Myostatin deficiency is associated with lipidomic abnormalities in skeletal muscles

Myostatin (Mstn) deficiency leads to skeletal muscle overgrowth and Mstn inhibition is considered as a promising treatment for muscle-wasting disorders. Mstn gene deletion in mice also causes metabolic changes with decreased mitochondria content, disturbance in mitochondrial respiratory function and increased muscle fatigability. However the impact of MSTN deficiency on these metabolic changes is not fully elucidated. Here, we hypothesized that lack of MSTN will alter skeletal muscle membrane lipid composition in relation with pronounced alterations in muscle function and metabolism. Indeed, phospholipids and in particular cardiolipin mostly present in the inner mitochondrial membrane, play a crucial role in mitochondria function and oxidative phosphorylation process. We observed that Mstn KO muscle had reduced fat membrane transporter levels (FAT/CD36, FABP3, FATP1 and FATP4) associated with decreased lipid oxidative pathway (citrate synthase and β-HAD activities) and impaired lipogenesis (decreased triglyceride and free fatty acid content), indicating a role of mstn in muscle lipid metabolism. We further analysed phospholipid classes and fatty acid composition by chromatographic methods in muscle and mitochondrial membranes. Mstn KO mice showed increased levels of saturated and polyunsaturated fatty acids at the expense of monounsaturated fatty acids. We also demonstrated, in this phenotype, a reduction in cardiolipin proportion in mitochondrial membrane versus the proportion of others phospholipids, in relation with a decrease in the expression of phosphatidylglycerolphosphate synthase and cardiolipin synthase, enzymes involved in cardiolipin synthesis. These data illustrate the importance of lipids as a link by which MSTN deficiency can impact mitochondrial bioenergetics in skeletal muscle. Due to energetic and structure function of fatty acids and phospholipids in muscle, we hypothesized that in MSTN deficient hypertrophied skeletal muscle, hence a change in the muscle and mitochondrial lipid composition could exist, in relation with metabolic and mitochondrial abnormalities observed in this phenotype. In this study, we evaluated the lipid profile of muscle and mitochondrial membranes from Mstn KO mice, and explored muscle lipid metabolism pathway. We demonstrated in MSTN deficient muscle a decrease of both lipogenesis and lipid oxidation associated with important modifications in the fatty acid composition of total muscle lipid profile. Interestingly, we also found that muscle mitochondrial proportion of cardiolipin is significantly decreased in Mstn KO mice correlating with the reduced gene expression of two enzymes involved in cardiolipin synthesis. These results reveal alterations of the phospholipid profile in skeletal muscle of Mstn KO mice. We suggest that this could be mechanism through which lack of MSTN impacts mitochondrial function. muscle homogenate on silica gel 60 HPTLC plates pre-treated with 1.5% w/v boric acid in ethanol, was automatically performed on a 4-mm band width using a CAMAG ATS4 apparatus Switzerland). Two developments were performed, first with pentane–chloroform–methanol (52:45:3, by volume) on 60-mm migration distance and second with pentane–diethyl ether (97:3 by volume) on 67-mm migration distance, which allowed the separation of neutral lipids. The scanning of the plates was carried out using a CAMAG TLC scanner 3 (Muttenz, Switzerland) operating in the reflectance mode. The plates were scanned at 550 nm after dipping in a solution of copper sulphate 640 mM in H3PO4 1·18 M and heating for 20 min at 180°C. The 1-monoacylglycerol (MAG), 1,2- diacylglycerol (DAG) and 1,3-DAG, TG, cholesterol (CHL) and FFA contents were finally identified by comparing their retention factor (Rf) with authentic standards and were quantified using calibration CL CL mstn WT CL/mg mitochondria quantitative mstn CL. on CL control mitochondria skeletal cardiolipin and mitochondria mstn to WT µmoL CL observed a decrease in the proportion of CL in provide evidence that and mitochondrial


Introduction
Myostatin (Mstn), a secreted growth factor and member of the TGF-β superfamily, regulates levels of lean muscle mass and also body fat content in mice [1]. Inactivation of the Mstn gene in mice, or mutations in the bovine [2,3] ovine [4] and human Mstn [5] genes result in a similar phenotype of increased muscle growth. Thus, targeted inhibition of Mstn gene has been considered as a promising treatment for various muscle-wasting disorders [6,7,8,9]. In this case, strategies have been developed to treat muscle dystrophies, muscle wasting and myopathies by blocking the Myostatin/ActRIIB pathway. Many clinical trials are in process but results are associated with undesirable adverse effects [10]. Therefore, a better understanding of the mechanisms underlying the muscle phenotype in Mstn knock-out (KO) model is warranted that may ultimately improve treatments.
Indeed, beyond muscle hypertrophy, Mstn KO mice show a disturbed muscle function with loss of muscle strength and endurance in vivo or ex vivo accompanied with a decrease in mechanical performance, and ATP production during exercise [11,12,13,14]. In parallel, many other metabolic changes have been documented in Mstn KO muscle such as a decrease in mitochondrial content, disturbance in mitochondrial respiratory function with a decay in the respiratory control ratio in intermyofibrillar mitochondria, and a decline in porine activity [11,13,15]. In addition, a decrease of fatty acid translocase (FAT/CD36) level, a lipid transporter, and the citrate synthase (CS) activity, one of the main mitochondrial oxidative enzyme, associated with a lower peroxisome proliferatoractivated receptor gamma coactivator 1 alpha (PGC1-α) expression has been observed [15,16].
Although it is clear that lack of MSTN impairs muscle oxidative metabolism and mitochondrial function, the underlying mechanisms by which this alteration occurs have not been fully elucidated.
It is known that muscle mitochondrial function and oxidative metabolism are both highly reliant on the lipid pathway. Indeed, fatty acids (FA) are the main substrate for ATP production in muscle via the β-oxidation pathway [17]. Entering muscle fibers from blood via different transporters, lipids are then oxidized to Acyl-CoA in the cytosol then to acylcarnitine before translocating through mitochondria membranes by carnitine palmitoyltransferase I and II. These substrates in turn are subjected to oxidation by beta-hydroxyacyl CoA dehydrogenase (β-HAD) and CS to produce ATP.
Lipogenesis encompasses the processes of fatty acid and triglyceride synthesis from glucose [18].
After glycolysis, pyruvate is produced from glucose which then enters mitochondria. Through the tricarboxylic acid cycle (TCA), pyruvate is converted to acetyl-CoA and exported to cytosol to be used as a precursor for lipid synthesis. Indeed, acetyl-CoA carboxylase (ACC) converts acetyl-CoA into malonyl-CoA, which finally activates synthesis of triglycerides, using the fatty acid synthase enzyme (FAS) [18]. In the case of high-fat diet, fatty acids are also stored as triglycerides.
A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
5 of FA and phospholipid composition [19]. For example, being the most abundant phospholipid in both the outer mitochondrial membrane and the inner mitochondrial membrane, phosphatidylcholine (PC) modulates skeletal muscle mitochondrial function [20]. Genetic studies in mice or humans with defects in coline kinase β, enzyme catalysing the synthesis of PC, demonstrate that reduction of mitochondrial membrane PC content induced enlarged mitochondria, reduced inner membrane potential and caused muscle weakness [20][21][22][23]. The proper lipid composition of the membranes is crucial for membrane-bound protein machineries to maintain optimal mitochondrial biogenesis and function. The recent review of Martensson et al., 2016 illustrated that phospoholipids affect outer (TOM complex, MIM, SAM) and inner (TIM23, TIM22 and OXA) membrane protein translocases and thus play a crucial role in mitochondrial protein import [24].The role of phospholipids on mitochondria protein is also highlighted by their impact on respiratory chain supercomplexes [24]. In these studies, one phospholipid, cardiolipin (CL) attracted scientific attention. Mostly present in the inner mitochondrial membrane, CL undergoes two main phases for its activity [24,25]. The first is its biosynthesis that takes place on the matrix face of the inner mitochondrial membranes, pathway catalysed by numerous and successive enzymes as Tam41, Pgs1, PTPMT1 and finally by CL synthase (also called Crd1), to form premature (nascent) CL. The second phase is the remodelling of its fatty acids, with recruitment of a new set of tissue-specific fatty acids, and its assembly into the inner membrane. Nascent CL is deactyled by CL deacylases/phospholipases to monolyso-CL which is reactylated principally by the taffazin to mature CL [19,25]. CL interacts with a number of inner mitochondrial membrane proteins, enzymes and metabolite carriers, such as the electron transport chain complexes involved in oxidative phosphorylation and ADP/ATP carriers, which optimize efficient electron/proton flux and hence, ATP synthesis [26]. CL is also involved in the β-oxidation of fatty acids, as a reduction of mature CL impairs skeletal muscle FA oxidation [27,28] and a positive correlation has been demonstrated between mitochondrial COX activity and CL content [29]. Patients with Barth Syndrome (a X-linked disorder resulting from mutations in the gene encoding for tafazzin) or mice models which lack mature CL, have shown exercise intolerance, cardiac and skeletal muscle dysfunction, mitochondrial dysfunction, ATP deficiency and premature mortality [30,31,32] In vitro, mitochondrial respiration has been shown to be impaired in inducible pluripotent stem cell (iPSC)derived human cardiomyocytes from Barth Syndrome patients [33]. Possible mechanisms for this disease-associated mitochondrial dysfunction include mitochondrial supercomplex destabilization [34], higher reduced levels of mitochondrial cardiolipin [35] and abnormal mitochondrial morphology [30]. rely on glycolytic metabolism to a greater extent than unaffected, age-matched controls, likely due to an impairment in fat oxidation. These findings highlight the crucial role that mitochondrial CL plays in skeletal muscle oxidative metabolism and mitochondria bioenergetics processes [26].

