exercise preconditioning may synergize with intermittent hypoxic exposure to optimize hypoxia tolerance. A recent study highlighted that high-intensity interval training (HIIT) under normobaric hypoxia enhances the expression of mitochondrial biogenesis and dynamics-related proteins in the GAS muscle of rats, thereby improving exercise performance [56].

Overall, our findings demonstrate that chronic hypobaric hypoxia promotes skeletal muscle dysfunction and pre-exercise training enhances skeletal muscle adaptation to hypobaric hypoxia through Sirt1-mediated myotube functional and mitochondrial remodeling. This supports the translational potential of exercise preconditioning as a non-pharmacological strategy to mitigate altitude-related myopathies in high-altitude workers (e.g., miners, mountaineers) and athletes. Future studies should establish standardized protocols for exercise preconditioning (e.g., modality, duration, intensity) tailored to occupational or athletic demands at altitude.

Limitations of study

While our findings provide novel insights into exercise-mediated protection against hypobaric hypoxia-induced skeletal muscle dysfunction, several limitations should be noted. First, our murine model cannot fully recapitulate the multifaceted stressors of natural high-altitude environments. Second, skeletal muscle functions as a central metabolic hub that interacts dynamically with other organs under hypoxia; however, our study focused primarily on intramuscular mechanisms. Finally, future studies should validate these mechanisms in human trials and explore combined interventions (e.g., exercise plus Sirt1 agonists) for extreme-altitude populations.

Data Availability

All data in this study are provided in the paper and Supplementary Data.

Funding

This work was supported by National Key R&D Program of China (2023YFA1801200), National Major Science and Technology Special Project (2023ZD0505300), and the Fujian Provincial Natural Science Foundation (Grant No. 2021J011452 and 2022J011513) .

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: