The TRPV1 Receptor Is Up-Regulated by Sphingosine 1-Phosphate and Is Implicated in the Anandamide-Dependent Regulation of Mitochondrial Activity in C2C12 Myoblasts

The sphingosine 1-phosphate (S1P) and endocannabinoid (ECS) systems comprehend bioactive lipids widely involved in the regulation of similar biological processes. Interactions between S1P and ECS have not been so far investigated in skeletal muscle, where both systems are active. Here, we used murine C2C12 myoblasts to investigate the effects of S1P on ECS elements by qRT-PCR, Western blotting and UHPLC-MS. In addition, the modulation of the mitochondrial membrane potential (ΔΨm), by JC-1 and Mitotracker Red CMX-Ros fluorescent dyes, as well as levels of protein controlling mitochondrial function, along with the oxygen consumption were assessed, by Western blotting and respirometry, respectively, after cell treatment with methanandamide (mAEA) and in the presence of S1P or antagonists to endocannabinoid-binding receptors. S1P induced a significant increase in TRPV1 expression both at mRNA and protein level, while it reduced the protein content of CB2. A dose-dependent effect of mAEA on ΔΨm, mediated by TRPV1, was evidenced; in particular, low doses were responsible for increased ΔΨm, whereas a high dose negatively modulated ΔΨm and cell survival. Moreover, mAEA-induced hyperpolarization was counteracted by S1P. These findings open new dimension to S1P and endocannabinoids cross-talk in skeletal muscle, identifying TRPV1 as a pivotal target.

Once generated, S1P can be degraded by S1P lyase (SPL) [3,4], or alternatively recycled into ceramide through S1P phosphatases (SPP) and ceramide synthases. In addition, S1P can be released outside the cell via unspecific transporters, such as ATP-binding cassette transporters and the specific Spinster homolog 2 [5], and MFSD2B [6] carriers. All these metabolic pathways are necessary to maintain the sphingolipid biostat, crucial for a wide range of physiological processes, including determination of cell fate [7]. Once exported outside the cell, S1P can bind to five different G-protein coupled receptors, called in skeletal muscle C2C12 cells, and that TRPV1 transmits a dual action of mAEA depending on its concentration: at high dose (10 μM) mAEA decreases cell survival and inhibits mitochondrial activity, whereas at low doses (<5 μM) it increases mitochondrial membrane potential and induces the expression of peroxisome proliferator-activated receptor gamma coactivator 1-α (PGC1α), a transcriptional coactivator of genes implicated in the regulation of mitochondrial biogenesis and function [41]. Overall, this study discloses a new modulatory role of S1P in skeletal-muscle cells, which engages a distinct ECS element and supports the concept that different bioactive lipids can indeed cross-talk with each other.

Effect of S1P on the Expression of ECS Elements
The mRNA expression of eCB-binding receptors (CB1, CB2, GPR55, TRPV1) and AEA and 2-AG metabolic enzymes (NAPE-PLD, FAAH, DAGLα, DAGLβ, and MAGL) was evaluated by qRT-PCR in C2C12 cells treated for 24 and 48 h with S1P at 1 μM, a dose that was shown to be effective to induce myogenic differentiation [32]. Upon exposure to S1P for 24 h, CB1 and CB2 mRNA expression showed a trend towards increase, yet not statistically significant, whereas GPR55 and TRPV1 mRNA levels were significantly enhanced (p < 0.01 vs. control) (Figure 1a). FAAH, DAGLα and DAGLβ were also significantly increased (p < 0.01 vs. control, for FAAH and DAGLβ; p < 0.05 vs. control, for DAGLα), unlike NAPE-PLD and MAGL that remained unaffected (Figure 1b). Such an increase in selected elements of the ECS returned to control levels when S1P exposure was prolonged for additional 24 h (Figure 2a,b), with the exception of GPR55 gene expression, which was significantly reduced (p < 0.01 vs. control) (Figure 2a). Such an increase in selected elements of the ECS returned to control levels when S1P exposure was prolonged for additional 24 h (Figure 2a Then, the protein expression of eCB-binding receptors, AEA and 2-AG metabolic enzymes was evaluated by Western blot in C2C12 cells treated with 1 μM S1P for 24 h. In treated cells CB2 protein level was significantly decreased (p < 0.059) compared to controls (Figure 3a), whereas TRPV1 protein level was significantly increased (p < 0.05 vs. control), in keeping with qRT-PCR data (Figure 1a). Remarkably, protein expression of any other Then, the protein expression of eCB-binding receptors, AEA and 2-AG metabolic enzymes was evaluated by Western blot in C2C12 cells treated with 1 µM S1P for 24 h. In treated cells CB 2 protein level was significantly decreased (p < 0.059) compared to controls (Figure 3a), whereas TRPV1 protein level was significantly increased (p < 0.05 vs. control), in keeping with qRT-PCR data ( Figure 1a). Remarkably, protein expression of any other ECS element was unaffected (Figure 3b-g), while CB 1 and DAGLα were below the detection limit.  Densitometric analysis values are expressed as relative optical density and normalized to β-Actin. The values represent the mean ± SEM of three independent experiments (n = 3). Statistical analysis was performed by ONE-WAY ANOVA test followed by Bonferroni post hoc test. [* p < 0.05 vs. control cells (CTRL)]. "+" with S1P, "−" without S1P.
To further investigate the possible effect of S1P on the endogenous levels of AEA and 2-AG, LC/MS analysis was performed. In keeping with protein expression of metabolic enzymes, no significant differences in the levels of these two eCBs were observed in cells treated with 1 μM S1P for 24 h (Figure 4). Densitometric analysis values are expressed as relative optical density and normalized to β-Actin. The values represent the mean ± SEM of three independent experiments (n = 3). Statistical analysis was performed by ONE-WAY ANOVA test followed by Bonferroni post hoc test. [* p < 0.05 vs. control cells (CTRL)]. "+" with S1P, "−" without S1P.
To further investigate the possible effect of S1P on the endogenous levels of AEA and 2-AG, LC/MS analysis was performed. In keeping with protein expression of metabolic enzymes, no significant differences in the levels of these two eCBs were observed in cells treated with 1 µM S1P for 24 h (Figure 4).

Effect of Methanandamide on Mitochondrial Membrane potential
As shown by gene and protein expression data, C2C12 cells did express FAAH, the main hydrolytic enzyme of AEA (Figures 1-3). Therefore, to rule out any interference of this enzyme activity, for subsequent cell treatments the non-hydrolysable, stable analogue of AEA methanandamide (mAEA) [40] was used. In particular, the mitochondrial membrane potential (ΔΨm) was evaluated in C2C12 cells treated with mAEA at different doses (2.5-5.0-10 μM) for 24 h, because of the key role of AEA on energy homeostasis [42,43]. ΔΨm values were measured with JC-1, a fluorescent dye that forms aggregates in the mitochondria with a red emission (~590 nm) but reverts to monomers with a green emission (~526 nm) when the mitochondrial membrane starts to depolarize [44]. Treatment for 24 h at lower doses of mAEA (2.5 μM and 5.0 μM) significantly (p < 0.01 and p < 0.05, respectively) increased ΔΨm values compared to controls ( Figure 5). Yet, at the highest dose of 10 μM mAEA significantly (p < 0.05) decreased the ratio of aggregated to monomeric form of JC-1, by 48 ± 5.9%, compared to untreated controls ( Figure 5). The depolarization induced by 10 μM mAEA was further investigated in the presence of S1P, and interestingly mAEA failed to affect ΔΨm when cells were incubated in the presence of S1P, and vice versa S1P was unable to alter ΔΨm in the absence of mAEA ( Figure 6). Taken together, these data suggested a possible crosstalk between the two lipid signaling systems. Moreover, to interrogate whether the mAEA depolarizing effect was mediated by eCB-binding receptors, the analysis was performed also in the presence of

Effect of Methanandamide on Mitochondrial Membrane potential
As shown by gene and protein expression data, C2C12 cells did express FAAH, the main hydrolytic enzyme of AEA (Figures 1-3). Therefore, to rule out any interference of this enzyme activity, for subsequent cell treatments the non-hydrolysable, stable analogue of AEA methanandamide (mAEA) [40] was used. In particular, the mitochondrial membrane potential (∆Ψm) was evaluated in C2C12 cells treated with mAEA at different doses (2.5-5.0-10 µM) for 24 h, because of the key role of AEA on energy homeostasis [42,43]. ∆Ψm values were measured with JC-1, a fluorescent dye that forms aggregates in the mitochondria with a red emission (~590 nm) but reverts to monomers with a green emission (~526 nm) when the mitochondrial membrane starts to depolarize [44]. Treatment for 24 h at lower doses of mAEA (2.5 µM and 5.0 µM) significantly (p < 0.01 and p < 0.05, respectively) increased ∆Ψm values compared to controls ( Figure 5). Yet, at the highest dose of 10 µM mAEA significantly (p < 0.05) decreased the ratio of aggregated to monomeric form of JC-1, by 48 ± 5.9%, compared to untreated controls ( Figure 5).

Effect of Methanandamide on Mitochondrial Membrane potential
As shown by gene and protein expression data, C2C12 cells did express FAAH, the main hydrolytic enzyme of AEA (Figures 1-3). Therefore, to rule out any interference of this enzyme activity, for subsequent cell treatments the non-hydrolysable, stable analogue of AEA methanandamide (mAEA) [40] was used. In particular, the mitochondrial membrane potential (ΔΨm) was evaluated in C2C12 cells treated with mAEA at different doses (2.5-5.0-10 μM) for 24 h, because of the key role of AEA on energy homeostasis [42,43]. ΔΨm values were measured with JC-1, a fluorescent dye that forms aggregates in the mitochondria with a red emission (~590 nm) but reverts to monomers with a green emission (~526 nm) when the mitochondrial membrane starts to depolarize [44]. Treatment for 24 h at lower doses of mAEA (2.5 μM and 5.0 μM) significantly (p < 0.01 and p < 0.05, respectively) increased ΔΨm values compared to controls ( Figure 5). Yet, at the highest dose of 10 μM mAEA significantly (p < 0.05) decreased the ratio of aggregated to monomeric form of JC-1, by 48 ± 5.9%, compared to untreated controls ( Figure 5). The depolarization induced by 10 μM mAEA was further investigated in the presence of S1P, and interestingly mAEA failed to affect ΔΨm when cells were incubated in the presence of S1P, and vice versa S1P was unable to alter ΔΨm in the absence of mAEA ( Figure 6). Taken together, these data suggested a possible crosstalk between the two lipid The depolarization induced by 10 µM mAEA was further investigated in the presence of S1P, and interestingly mAEA failed to affect ∆Ψm when cells were incubated in the presence of S1P, and vice versa S1P was unable to alter ∆Ψm in the absence of mAEA ( Figure 6). Taken together, these data suggested a possible crosstalk between the two lipid signaling systems. Moreover, to interrogate whether the mAEA depolarizing effect was mediated by eCB-binding receptors, the analysis was performed also in the presence of their selective antagonists: 0.1 µM SR1 for CB 1 [45], 0.1 µM SR2 for CB 2 [45], 0.1 µM ML193 for GPR55 [46], and 0.1 µM I-RTX for TRPV1 [47,48]. None of the antagonists affected ∆Ψm when applied alone, yet pre-treatment with the specific antagonists of eCB-binding receptors prevented the depolarizing activity of 10 µM mAEA, whereas in the presence of I-RTX, selective TRPV1 antagonist, mAEA increased ∆Ψm (#### p < 0.0001 vs mAEA alone) ( Figure 6).  [45], 0.1 μM SR2 for CB2 [45], 0.1 μM ML193 for GPR55 [46], and 0.1 μM I-RTX for TRPV1 [47,48]. None of the antagonists affected ΔΨm when applied alone, yet pre-treatment with the specific antagonists of eCB-binding receptors prevented the depolarizing activity of 10 μM mAEA, whereas in the presence of I-RTX, selective TRPV1 antagonist, mAEA increased ΔΨm (#### p < 0.0001 vs mAEA alone) ( Figure 6). Since 10 μM mAEA markedly depolarized the mitochondrial membrane, we checked whether it could impair C2C12 cell viability. When added for 24 h, mAEA showed a dosedependent cytotoxicity, and at 10 μM, it significantly reduced cell viability by 30 ± 3.0% ( Figure 7). In order to further analyze the effect of different concentrations of mAEA on mitochondrial activity, ΔΨm was measured in growing and serum-starved C2C12 myoblasts upon treatment for 24 h with mAEA at 5 μM, the dose immediately below the toxic one. Since 10 µM mAEA markedly depolarized the mitochondrial membrane, we checked whether it could impair C2C12 cell viability. When added for 24 h, mAEA showed a dose-dependent cytotoxicity, and at 10 µM, it significantly reduced cell viability by 30 ± 3.0% ( Figure 7).  [45], 0.1 μM SR2 for CB2 [45], 0.1 μM ML193 for GPR55 [46], and 0.1 μM I-RTX for TRPV1 [47,48]. None of the antagonists affected ΔΨm when applied alone, yet pre-treatment with the specific antagonists of eCB-binding receptors prevented the depolarizing activity of 10 μM mAEA, whereas in the presence of I-RTX, selective TRPV1 antagonist, mAEA increased ΔΨm (#### p < 0.0001 vs mAEA alone) ( Figure 6). Since 10 μM mAEA markedly depolarized the mitochondrial membrane, we checked whether it could impair C2C12 cell viability. When added for 24 h, mAEA showed a dosedependent cytotoxicity, and at 10 μM, it significantly reduced cell viability by 30 ± 3.0% ( Figure 7). In order to further analyze the effect of different concentrations of mAEA on mitochondrial activity, ΔΨm was measured in growing and serum-starved C2C12 myoblasts upon treatment for 24 h with mAEA at 5 μM, the dose immediately below the toxic one. For these assays, the Mitotracker Red CMX-Ros fluorescent dye was used, which stains In order to further analyze the effect of different concentrations of mAEA on mitochondrial activity, ∆Ψm was measured in growing and serum-starved C2C12 myoblasts upon treatment for 24 h with mAEA at 5 µM, the dose immediately below the toxic one. For these assays, the Mitotracker Red CMX-Ros fluorescent dye was used, which stains mitochondria proportionally to their membrane potential. Laser-scanning confocal microscopy allowed the detection of fluorescence intensity, which was normalized to the cell number in each field. As shown in Figure 8, in 24 h serum starved C2C12 myoblasts ∆Ψm significantly increased by 77.7 ± 7.5% compared to growing myoblasts (**** p < 0.0001), and it was further increased by 33.5 ± 2.9% (*** p < 0.001) upon challenge with 5 µM mAEA. mitochondria proportionally to their membrane potential. Laser-scanning confocal microscopy allowed the detection of fluorescence intensity, which was normalized to the cell number in each field. As shown in Figure 8, in 24 h serum starved C2C12 myoblasts ΔΨm significantly increased by 77.7 ± 7.5% compared to growing myoblasts (**** p < 0.0001), and it was further increased by 33.5 ± 2.9% (*** p < 0.001) upon challenge with 5 μM mAEA.  With the aim of interrogating whether S1P might affect mAEA activity on ∆Ψm, the 5 µM mAEA challenge was performed also in the presence of 1 µM S1P for 24 h. Under these conditions, the mAEA-dependent increase in ∆Ψm was fully reverted by S1P (Figure 9). With the aim of interrogating whether S1P might affect mAEA activity on ΔΨm, the 5 μM mAEA challenge was performed also in the presence of 1 μM S1P for 24 h. Under these conditions, the mAEA-dependent increase in ΔΨm was fully reverted by S1P (Figure 9). Moreover, based on the protein expression data, the role of eCB-binding receptors in 5 μM mAEA-induced increase in ΔΨm was studied. The specific blockade of TRPV1 by I-RTX abrogated the activity of 5 μM mAEA on ΔΨm, whereas SR1 and SR2 were ineffective ( Figure 10). ✱ Figure 9. Effect of S1P on mAEA-induced mitochondrial membrane potential in C2C12 myoblasts. Confluent serum-starved cells were treated with 5 µM mAEA in the presence or absence of 1 µM S1P for 24 h. Fluorescence intensity of Mitotracker Red CMX-Ros (Ex/Em: 579/599) was normalized to cell number and data are reported as mean ± SEM of six field per condition in three independent experiments. Statistical analysis was performed by TWO-WAY ANOVA followed by Bonferroni post-hoc test [* p < 0.05 mAEA vs CTRL, # p < 0.05 S1P+mAEA vs mAEA].
Moreover, based on the protein expression data, the role of eCB-binding receptors in 5 µM mAEA-induced increase in ∆Ψm was studied. The specific blockade of TRPV1 by I-RTX abrogated the activity of 5 µM mAEA on ∆Ψm, whereas SR1 and SR2 were ineffective ( Figure 10).

Effect of Methanandamide on PGC1α, Respiratory Chain Complex Expression and Oxygen Consumption
With the goal to analyze the molecular mechanism involved in the effects of 5 µM mAEA on mitochondrial function, the expression of PGC1α and respiratory chain complexes was measured by Western blot analysis upon treatment for 24 h. As shown in Figure 11, the protein content of PGC1α was significantly enhanced, whereas that of the respiratory chain complexes were not affected by 5 µM mAEA challenge.

Effect of Methanandamide on PGC1α, Respiratory Chain Complex Expression and Oxygen Consumption
With the goal to analyze the molecular mechanism involved in the effects of 5 μM mAEA on mitochondrial function, the expression of PGC1α and respiratory chain complexes was measured by Western blot analysis upon treatment for 24 h. As shown in Figure 11, the protein content of PGC1α was significantly enhanced, whereas that of the respiratory chain complexes were not affected by 5 μM mAEA challenge. Figure 11. Effect of mAEA on PGC1α (a) and respiratory chain complexes' (b) expression. C2C12 myoblasts were treated with mAEA (5 μM) for 24 h. Samples were subjected to SDS-PAGE and Western blot analysis using anti-PGC1α, as well as OXPHOS WB Antibody Cocktail specific for respiratory chain complexes and anti-β-actin antibody as housekeeping. The data were reported after densitometric analysis of the bands as means ± SEM normalized to β-actin of three independent experiments performed in duplicate. The data were reported after densitometric analysis of the bands as means ± SEM normalized to β-actin of three independent experiments. Statistical analysis was performed by analysis of variance using Student t-test [* p < 0.05 vs. control cells (CTRL)].
Moreover, the ability of 5 μM mAEA to affect oxygen consumption was examined by high-resolution respirometry. To this end, C2C12 myoblasts were treated with 5 μM mAEA for 24 h, then were collected and subjected to basal oxygen consumption (ROU-TINE), proton leak (LEAK) and maximal oxygen consumption (E) measurement by the Oroboros-O2K system. At 5 μM, mAEA was unable to affect any of these respiratory parameters in C2C12 myoblasts ( Figure 12). Figure 11. Effect of mAEA on PGC1α (a) and respiratory chain complexes' (b) expression. C2C12 myoblasts were treated with mAEA (5 µM) for 24 h. Samples were subjected to SDS-PAGE and Western blot analysis using anti-PGC1α, as well as OXPHOS WB Antibody Cocktail specific for respiratory chain complexes and anti-β-actin antibody as housekeeping. The data were reported after densitometric analysis of the bands as means ± SEM normalized to β-actin of three independent experiments performed in duplicate. The data were reported after densitometric analysis of the bands as means ± SEM normalized to β-actin of three independent experiments. Statistical analysis was performed by analysis of variance using Student t-test [* p < 0.05 vs. control cells (CTRL)].
Moreover, the ability of 5 µM mAEA to affect oxygen consumption was examined by high-resolution respirometry. To this end, C2C12 myoblasts were treated with 5 µM mAEA for 24 h, then were collected and subjected to basal oxygen consumption (ROUTINE), proton leak (LEAK) and maximal oxygen consumption (E) measurement by the Oroboros-O2K system. At 5 µM, mAEA was unable to affect any of these respiratory parameters in C2C12 myoblasts (Figure 12).

Discussion
eCBs and S1P are bioactive lipids, which are critically important for the regulation of a plethora of key biological processes [8,13]. Intriguingly, these two lipid systems are composed by multiple enzymes that regulate their cellular tone, and specific receptor targets that transmit their effects, via a complex array of signal transduction pathways that remain only partly understood [8,13].
Here, unprecedented evidence is provided for a modulatory role of S1P on selected elements of ECS in cultured murine skeletal-muscle C2C12 cells, which are are widely used as a valuable model for interrogating, at the molecular level, key processes in skeletal muscle.
Importantly, S1P and eCB systems regulate fundamental biological events in skeletal muscle cell biology [10,35]. Indeed, S1P is known to activate muscle resident stem cells, favoring tissue repair by promoting also their proliferation and differentiation [49][50][51]. Moreover, it regulates tissue mechanical properties and responsiveness to insulin [10,52]. Myogenesis is also modulated by ECS, since CB1 expression has been reported to be upregulated during myoblast differentiation and 2-AG was found to act via CB1 as endogenous repressor of myoblast differentiation, while mAEA treatment inhibited myogenesis [29,35]. The key role of the two families of lipid mediators in the control of skeletal-muscle cell biology is further underscored by the evidence that both signaling systems are dysregulated in a major skeletal-muscle disease, namely Duchenne muscular dystrophy [53,54]. In this context, CB1 was found to be overexpressed in murine dystrophic muscles, as well