Animals
Male Mstn KO mice (10 wk old) used in this study have been described previously and were generously provided by L. Grobet (Faculty of Veterinary Medicine, University of Liège, Belgium) [37]. These mice harbor a constitutive deletion of the third Mstn exon leading to the deletion of the entire mature COOH-terminal region of MSTN and were therefore null for MSTN function. The mice were generated on a FVB/N-C57BL genetic background. WT (n=20) and Mstn KO mice (n=20) were produced from homozygous matings. Parental genotyping were determined by polymerase chain reaction analysis of tail DNA. Mice were fed ad libitum and kept under a 12:12-h light-dark cycle.
The experimental protocols of this study were handled in strict accordance with European directives (86/609/CEE) and approved by the Ethical Committee of Region Languedoc Roussillon (APAFIS#2551-2015110311365663v2).

Muscle preparation and sample
Mice were weighed and killed by cervical elongation. Tibialis anterior, extensor digitorum longus (EDL), and quadriceps muscles were quickly excised and immediately placed in ice-cold buffer

Mitochondria isolation
Total mitochondria were fractionated by differential centrifugation as described previously [38].
Briefly, muscles were freed of connective tissues, minced, and homogenized with a Potter-Elvehjem homogenizer. Then, mitochondria were obtained from the initial pellet following a treatment with Subtilisin A (0.25 mg/g wet muscle) and subsequent centrifugations at 800 and 9,000 g. Total mitochondria were resuspended in 100 mM KCl and 10 mM MOPS, pH 7.4. Mitochondrial protein content was determined using the Bradford assay, and the yield was expressed as milligram of mitochondrial proteins per gram of muscle wet weight.

Lipids extraction and thin layer chromatography (TLC) Triglyceride (TG) and free fatty acids (FFA) content in muscle lipid extracts by TLC-densitometry
Muscle lipids were extracted using a mixture of chloroform-methanol (2:1 by volume) in the presence of butylated hydroxytoluene (50mg/l) according to [39]. The application of lipid extracts of were quantified using calibration curves of the same standards [40].

