Role of p66shc in skeletal muscle function

p66shc is a growth factor adaptor protein that contributes to mitochondrial ROS production. p66shc is involved in insulin signaling and its deletion exerts a protective effect against diet-induced obesity. In light of the role of skeletal muscle activity in the control of systemic metabolism and obesity, we investigated which is the contribution of p66shc in regulating muscle structure and function. Here, we show that p66shc−/− muscles are undistinguishable from controls in terms of size, resistance to denervation-induced atrophy, and force. However, p66shc−/− mice perform slightly better than wild type animals during repetitive downhill running. Analysis of the effects after placing mice on a high fat diet (HFD) regimen demonstrated that running distance is greatly reduced in obese wild type animals, but not in overweight-resistant p66shc−/− mice. In addition, muscle force measured after exercise decreases upon HFD in wild type mice while p66shc−/− animals are protected. Our data indicate that p66shc affect the response to damage of adult muscle in chow diet, and it determines the maintenance of muscle force and exercise performance upon a HFD regimen.

and metabolic parameters demonstrated that p66shc deletion does not protect lep Ob/Ob mice from glucose intolerance and insulin resistance 21 , suggesting a different role of p66shc in different obesity models.
Together with liver and fat, skeletal muscle greatly affects organism metabolism. Muscle activity and exercise contribute to the maintenance of glucose tolerance, while inactivity is one of the major risk factors of metabolic syndrome development, type 2 diabetes and subsequent complications 22,23 . p66shc regulates glucose transport in skeletal muscle myoblasts, by modulating the expression of glucose trasporters GLUT1 and GLUT3. In particular, expression of antisense p66shc is sufficient to induce GLUT1 and GLUT3 expression, while overexpression of p66shc causes a reduction of transporters expression and of glucose uptake rate 24 . Studies on the association of exercise with HFD highlighted the positive effect of training to HFD-induced insulin resistance and impairment of glucose uptake, although the scenario remains complex, especially when translated to humans 25,26 . Mitochondria are undoubtedly the main source of ROS in skeletal muscle, and acute exercise has been shown to trigger increased p66shc expression and H 2 O 2 content 27 . However, whether p66shc-induced mitochondrial ROS (mROS) production affects skeletal muscle homeostasis remains to be determined.
In this manuscript, we aimed to determine whether p66shc deletion affects adult skeletal muscle structure and function, both in chow and in HFD, and to clarify whether skeletal muscle contributes to the p66shc −/− metabolic phenotype. Our results suggest that p66shc is required to maintain performance during downhill running in chow diet, while its absence exerts a protective effect against the decrease in running distance and muscle force caused by HFD.

Results
Skeletal muscle fiber size and myosin composition are unaffected by p66shc deletion. To carry out a comprehensive study of the role of p66shc-induced mROS on skeletal muscle homeostasis, we analyzed structure and function of adult muscles of total p66shc −/− mice 5 . After performing a Hematoxylin/Eosin (H&E) staining of transversal cryosections, which detected a normal structure in p66shc −/− muscles (Fig. 1A), we asked whether p66shc deletion affects muscle trophism in resting conditions. Fiber size measurements of tibialis anterior (TA) muscles did not reveal any difference between p66shc −/− and wild type (wt) animals (Fig. 1B), indicating that p66shc does not contribute to healthy muscle size. However, whether reduction in mROS production by p66shc deletion prevents muscle loss is still unclear. Indeed, the role of ROS in muscle atrophy is controversial and greatly depends on the etiology of muscle loss 3 . To clarify this issue, we used a denervation-induced model of muscle atrophy. p66shc mRNA levels were greatly increased by denervation (Fig. 1C), indicative of a denervation-induced redox response. However, fiber size analysis, performed 12 days after cutting the sciatic nerve, demonstrated no differences in size between p66shc −/− and wt hind limb muscles (Fig. 1D), despite subtle differences in fiber size distribution (Fig. 1E). This result indicates that p66shc does not regulate fiber size, neither in normal conditions nor in denervation-induced atrophy. Still we wondered whether p66shc deletion modifies muscle contractile properties and metabolism. Muscles are composed of different fiber types, according to their myosin composition and metabolic features. Myosins are classified from slow to fast according to the scheme 1-2A-2X-2B and myosin composition determines the contractile properties of a fiber 28 . In addition, fibers are also classified based on their metabolic properties, ranging from purely glycolytic to mitochondria-rich oxidative fibers. We evaluated all these parameters in p66shc −/− hind limb muscles. Firstly, we performed an electrophoresis analysis in order to determine myosin composition of gastrocnemius muscles (Fig. 1F). Then, we measured p66shc expression levels in muscles with different metabolic properties, ranging from glycolytic (EDL) to oxidative (soleus) muscles (Fig. 1G). In neither case we observed differences between wt and p66shc −/− muscles.
Next, we investigated whether p66shc deletion affects fiber oxidative metabolism. We discerned glycolytic and oxidative fibers on the basis of a colorimetric assay for the activity of the mitochondrial enzyme succinyl-dehydrogenase (SDH). The proportion of glycolytic versus oxidative fibers was unaffected in p66shc −/− TA muscles, indicating that p66shc does not determine a metabolic shift of muscle fibers ( Fig. 2A). To clarify whether p66shc regulates glucose metabolism of healthy muscles, intracellular glycogen content was evaluated. In both Periodic Acid-Schiff (PAS) staining and glycogen level measurements, no differences were observed between wt and p66shc −/− muscles (Fig. 2B,C). In order to understand whether mitochondrial function was affected by p66shc deletion, we measured pyruvate dehydrogenase (PDH) phosphorylation levels (Fig. 2D), which are in inverse correlation with the activity status of PDH. In addition, we measured Oxygen Consumption Rate (OCR) in single isolated myofibers (Fig. 2E) and ATP production (Fig. 2F). Our data demonstrate that deletion of p66shc does not impinge on basic mitochondrial functions.
Altogether these data demonstrate that p66shc does not contribute to muscle homeostasis of healthy animals, neither in terms of fiber size nor of fiber type and metabolism. Moreover, p66shc does not play a fundamental role in denervation-induced atrophy. p66shc −/− mice exhibit unaltered endurance exercise performance but perform slightly better during downhill running. Whether exercise-induced ROS production causes muscle damage and contributes to muscle fatigue is still debated. Several lines of evidence suggest that this is not always the case. For example, in specific settings, treatment with antioxidants reduces exercise performance 2,29 . We asked whether p66shc-dependent ROS production contributes to exercise performance. To answer this question, we decided to compare maximal running distance and muscle force of p66shc −/− and wt mice. Exhaustion exercise performance was evaluated by a single bout of run on a treadmill. Wt and p66shc −/− mice performed similarly in terms of running distance (Fig. 3A). However, the beneficial effects of ROS on exercise adaptation are particularly evident during repeated bouts of damaging eccentric exercise 2 . Thus, we hypothesized that p66shc deletion would blunt this effect, causing a progressive loss of muscle performance, as occurs in anti-oxidant treated mice 2 . Accordingly, experiments of exhaustive downhill running were performed in three consecutive days and running distances SCIeNtIfIC RepoRTs | 7: 6383 | DOI:10.1038/s41598-017-06363-0 were recorded. p66shc deletion caused an almost significant reduction in running performance, indicating that p66shc-dependent ROS production is instrumental for muscle recovery upon damage (Fig. 3B).
Trophic and metabolic properties of p66shc −/− muscles upon HFD. Increased oxidative stress is both a trigger and a consequence of obesity and antioxidant treatment improves metabolic parameters 30 . It has been demonstrated that p66shc −/− mice are protected from HFD-induced obesity 14,15 . Similarly, protection by p66shc deletion was observed in lep Ob/Ob mice 20 , although the beneficial effects on glucose intolerance and insulin resistance have been questioned 21 . Muscle atrophy and weakness have been linked to obesity 31,32 . We wondered whether skeletal muscle fiber size, myosin composition and metabolic parameters are affected by an experimental HFD regimen in which 60% of the total energy derives from fat, and whether p66shc plays a role in this context. Female mice were fed with HFD for 4 and 9 months. As reported 14 , increased body weight was apparent in wt mice, while it was less pronounced in p66shc −/− mice (Fig. 4A). We investigated whether this HFD treatment triggers ROS production in skeletal muscle and whether p66shc is required. Accordingly, analysis of carbonylated proteins on muscle extracts was performed. HFD caused an increase in ROS production only in wt samples (Fig. 4B), indicating that p66shc is required for HFD-induced ROS generation in skeletal muscle. However, muscle weight was similar between chow and HFD-fed mice, in both backgrounds (Fig. 4C). Moreover, fiber size analysis showed no differences between chow (reported in Fig. 1B) and HFD-treated mice (Fig. 4D) in both genotypes. We also asked whether HFD regimen affects myosin composition, compared to data reported in Fig. 1F. Quantitative myosin analysis demonstrated that HFD does not affect the relative myosin distribution, neither in wt, nor in p66shc −/− muscles (Fig. 4E). We reasoned that subtler metabolic parameters could be affected. However, H&E staining of HFD muscle cryosections, both wt and p66shc −/− , were compatible with a healthy tissue structure (Fig. 4F), although lipids visualization by Oil Red O (ORO) staining demonstrates lipid accumulation in interstitial spaces (Fig. 4G). We monitored mitochondrial activity by SDH staining, without observing important differences between p66shc −/− and wt muscles (Fig. 4H). Lastly, qualitative and quantitative measurements of glycogen content (as reported in Fig. 2B,C) demonstrated that HFD regimen does not significantly affect this parameter (Fig. 4I-J).
Overall these data demonstrate that, in our experimental conditions, HFD, independently of the contribution of p66shc-induced ROS, does not significantly alter trophic and metabolic properties of skeletal muscle. p66shc −/− mice maintain muscle force and exercise performance on HFD regimen. We wondered whether HFD-induced obesity affects exercise performance and force generation, and whether these are controlled by p66shc. Mice were subjected to a single bout of strenuous exercise on a treadmill and running distance was measured. As reported in Fig. 3A, wt animals on chow diet run about 1,200 meters. When kept in HFD for 4 months, running distance drastically decreased to about 870 meters, and after 9 months of HFD average running distance was only 400 meters (Fig. 5A). On the contrary, p66shc −/− mice, which showed a maximum average running distance of 1,200 meters in chow diet (see Fig. 3A), maintained similar running distance after 4 and 9 months of HFD (Fig. 5A). In order to understand if the reduction in exercise performance in wt mice fed with HFD for 9 months could be explained, at least partially, by a reduction in muscle force, we first examined the effects of the diet on tetanic strength. HFD did not affect gastrocnemius tetanic force production of sedentary mice in vivo (Fig. 5B). However, when force was measured immediately after one bout of strenuous exercise, wt mice showed a marked reduction (Fig. 5C). Interestingly, despite the fact that p66shc −/− mice run significantly further, they didn't show any reduction in normalized force after a treadmill run until exhaustion.
These results suggest that diet-induced obesity affects both the performance during resistance training and the muscle force generation after exercise, and that p66shc deletion exerts a protective effect, both impinging on whole organism metabolism, and on skeletal muscle per se.
Regulation of muscle autophagy and UPR markers by p66shc. The autophagy-lysosome system plays a major role in the maintenance of skeletal muscle homeostasis. A basal level of autophagy is constantly present in resting conditions, where it guarantees the removal of dysfunctional macromolecules and organelles. Physiologic and pathologic triggers are responsible for an increase of autophagy flux beyond the resting level. For example, muscle autophagy is induced by starvation 33 . Autophagy is also induced by exercise 34,35 and contributes to its beneficial metabolic adaptations, as demonstrated in a mouse model in which autophagy induction is prevented 35 . A detailed analysis of different exercise protocols demonstrated that muscle-specific inhibition of autophagy does not compromise running performance in a uphill treadmill exercise 2 . However, the same study demonstrated that muscle autophagy is required for the maintenance of mitochondrial function during damaging downhill exercise, which is characterized by increased ROS production 2 . Thus, we asked whether p66shc contributes to exercise-induced autophagy. Muscles were withdrawn immediately after a single bout of exhaustion running on a treadmill and protein and mRNA levels of autophagy markers were measured. As already reported 2,34,35 , also in our model exercise induced LC3B lipidation in wt muscles. A similar effect was observed in p66shc −/− muscles (Fig. 6A). In addition, although Beclin expression was mostly unaffected by the exercise protocol (Fig. 6A), Bnip3 and Bnip3l transcription was induced by exercise both in wt and in p66shc −/− muscles (Fig. 6B,C). These data confirm the induction of autophagy upon exercise but do not support a role of p66shc in this process.