Discussion
eCBs and S1P are bioactive lipids, which are critically important for the regulation of a plethora of key biological processes [8,13]. Intriguingly, these two lipid systems are composed by multiple enzymes that regulate their cellular tone, and specific receptor targets that transmit their effects, via a complex array of signal transduction pathways that remain only partly understood [8,13].
Here, unprecedented evidence is provided for a modulatory role of S1P on selected elements of ECS in cultured murine skeletal-muscle C2C12 cells, which are are widely used as a valuable model for interrogating, at the molecular level, key processes in skeletal muscle.
Importantly, S1P and eCB systems regulate fundamental biological events in skeletal muscle cell biology [10,35]. Indeed, S1P is known to activate muscle resident stem cells, favoring tissue repair by promoting also their proliferation and differentiation [49][50][51]. Moreover, it regulates tissue mechanical properties and responsiveness to insulin [10,52]. Myogenesis is also modulated by ECS, since CB 1 expression has been reported to be up-regulated during myoblast differentiation and 2-AG was found to act via CB 1 as endogenous repressor of myoblast differentiation, while mAEA treatment inhibited myogenesis [29,35]. The key role of the two families of lipid mediators in the control of skeletal-muscle cell biology is further underscored by the evidence that both signaling systems are dysregulated in a major skeletal-muscle disease, namely Duchenne muscular dystrophy [53,54]. In this context, CB 1 was found to be overexpressed in murine dystrophic muscles, as well as in muscle biopsies of dystrophic patients [53], while muscles of dystrophic mice exhibited S1P lyase (SPL) upregulation and S1P deficiency, and accordingly pharmacological blockade of SPL in dystrophic mouse muscles exerted beneficial effects mediated by enhanced S1P, which in turn binds to and inhibits histone deacetylase [54].
In the present study, the possible functional interplay between the S1P and eCB systems has been explored in murine myoblasts by examining, as a first approach whether S1P treatment can modify receptor-mediated responsiveness to eCBs or the endogenous metabolism of these lipids. To this end, the expression of individual receptors and eCBmetabolic enzymes upon S1P challenge has been investigated both at mRNA and protein levels, and AEA and 2-AG cell content has been quantified. Notably, S1P at 24 h was found responsible of the selective up-regulation of TRPV1. Conversely, the protein content of CB 2 was reduced by S1P, even though this variation was not mirrored by mRNA levels. In addition, mRNA levels of GPR55 receptor, as well as FAAH and DAGLβ enzymes in S1P-treated myoblasts were transiently increased by S1P at 24 h, and only GPR55 was found to be decreased at 48 h. It seems noteworthy that the present gene expression data on DAGL isozymes are consistent with the notion that non-neuronal cells (such as myoblasts) predominantly express DAGLβ, whereas DAGLα is more abundant within the brain [55]. In this context, it should also be noted that discrepancies between mRNA and protein expression of a specific target in treated cells are not quite unprecedented. Indeed, similar disparities between changes in mRNA abundance and protein content have been already reported [56], also within the ECS, by others [57] and by us [58,59]. Interestingly, an interaction between GPR55 and S1P5 receptor has already been demonstrated in a colon cancer cell line [60]. In particular, S1P5 strongly and specifically interacts with GPR55, and the activation of each receptor led to increased cell proliferation, ERK phosphorylation and cancer-associated gene expression. Conversely, co-activation of both receptors inhibited the above mentioned events, supporting the occurrence of a functional crosstalk [60]. In the herein-investigated model of murine myoblasts, S1P significantly rearranges cell sensitivity to eCBs without changing their endogenous content, suggesting a potential impact on the signal transduction driven by these lipid mediators.
Although important molecular effects of eCBs in skeletal muscle have been identified [53], so far it is not known if they engage the regulation of mitochondrial activity, which is essential for skeletal muscle metabolism and proper tissue plasticity. It should be recalled that AEA is known to play a key role in energy homeostasis [37][38][39], and indeed it can affect the shape and function of isolated mitochondria [42]. In line with this, some reports have even suggested a mitochondrial localization of CB 1 , supporting the view that AEA-dependent effects on energy production are triggered by these particular CB 1 receptors [61,62]. Against this background, here we sought to ascertain whether the stable anandamide analogue mAEA could affect mitochondrial membrane potential, a key functional parameter of these organelles. Interestingly, our findings show that mAEA at the lowest concentration of 5 µM markedly enhanced ∆Ψm, whereas the highest dose of 10 µM exerted an opposite effect. Moreover, in both cases the changes of ∆Ψm elicited by mAEA were found to be mediated by TRPV1, since they were abrogated in the presence of its selective antagonist I-RTX. Likewise, treatment of mitochondria isolated from mice liver with higher concentrations (up to 50 µM) of AEA significantly decreased ∆Ψm, down tõ 70% of untreated controls, and they also increased membrane fluidity [42]. However, these effects were independent of CB 1 , CB 2 or TRPV1 [42]. Similarly, in rat cortical neurons AEA, at the dose of 30 µM, was able to reduce mitochondrial membrane potential, induce the translocation of cytochrome c and activate caspase-3, overall leading to cell death independently of its receptor targets [63]. Instead, in human neuroblastoma SH-SY5Y cells, AEA dose-dependently, and again in the micromolar range, induced a mitochondria-dependent apoptosis, which in these cells was mediated by CB 1 [38]. Overall, there seem to be different mechanisms by which AEA can affect mitochondrial function, possibly also in relation to cell type and species-specificity. Moreover, the ability of S1P to counteract the depolarizing effect of mAEA in C2C12 cells appears to be noteworthy. Yet, S1P alone did not affect the mitochondrial membrane potential, which is in agreement with its ability to act as pro-survival factor and to control mitophagy and mitochondria dynamics [64][65][66]. Altogether, these results add a new dimension to the signaling pathways triggered by S1P and eCBs, and identify the Ca 2+ channel TRPV1 as a pivotal target crucially involved in their crosstalk. The latter observation is quite remarkable, because so far most of the effects of eCBs in murine myoblasts have been ascribed to CB 1 /CB 2 receptors with one exception describing TRPV1 as responsible for Ca 2+ increase and subsequent promyogenic effect induced by cannabidiol [67]. The molecular mechanism by which S1P modulates responsiveness of myoblasts to eCBs has not been here investigated, paying special attention to the biochemical effects transmitted by S1P via eCB system, rather than its mechanistic action. C2C12 myoblasts are known to express four out of five S1P receptor subtypes, namely S1P1, S1P2, S1P3 and S1P4 [68], hence future studies will be necessarily addressed to dissect the exact role exerted by one or more receptor subtypes in the observed effects, even if it is conceivable to hypothesize that S1P2 is implicated. Indeed, this receptor subtype, although less represented than other isoforms, at least at mRNA level, [68], is regarded as the dominant receptor subtype in this cell type, being capable of transmitting key biological effects such as cell differentiation, inhibition of cell motility and cell proliferation [32,69].
Another relevant issue is the possibility that, instead of a cross-talk, mAEA and S1P may interact directly with TRPV1. Indeed, recent evidence supports the direct activation of TRPV1 by S1P in the context of pain and itch [70]. However, the hypothetic agonism of S1P at TRPV1 could not account for the observed up-regulation of TRPV1 at protein and mRNA level elicited by S1P at 24 h. Unfortunately, the analysis of the elevation of intracellular Ca 2+ in transfected cells does not appear a suitable read-out to clarify this issue, because extracellular S1P itself, acting via its specific receptors, may potently induce such an elevation in many cell types, C2C12 myoblasts included [71]. More sophisticated approaches, such as silencing S1P receptor subtypes in C2C12 myoblasts, would be more informative, and should be the subject of independent investigations.
It is tempting to speculate that the herein-observed up-regulation of TRPV1 induced by S1P can counteract the increase in mitochondrial membrane potential induced by low mAEA, causing calcium overload that impairs mitochondrial function, thus mimicking the effect elicited by high dose of mAEA. In this context, it is well known that the large calcium gradient maintained across the mitochondrial inner membrane represents a critical signaling potential for this cation [72]. It is noteworthy that TRPV1 activation by capsaicin has been proven to increase free cytosolic calcium and improve energy metabolism by upregulating PGC1α in C2C12 myotubes, as well as in skeletal muscles [73]. Furthermore, it has been demonstrated that TRPV1 ligands decrease mitochondrial membrane potential and oxygen consumption in isolated mitochondria from rat cardiomyocytes [74]. Notably, in the present study, we found that low doses of mAEA are responsible for the increase in mitochondrial membrane potential and induction of PGC1α expression, without affecting oxygen consumption, whereas the highest dose of mAEA turns the TRPV1 action to negatively modulate mitochondrial membrane polarization and cell survival. The opposite functional changes mediated by TRPV1 may rely on concentration-dependent responses that via subtle increases in intracellular calcium positively affect mitochondrial membrane potential, whereas through mitochondrial calcium overload cause dysregulation of mitochondrial membrane potential and subsequent mitochondrial dysfunction. On a final note, markedly different consequences of low or high doses of AEA are not unprecedented. Indeed, in primary human melanocytes low (<1 µM) AEA leads to melanogenesis, whereas high (>5 µM) AEA causes development of melanoma tumors [75].
Overall, this study reveals a new critical cross-talk between S1P and eCB systems in skeletal-muscle cells, identifying the Ca 2+ channel TRPV1 as a pivotal target and thus opening the avenue to new molecular approaches to control skeletal-muscle disorders characterized by calcium dyshomeostasis.