FFA composition of muscle by gas chromatography
Fatty acid methyl esters were prepared by incubation with acidified methanol, according to the method of Lepage and Roy [41] , after the addition of an internal standard (500 µg C17:0/mL) to lipids in the Folch extract. Briefly, the methylation reagent was generated by mixing sulphuric acid

Tissue desaturase indices and unsaturation index
Since it is not possible to directly measure the activity of the enzymes that catalyze the desaturation 6)]. The unsaturation index (UI) was calculated from the relative percentage of each type of mono-and polyunsaturated fatty acid multiplied for the number of double bonds present in the molecule [43].

Phospholipid composition of muscle mitochondrial membrane by TLC-densitometry
Muscle mitochondrial suspensions were extracted by a mixture of chloroform/methanol (2:1 by volume) according to [39] in the presence of 50 mg/L of butylated hydroxytoluene. Phosphorus was quantified on Folch extracts of mitochondrial suspensions in order to determine total phospholipid quantity as previously described [44]. and quantified using calibration curves of the same standards [45]. Results are expressed as percentage of nmol of phosphorus in total phospoholipids, and indicate for each phospholipid its proportion (%) in the composition of mitochondrial membrane. We also estimate the mitochondria CL content expressing in µmoL pf phosphorus in CL per mg of mitochondrial protein.

Enzyme activity assays
Lipids oxidation was evaluated by CS and β-HAD activities. CS activity was measured according to Srere et al., 1969 [46] : the activity of the enzyme was measured spectrophotometrically at 412 nm by following the yellow color of 5-thio-2-nitrobenzoic acid, which is generated from the reaction of 5,5'dithiobis-2-nitrobenzoic acid with the free -CoA liberated from the reaction of acetyl-CoA and oxaloacetate under citrate synthase action to synthesize citrate. β-HAD activity was measured according to [47]. The activity of the enzyme was measured spectrophotometrically at 340 nm by following the disappearance of NAD. The β-HAD converted the acetoacetyl-CoA in alcohol inversely β-oxidation reaction.

Quantification of mRNA expression by real time reverse transcription (RT-qPCR)
Total RNAs were isolated from gastrocnemius muscle powder using the RNeasy Fibrous  All PCR efficiencies were above 95%. The relative abundance of each sample was then normalized according to the equation: Relative Quantity = 2 - [48].

Statistical analysis
All data are presented as means ± SEM. Data were compared between the two groups of mice WT and Mstn KO using an unpaired t-test or a Mann-Whitney rank sum test when normality was not obtained. The significance level was set at 0.05. The data were analyzed using the statistical package Graphpad Prism version 6.02 for windows (GraphPad Software, La Jolla, California).

Animals and muscle characteristics
Consistent with the hypermuscular phenotype, at 10 weeks of age, Mstn KO mice displayed greater body and muscle (gastrocnemius, EDL, soleus) weights compared to their WT littermates ( Table 2).

Decreased lipid oxidation in Mstn KO muscle
We and others have recently shown that inhibition of Mstn represses muscle endurance and metabolic mitochondrial function [13,15]. As this metabolic phenotype is greatly dependent on lipids, we first investigated whether myostatin deletion would impact mitochondrial lipid oxidation in muscle.
Mitochondrial yield and CS quantity were significant reduced in mstn gastrocnemien muscles ( Figure 1A and B) confirming the reduced mitochondrial pool in mstn KO mice vs WT. The activity of isolated key mitochondrial enzymes, CS and β-HAD, was decreased in gastrocnemius muscle homogenates from KO compared to WT littermates ( Figure 1C). Supporting the decrease of lipid oxidation, we found reduced mRNA level of Cpt1, a membrane mitochondrial lipid transporter ( Figure 1D). Furthermore, mRNA levels of Ppar-, a key transcription factor promoting oxidative metabolism is also markedly reduced in mstn deficient muscle ( Figure 1D). Altogether, these data suggest that oxidative metabolism and lipid β-oxidation is down-regulated in the absence of myostatin.
In agreement with these results, we observed decreased protein content of CD36, FABP3 (fatty acid binding protein 3), FATP1 and FATP4 (fatty acid transport proteins 1 and 4), all of which are involved in lipid membrane and cytosolic transport, in gastrocnemius muscle of KO mice compared to WT mice (Figures 2A, 2B). Inversely, we showed that protein involved in muscle glucose uptake, as hexokinase II (HKII) and glucose transporter type 4 (GLUT4) protein content were respectively significant increased or tended to be increased in mstn KO muscle in comparison with control group (Figure 3).