We then hypothesized that p66shc may play a role in the regulation of the unfolded protein response (UPR), which maintains endoplasmic reticulum (ER) homeostasis upon stress, and has been reported to contribute to adaptation of skeletal muscle upon exercise. Indeed, ER stress and UPR markers are induced upon a bout of exhaustive treadmill running 36 and deletion of ATF6α, one of the UPR sensors 37 , hampers recovery from muscle damage upon repeated bouts of exhaustive exercise 36 . In agreement with previous data 36 , expression of ATF3 and ATF4, two ER stress markers, was induced in hind limb muscles upon one bout of exhaustive exercise, however no differences were observed between wt and p66shc −/− muscles (Fig. 6D,E).
In light of the positive effect of p66shc deletion on muscle performance and force in HFD, we wondered whether these two pathways, i.e. autophagy and UPR, are dysregulated in these conditions and whether p66shc contributes to their modulation.
Analysis of LC3B lipidation demonstrated that autophagy is induced by exercise also in HFD-treated mice. However, protein levels of Beclin and mRNA expression levels of LC3, Bnip3 and Bnip3l were not significantly affected by exercise. In addition, p66shc deletion was not sufficient to restore autophagy gene expression upon  exercise (Fig. 7A-D). These data demonstrate that the differences in muscle force between HFD-fed p66shc −/− and wt mice after exercise could not be explained by variations in autophagy modulation among the two genotypes.
Concerning ER stress and UPR response, an interesting indication came from the analysis of ATF3 expression. Indeed, as in chow diet, ATF3 was induced in wt muscles after exercise upon 9 months of HFD. However, the increase in ATF3 expression was less pronounced in p66shc −/− muscles, suggesting a potential role of ER stress in the differential response to exercise of HFD treated p66shc compared to wt mice (Fig. 7E).
Finally, basis on the importance that ROS play in muscle adaptation to exercise, we investigated the role of ROS scavengers. However, mRNA expression analysis confirmed the absence of a modulation of endogenous antioxidant systems by exercise in HFD, regardless of the genotype (Fig. 7F-I).

Discussion
ROS play a dual role in tissue homeostasis. At low, physiological concentrations, ROS act as positive and indispensable signaling molecules. On the other hand, dysregulated ROS production triggered by pathological stimuli causes damages to several cell structures which, if not adequately counteracted by ROS scavengers and repair enzymes, negatively affects cell function 38 . The boundary between these two opposite scenarios is not completely defined and, in specific settings, it is difficult to discern whether ROS are playing a beneficial or rather a damaging role. For example, in the case of skeletal muscle, the role of ROS is still controversial. If on one side the consensus is that damaging ROS production is increased by exercise and contributes to muscle fatigue, on the other hand the use of antioxidants has been proven to negatively affect muscle performance 1, 2 . Normalized maximal force production from gastrocnemius muscles of mice fed with chow diet or HFD for 9 months. Data are presented as mean ± SEM (7 animals per group). (C) Normalized maximal force production from gastrocnemius muscles of mice fed with HFD for 9 months before and after a treadmill run until exhaustion. ***p < 0.005, pairwise multiple comparison (Holm-Sidak method) of 7 animals per condition. Data are presented as mean ± SEM.
SCIeNtIfIC RepoRTs | 7: 6383 | DOI:10.1038/s41598-017-06363-0 A significant amount of ROS originates as side product of mitochondrial respiration 39 . In vivo studies of mROS-related effects have been facilitated by the p66shc −/− mouse model 5 . p66shc belongs to the shc family of adaptor proteins encoded by SHC1 gene which includes p66shc, p46shc and p52shc. Unique among the family, p66shc translocates to the mitochondrial matrix upon oxidative stress and induces mROS production 6 . Thus, the p66shc −/− mouse appears to be an ideal model for the study of mROS in different organs and tissues and in various pathophysiological conditions. The most striking phenotype of p66shc −/− mice is the fact that they are protected from HFD-induced obesity 14 .