Materials and Reagents
Dulbecco's modified Eagle's medium, foetal calf serum and penicillin/streptomycin were from Corning (Corning, NY, USA). Bovine serum albumin (BSA) was purchased from Sigma-Aldrich (

Quantitative Real Time-Reverse Transcriptase-Polymerase Chain Reaction (qRT-PCR) Analyses
For the quantification of gene-expression levels of receptors and metabolizing enzymes of eCBs, mRNA was extracted from control and S1P-treated C2C12 cells by using TRIzol, according to the manufacturer's instructions (Life Technologies, Grand Island, NY, USA), and was quantified by using Thermo Scientific NanoDrop 2000 c UV-Vis spectrophotometer at 260 nm (Waltham, MA, USA). Subsequently, 1 µg of total mRNA was retrotranscribed in cDNA by using the RevertAid H Minus First Strand cDNA Synthesis Kit (Life Technologies, Grand Island, NY, USA). SensiFASTTM SYBR Lo-ROX kit was used to assess the relative abundance of CB 1 , CB 2 , GPR55, TRPV1, FAAH, DAGLα, DAGLβ, NAPE-PLD and MAGL on a 7500 Fast Real-time PCR System (Life Technologies, Grand Island, NY, USA), as described previously [31,76]. Primer sequences are reported in Table 1. The relative expression of different amplicons was calculated by the ∆∆Ct method and converted to relative expression ratio 2 (−∆∆Ct) for statistical analysis [77]. All data were normalized to the endogenous reference genes GAPDH and β-Actin.

Western Blotting
Control and S1P-or mAEA-treated cells were lysed in RIPA buffer in the presence of a protease inhibitors cocktail (Sigma-Aldrich, St. Louis, MO). Then, they were sonicated three times and centrifuged at 9500 g for 15 min at 4 • C; supernatants were collected and protein content was determined by the Bio-Rad Protein assay (Bio-Rad Laboratories, Hemel Hempstead, UK). Cell lysates were mixed with Laemli sample buffer (heated for 10 min at 60 • C) so that an equal amount of protein per lane (20-70 µg) were subjected to a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were then electroblotted onto polyvinylidene difluoride (PVDF) membrane (Amersham Hybond, GE Healthcare Life Science, Piscataway, NJ, USA). Subsequently, PVDF membranes were blocked with 5% milk, incubated overnight with the primary antibodies in cold room and then with the appropriate horseradish peroxidase-conjugated secondary antibody (cod.31461, Thermo Fisher Scientific, Waltham, MA, USA) for 1 h at room temperature.  Detection was performed by Clarity Western ECL substrate (Bio-Rad, Hercules, CA, USA) as developed by Azure Biosystems c400 (Sierra Ct, Dublin, CA, USA). Immunoreactive band intensities was quantified by densitometric analysis through the ImageJ software (NIH, Bethesda, MD, USA), as reported [45].