Decreased lipogenesis in Mstn KO muscles
Since lipogenesis involves synthesis of fatty acids from excess carbohydrates leading to TG storage, we next tested whether MSTN deficiency altered this metabolic pathway. Western blot analysis showed that FAS levels were decreased in Mstn KO gastrocnemius compared to WT (Figure 4A). In agreement with this, we found that TG levels were 4-fold lower in Mstn KO mice compared with WT ( Figures 4B, 4C).

Modification of fatty acid composition in Mstn KO muscle
The structural unit of lipids is the FA, which is divided in three groups: saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs).

Modification in phospholipids composition in Mstn KO muscle
Then, we identified and compared the lipid patterns of gastrocnemius muscle from wild type and Mstn KO mice between the different phospholipids quantity by thin layer chromatography. Among different phospholipids analyzed, we observed a significant decrease in CL levels in Mstn KO suggesting modifications in mitochondria membrane phospholipid composition (Figure 6).

Impaired phospholipid mitochondrial membrane composition in Mstn KO muscle.
As Mstn deficiency led to a congenital shift towards fast glycolytic muscle fibers and a reduction in mitochondrial content, afterwards we identified and compared the lipid patterns of mitochondria membranes from wild type and Mstn KO mice using the same technique using mixed muscle samples.

Cardiolipin synthesis pathway in Mstn KO muscle mitochondria
Thereafter, we examined the enzymes involved in CL synthesis (CLS: Cardiolipin synthase, Pgps: phosphatidylglycerolphosphate synthase) and CL remodeling (Alcat1: ER-associated monolyso-CL acyltransferase, Taz: tafazzin). At the protein level, the amount of CLS is lower in the mstn deficient muscle normalized by -tubulin or by CS abundance (Figure 8A). Furthermore mRNA level of Pgps was also significantly decreased in Mstn KO muscle (Figure 8B) whereas the expression of the genes involved in the remodeling of CL acyl chains (Alcat1, Taz) did not change.

Myostatin and muscle lipid metabolism
Here, we showed that Mstn gene deletion results in reduced lipolytic machinery in Mstn deficient hypertrophic muscles (impaired mitochondrial yield, CS and β-HAD activities and Cpt1 and Ppar- gene expressions), in line with previous studies showing that oxidative metabolism is diminished in several models of Mstn deficiency [12,13,15,16,51]. In addition, we demonstrated that the levels of cytosolic (FABP3) and sarcolemmal lipid transporters (FATP1, FATP4) are reduced, in agreement with our previous observation showing a significant reduction of FAT/CD36 levels [16]. It should be noted that CD36, FABP3, FATP1 and FATP4 are co-expressed in skeletal muscle, and all of these proteins contributed to transport, with recent evidence suggesting that CD36 and FATP4 are qualitatively the most important [53]. Our results indicated that the process regulating fatty acids uptake into muscle by protein-mediated transport is negatively and strongly affected by MSTN and insulin sensitivity thanks to an increase in glucose uptake in muscle and adipose tissues [54]. In the same vein, Relizani et al. reported that the blockade of mstn decreased mRNA levels of Pdk4, a regulatory protein switch of substrate utilization from glucose towards fatty acids [15]. In a previous study, we observed in aged Mstn KO EDL muscle a significant increase in GLUT4 protein levels, and a decrease in FAT/CD36 protein levels compared to aged WT mice [16]. In the present study, we confirm the increased content of muscle glucose uptake markers in mstn KO muscle. Given that FA and glucose are the main energy sources in skeletal muscle, these results suggest that MSTN could contribute to the fine regulation for the use of these two substrates. Our results demonstrate that mstn deficient muscle malfunctions for fat handling that could explain metabolic shift towards glycolytic pathway.