We decided to explore whether p66shc deletion affects adult skeletal muscle structure and function, both in chow and in HFD in order to understand whether skeletal muscle contributes to the p66shc −/− metabolic phenotype. In chow diet, p66shc appeared totally dispensable for the determination of muscle structure and metabolic properties, indicating that p66shc-dependent mROS do not play a major role in resting adult skeletal muscles. The response to denervation-induced atrophy was also unaffected by loss of p66shc. However, whether increased ROS contribute to muscle loss is not completely understood, and certainly depends on the atrophy model 3,4 . We reasoned that p66shc-dependent mROS could still play a role during muscle activity. Running performance of p66shc −/− mice during strenuous exercise was undistinguishable from wt animals, however, during downhill running, a condition that has been associated to ROS-induced muscle damage 40 , p66shc −/− mice were slightly less fatigued than wt. Thus, despite the minor role of p66shc in resting adult skeletal muscle, its deletion has relevant effects in conditions of high levels of cell stress. The second aspect that we decided to investigate was the contribution of skeletal muscle to p66shc −/− phenotype in HFD. p66shc −/− mice are protected from diet-induced obesity and maintain glucose tolerance and insulin sensitivity 14,15 . In agreement with previous reports, the HFD protocol used in our study caused obesity in wt animals but not in p66shc −/− animals. Analysis of carbonylated proteins indicates increased ROS production by HFD, which is hampered in the absence of p66shc. However, a structural analysis of skeletal muscles upon HFD did not reveal substantial differences with chow-treated mice, neither in wt nor in p66shc −/− mice, suggesting that induction of ROS within this HFD protocol is not sufficient to trigger atrophy and metabolic dysfunctions. While obese wt animals were characterized by impaired running ability, muscle performance of p66shc −/− animals was unaffected. The most obvious reason could be the fact that p66shc −/− mice were leaner. However, the direct contribution of skeletal muscle force to this result is demonstrated by the fact that measurements of normalized muscle strength after exercise revealed a decrease in tetanic force only in wt animals, while no reduction in force was observed in p66shc −/− muscles.
Concerning the mechanism, our analysis of exercise-induced stress-related pathways, such as autophagy and UPR, did not reveal a strong connection with the observed phenotypes, indicating that further investigation is required.
Thus, our results indicate that the contribution of p66shc to muscle function becomes evident in conditions of maximal stress, like those occurring during strenuous exercise, while its role is dispensable for muscle homeostasis in resting conditions. A dual role is underlined: while in chow diet p66shc plays a protective role during downhill exercise, presumably contributing with essential ROS, in HFD the associated metabolic alterations prevail, and p66shc deletion protects muscles from loss of force and performance. Further studies will clarify the mechanism underlying these effects. The p66Shc −/− mouse strain carries a targeted mutation of the Shc CH2 exon, as described in 5 . C57BL/6 J mice were used as wt controls.

In vivo
Denervation was achieved by cutting the sciatic nerve high in the thigh. 12 days later histological analysis on muscle cryosections were performed as reported below.
For HFD experiments, after weaning animals were fed for 4 or 9 months with a diet in which 60% of the total energy derives from fat (Mucedola). Wt and p66Shc −/− mice were randomly assigned to the HFD or to the standard diet group.
Exercise was performed by subjecting mice to a forced run on a treadmill device until exhaustion. Briefly, mice were first left to adapt to the treadmill for 5 min at a speed of 5 cm/sec. After this, exercise was performed by starting with a speed of 10 cm/sec that was increased with 2 cm/sec every 2 min until a maximum of 40 cm/sec. Once mice were unable to run for 5 seconds consecutively, they were removed from the treadmill. The eccentric training protocol consisted of 1 to 3 days of treadmill running to exhaustion, with a downhill 10° decline. After 15′ of adaptation from 7 to 17 cm/sec, exercise was performed by starting with a speed of 17 cm/sec that was increased with 2 cm/sec every 10 min until mice were unable to run for 5 seconds consecutively. Total running distance of each animal were recorded.