LC/MS Analysis
The lipid fractions from control and S1P-treated C2C12 cells were extracted with chloroform/methanol (2:1 v/v) in the presence of internal standards (1 ng mL−1 AEA-d8, 200 ng mL−1 2-AG-d8). Then, a clean-up step was performed as previously reported [78]. Briefly, the organic phase was dried and then subjected to micro-solid phase extraction (µSPE) for a rapid clean-up, by using OMIX C18 tips from Agilent Technologies (Santa Clara, CA, USA). All analyses were performed on a Nexera LC 20 AD XR UHPLC system (Shimadzu Scientific Instruments, Columbia, MD, USA) with NUCLEODUR ® C18 Isis column from Macherey-Nagel GmbH and Co. (Neumann, Germany), coupled with a 4500 Qtrap mass spectrometer from Sciex (Concord, ON, Canada) equipped with a Turbo V electrospray ionization (ESI) source, operating in positive mode. The levels of AEA and 2-AG were then calculated on the basis of their area ratios with the internal deuterated standard signal areas, and their amounts in pmoles were normalized to the number of cells, as reported [75].
The ∆Ψm of C2C12 myoblasts was also assessed by Mito Tracker Red CMXRos (#M7512) probe (Thermo Fisher Scientific INC, MA, USA) and Laser scanning confocal microscopy after the treatment with mAEA (5 µM), in the presence of selective eCB-binding receptor antagonists, and/or 1 µM S1P for 24 h. Briefly, cells were seeded on microscope slides and grown for 24 h before being shifted in serum-free 0.1% BSA-containing medium in the presence or absence of mAEA. Specific and selective eCB-binding receptor antagonists were used as above described. The probe was diluted in DMEM medium without phenol red at a concentration of 50 nM, incubated for 30 min at 37 • C and 5% CO 2 in the dark, and then slides were fixed in 2% paraformaldehyde. Subsequently, cells were washed twice and incubated with a permeabilization and quenching solution (0.1% Triton X-100 and ethanolamine (1:165) in PBS). A DAPI solution (#MBD0015, Sigma-Aldrich, MA, USA) was used to detect nuclei. Slides were mounted by using Fluoromount Aqueous Mounting Medium (Sigma-Aldrich, MA, USA), and images were obtained in a Leica SP8 laser scanning confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany) using a 40× oil immersion objective.

MTT Assay
Cells were seeded into 96-well plates, at density of 1.5 × 104 per well, and incubated overnight. The day after, cells were incubated with different concentrations of mAEA (2.5-5.0-10 µM) for 24 h. Cell viability was assessed by the mitochondrial-dependent reduction of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT; Sigma, St. Louis, MO, USA) to purple formazan. After 3 h, the MTT solution was discarded and 100 µL of DMSO were added to dissolve the formazan crystals. Cells were analyzed by Enspire multimode plate reader (Perkin Elmer, MA, USA). The cell viability was calculated by subtracting the 630 nm OD background from the 570 nm OD total signal of cell-free blank of each sample, and was expressed as percentage of controls set to 100%.

Intact Myoblast Respiration Using High-Resolution Respirometry
Oxygen consumption was analyzed in 2 mL glass chambers at 37 • C using the Oroboros oxygraph-2K high-resolution respirometer (Oroboros Instruments, Innsbruck, Austria) and substrate, uncoupler, inhibitor, titration (SUIT) protocols, as reported [84]. The oxygen flux normalized on the cell number is calculated as the negative time derivative of the oxygen concentration, measured in sealed chambers, and normalized on the instrumental background (measured in a dedicated experiment before assaying the cells). C2C12 myoblasts treated with 5 µM mAEA or vehicle (ethanol) for 24 h were subjected to respirometry analysis. After instrumental air calibration, 300.000 cells resuspended in DMEM with 1 mg/mL BSA were introduced into the chambers and the basal respiratory activity was measured as routine respiration (R). The LEAK state (L) represents the nonphosphorylating state of uncoupled respiration due to proton leak, proton and electron slip, and cation cycling [85] after the inhibition of ATP synthase by oligomycin administration (5 nM). The capability of the electron transfer system (ETS) was measured by uncoupler titrations using the uncoupler CCCP (1.5 µM/titration steps) as the readout of the maximal capacity of oxygen utilization (E). The residual oxygen consumption (ROX) that remains after the inhibition of ETS was determined by antimycin A (2.5 µM) injection. Data acquisition and analysis were performed using DatLab software (version 7.4, Oroboros Instrument, Innsbruck, Austria), and the oxygen fluxes recorded in the individual titration steps were corrected for ROX.

Statistical Analysis
Data were analysed by the GraphPad Prism 9.3.1 (471) program (GraphPad Software, La Jolla, CA, USA), and were reported as means ± S.E.M of 3 to 5 independent experiments. The statistical analysis was performed through the Student's t-test and one-way or twoway analysis of variance (ANOVA), followed by Bonferroni post-hoc analysis. A level of p < 0.05 was considered statistically significant.