Myostatin and membrane phospholipids composition
Phospholipids affect function of protein machineries involved in respiratory metabolism, protein import, membrane architecture and mitochondrial dynamics [24]. At this point, we hypothesized that muscle lipid metabolism alterations related to MSTN deficiency can impact muscle lipid composition. We observed changes in FA composition, and components of phospholipids, in deficient Mstn muscles, concomitant with an increase in SFAs at the expense of MUFAs. This was accompanied by a down-regulation of SCD1. Of note, the promoter of Scd1 gene contains a Ppar response element sequence [55]; therefore the down-regulation of Ppar that we have seen in absence of mstn could result in decreased gene transcription of Scd1. As Scd1 converts SFA to MUFA, the down-regulation of this enzyme supports a shift away MUFA towards SFA. Biosynthesis PUFAs of the n-6 and n-3 series occurs via sequential desaturation, elongation and partial degradation step [56].
In our work, the impact of mstn deficiency on increased -5 and -6 desaturase activity is unlikely, as we noticed no significant change in -5 desaturase activity between mice groups. Moreover, 22:6n-3 proportion (docosahexaenoic acid), reflecting indirectly -6 desaturase activity, was no modified. The hypothesis of increased elongase and/or -8 desaturase activities in mstn KO muscle cannot be excluded. Finally, as mstn KO muscle exhibited higher significant proportion of 22:5n-3, but not of 22:6n-3, the hypothesis of an inhibited peroxysomal β-oxidation can be also advanced [56].
The lipid tails of the phospholipids composing the plasma membranes can affect mechanical properties. Membrane fluidity is promoted by lipids with short, unsaturated fatty acids [57]. In DMD   [59][60][61][62][63]. Our results suggest both qualitative (reduction of 12% in the proportion) and quantitative (reduction of 23% in quantity) impacts of mstn deficiency on mitochondrial cardiolipin pathway.
If it is difficult to distinguish between a causal role of cardiolipin and a correlative effect, in one study, replenishment with cardiolipin of hypothyroid mitochondria restored the activity of cytochrome oxidase to control levels [60]. Synthesized in the inner mitochondrial membrane, nascent CL undergoes a remodelling process to form the mature CL [24]. Our results on enzymes involved in CL formation indicated that MSTN deficiency likely negatively affected synthesis, rather than remodeling process, that induced the lower abundance of CL. It should be added that we showed in mstn KO gastrocnemius a tendency towards the reduction of linoleic acid (18:2n-6), the most abundant fatty acid in CL [64]. As a close interplay between CL biosynthesis and oxidative phosphorylation is recognized [24], our results raise the question of the link with the compromised bioenergetic capacity observed in this hypertrophic muscle. We recognize that direct evidence is still missing, and should be discussed with caution. However, our results suggest a specific effect on skeletal muscle. Indeed, no impact of mstn deficiency has been observed on mitochondrial CL proportion /content in heart, an organ without functional impairment in this mice model. These results strengthen and highlight correlation between cardiolipin pathway and mitochondrial impairment, as only skeletal muscle from mstn KO mice presents both qualitative and quantitative impacts on mitochondrial membrane cardiolipin pathway with oxidative mitochondrial and functional disorders.
CL plays an essential role in maintaining an adequate membrane potential [65], but arguably, the most important function of CL in mitochondria is the effect it has on the association of protein complexes and supercomplexes [25]. Actually, this specific phospholipid of the mitochondria plays an important role in respiratory chain (biogenesis, stability and activity of the respiratory complexes), in the βoxidation of fatty acids and is involved in the last step of energy conversion, namely the production of ATP [24][25][26]. Crystallographic and nuclear magnetic resonance studies showed the presence of CL, in  rich in 18:2n-6, associated with complex I, III and IV, with the ATPsynthase and the ANT [66,25] and its presence is essential for the functionality structure of these proteins, and thus for their optimal activity ( [25] For Review). For instance, the full transport activity of the ADP/ATP carrier depends on the presence of CL [67]. Inversely, the oligomerization state of ADP/ATP carrier in mitochondria, and its association with other protein complexes are altered in the absence of CL [68].