To measure muscle force in living animals, the contractile performance of gastrocnemius muscle in vivo was measured as described previously 41 . Briefly, anesthetized mice were placed on a thermostatically controlled table, keeping the knee stationary, and the foot firmly fixed to a footplate, which was connected to the shaft of the motor of a muscle-lever system (305B, Aurora Scientific). Contraction was elicited by electrical stimulation of sciatic nerve. Teflon-coated seven-stranded steel wires (AS 632, Cooner Sales) were implanted with sutures on either side of the sciatic nerve proximal to the knee before its branching. At the distal ends of the two wires, the insulation was removed, and the proximal ends were connected to a stimulator (S88, Grass). To avoid recruitment of the dorsal flexor muscles, the common peroneal nerve was cut.
Histology and fluorescence microscopy. For  For glycogen content, 20 μm thick cryosections were stained using PAS staining system (Sigma) according to manufacturer's instructions. Briefly, cryosections were incubated in Periodic acid solution for 5 min, washed, incubated in Schiff 's reagent for 15 min, washed again, and incubated in Gill's hematoxylin solution for 2 min. Finally, after washing, cryosections were dehydrated.
To detect lipids content, Oil Red O (ORO) staining procedure was used as previously reported 43 . Briefly, cryosections were incubated in a freshly prepared Oil Red O working solution for 15 min. After being washed, they were incubated in hematoxylin solution for 30 seconds. Finally, after 3 washes, cryosections were dehydrated.
Protein carbonyls detection. Oxidation of muscle proteins was determined by detecting the presence of carbonyl groups. For this purpose the OxyBlot Protein Oxidation Detection Kit was used (Millipore). Briefly, the carbonyl groups in the protein side chains were derivatized to 2,4-dinitrophenylhydrazone (DNP-hydrazone) by reaction with 2,4-dinitrophenylhydrazine (DNPH). The DNP-derivatized protein samples were separated by polyacrylamide gel electrophoresis followed by Western blotting with primary antibody specific to the DNP moiety of the proteins. All values were normalized for the housekeeping protein Actin (Santa Cruz).
Muscle ATP extraction and ATP quantification. ATP was extracted from frozen sections of TA muscles.
Tissue was homogenized with 1 ml of ice-cold phenol-chloroform-isoamylalchol (25:24:1). 200 µl of PBS were added to the homogenate, and the latter was centrifuged at 14000 rpm for 10 min at 4 °C for phase separation. 20 µl of the upper aqueous phase was used to measure ATP. ATP measurements were performed with ATPlite Luminescence Assay System (Perkin Elmer) according to manufacturer's instructions. Final results were normalized for muscle weight.
OCR (oxygen consumption rate) experiment. The rate of oxygen consumption was assessed in intact fibers using the XF24 Extracellular Flux Analyzer (Seahorse Biosciences), which allows the monitoring of OCR changes after up to four sequential additions of compounds (oligomycin 2 μM, FCCP 0.6 μM, rotenone 1 μM and antimycin A 1 μM). Fibers were prepared as reported above. A titration with the uncoupler FCCP was first performed, in order to optimize the FCCP concentration (0.6 μM) that maximally increases OCR. OCR measurements were normalized for the amount of fibers using Calcein as cellular marker. At the end of OCR measurements, fibers were loaded with Calcein-AM (2 μM, Sigma) for 20 min, and fluorescence (excitation 485/10 nm, emission 525/30 nm) was measured in well-scan mode using a Perkin Elmer EnVision plate reader.
SCIeNtIfIC RepoRTs | 7: 6383 | DOI:10.1038/s41598-017-06363-0 Statistical methods. Statistical data are presented as mean ± SEM; significance was calculated by t test or pairwise multiple comparison according to the number of independent factors present in each experiment, as detailed in the figure legends. *p < 0.05, **p < 0.01, ***p < 0.001. Raw data can be found as Supplementary  Table S1.