Interestingly, our previous data highlighted metabolic perturbations exclusively in the respiration of intermyofibrillar (IMF) mitochondria from Mstn-deficient muscle [13]. In diabetes mellitus type 1, cardiolipin content is decreased in the IMF subpopulation, but not in the in the sub-sarcolemmal subpopulation [69,70]. The decrease in CL mitochondria content, demonstrated in our study, could thus lead to the uncoupling in the mitochondria respiration (decay in the respiratory ratio control in the IMF mitochondria) and the decrease in ATP production previously reported in mstn KO muscle [13,71]. Consequently, we propose that the significant decrease in CL in Mstn KO mice, pointing to a defect of IMF, as one causative mechanism for the metabolic phenotype.
Our results reveal alterations in mitochondrial membrane lipids related to MSTN deficiency.
How can we link membrane lipid changes to MSTN? We showed that MSTN deficiency downregulates expression of PPAR transcription factor, which regulates fatty acid transport and oxidation in skeletal muscle [72][73][74][75]. Indeed, genes involved in energy metabolism including Fatp, Pdk4 are induced by a PPAR specific agonist. In addition, Wang et al. showed that the PPARδ-mediated reprogramming of muscle fiber involves the increased expression of genes related to fatty acid oxidation, mitochondrial respiration, oxidative metabolism, and slow-twitch contractile apparatus [73].
This suggests that MSTN might regulate membrane lipid changes via PPAR activity. Since Mstn KO mice have fiber hypertrophy in muscles composed predominantly of fast glycolytic fibers [12,[76], it is possible that the decrease in oxidative metabolism and β-oxidation is an indirect effect and related to the fast-twitch phenotype. In fact, PPARδ is preferentially found in oxidative rather than glycolytic myofibers [73]. Nevertheless, reduction of oxidative metabolism has been demonstrated using postnatal Mstn inhibition, where fiber type composition was not affected [15,69,77]. Furthermore, no difference has been noted in the percentage of CL in mitochondria from muscles with varying oxidative potentials, as well as in the percent of phospholipid head groups or major fatty acid subclasses [70]. However, rates of palmitate oxidation were positively correlated with both the unsaturation index and relative abundance of cardiolipin within mitochondria [69]. Overall, this suggests that the glycolytic shift cannot be the only mechanism to explain oxidative metabolism decline and membrane lipid changes in Mstn KO mice.

Conclusion and perspectives
To conclude, the results of the present study show that MSTN absence significantly alters the mitochondrial lipidome. This abnormality may be related to the impairment in mitochondrial function and the higher susceptibility to fatigue in this hypertrophic model, as lipids and in particular cardiolipin play central roles in optimal mitochondrial function. Future studies that modulate the membrane phosphoslipid composition in mitochondria and muscle in Mstn deficient models will be required for rigorous examination of how the abundance of each class of phospoholipids becomes optimized, and in turn affects mitochondrial respiratory function and oxidative metabolism. In this context, nutritional regimes or endurance training are two potential strategies to induce a remodelling of mitochondrial membrane phospholipid composition [78,79], and thus should be studied in the future to elucidate the link between membrane lipid changes and mstn signalling pathway.

Funding
This study was supported by funds from the Institut