Mitochondrial Oxidative Stress and Mitophagy Activation Contribute to TNF-Dependent Impairment of Myogenesis

Many muscular pathologies are associated with oxidative stress and elevated levels of the tumor necrosis factor (TNF) that cause muscle protein catabolism and impair myogenesis. Myogenesis defects caused by TNF are mediated in part by reactive oxygen species (ROS), including those produced by mitochondria (mitoROS), but the mechanism of their pathological action is not fully understood. We hypothesized that mitoROS act by triggering and enhancing mitophagy, an important tool for remodelling the mitochondrial reticulum during myogenesis. We used three recently developed probes—MitoTracker Orange CM-H2TMRos, mito-QC, and MitoCLox—to study myogenesis in human myoblasts. Induction of myogenesis resulted in a significant increase in mitoROS generation and phospholipid peroxidation in the inner mitochondrial membrane, as well as mitophagy enhancement. Treatment of myoblasts with TNF 24 h before induction of myogenesis resulted in a significant decrease in the myoblast fusion index and myosin heavy chain (MYH2) synthesis. TNF increased the levels of mitoROS, phospholipid peroxidation in the inner mitochondrial membrane and mitophagy at an early stage of differentiation. Trolox and SkQ1 antioxidants partially restored TNF-impaired myogenesis. The general autophagy inducers rapamycin and AICAR, which also stimulate mitophagy, completely blocked myogenesis. The autophagy suppression by the ULK1 inhibitor SBI-0206965 partially restored myogenesis impaired by TNF. Thus, suppression of myogenesis by TNF is associated with a mitoROS-dependent increase in general autophagy and mitophagy.


Introduction
Impaired muscle regeneration and atrophy of muscle fibers are observed in ageing and many chronic muscle disorders (Duchenne muscular dystrophy (DMD), facioscapulohumeral dystrophy (FSHD), cachexia, etc.) [1][2][3][4]. These disorders are accompanied by oxidative stress and a chronic increase in pro-inflammatory cytokines, including the tumor necrosis factor (TNF) [3,4]. An acute, self-limiting physiological inflammatory response is required for muscle repair after injury, while chronic inflammation impairs repair and causes muscle wasting [4]. Ectopic expression of a secreted form of the murine TNF causes cachexia and impaired muscle repair after injury [5].
High doses of TNF impair myogenic differentiation in cultured myoblasts [6][7][8][9]. TNF also induces muscle protein catabolism in mature myotubes [10,11], as well as apoptosis in myoblasts [7] and myotubes [12]. TNF activates several mechanisms leading to inhibition of myogenic differentiation, some of which are dependent on reactive oxygen species (ROS) [9]. ROS-dependent suppression of myogenesis is partly related to NFkB activation [9,13]. TNF also activates other redox-sensitive mechanisms of myogenesis suppression that are insufficiently studied [9,14]. by mTOR inhibitors or AMPK activators may impair myogenic differentiation [46][47][48]. Natural protective mechanisms can limit mitophagy due to a decrease in ROS levels. For example, the transcription coactivator PGC-1α which stimulates mitochondrial biogenesis inhibits mitophagy by stimulating the expression of antioxidant enzymes during myogenesis [45]. Antioxidant treatment improves the proliferation of muscle progenitor cells [49] and their capacity to form myotubes and to regenerate damaged muscles [50][51][52][53][54]. Exogenous antioxidants can be considered as possible therapeutic agents for the prevention of myogenesis dysregulation.
Here, we investigated the mechanisms of mito-ROS-dependent disruption of myogenic differentiation under the influence of TNF. We focused on the study of the mitoROSdependent effect of TNF on myoblasts at the stage of their preparation for fusion. To do this, we added TNF once, 24 h before changing the growth medium to the differentiation medium. Under these conditions, TNF disrupted the myogenesis of immortalized human MB135 myoblasts. TNF further enhanced the generation of mito-ROS and lipid peroxidation of mitochondria and mitophagy on days 0 and 1 of differentiation. The classic antioxidant Trolox and the mitochondria-targeted antioxidant SkQ1 prevented TNF-induced excess mito-ROS, lipid peroxidation, and mitophagy, and partially restored defects in myogenic differentiation. This indicated the key role of mito-ROS in the disruption of myogenesis at its early stage. AMPK/ULK1 signaling plays an important role in the induction of mitophagy during muscle regeneration [55,56]. We have shown that excessive stimulation of AMPK by AICAR leads to disruption of myogenesis, as well as stimulation of autophagy by the mTOR inhibitor rapamycin. At the same time, suppression of AMPK/ULK1 by the SBI 0206965 inhibitor, added once 30 min before TNF, partially restored myogenesis disturbed by this cytokine. This indicated that increased mitophagy may be the cause of TNF-induced impairment of myoblasts' ability to fuse. Thus, we have shown that TNF inhibits the early stage of differentiation by stimulating mitochondrial ROS production, lipid peroxidation of the mitochondrial inner membrane, and excessive mito-ROS-dependent mitophagy.
May-Grunwald Giemsa staining. After 3 days of differentiation, the cells were fixed with 4% PFA for 5 min (Euromedex, Souffelweyersheim, France). The wells with PFA-fixed cells were washed with phosphate-buffered saline (PBS), and stained with 200 µL of May-Grunwald dye for 5 min. Then, 1 mL of PBS was added to the wells without May-Grunwald dye removal and the cells were stained for additional 15 min in the diluted May-Grunwald solution. Afterwards, the wells were washed 3 times with distilled water, stained for 1 h with 1 mL of Giemsa stain (diluted 1:10 in PBS), washed with distilled water again and let dry. All procedures were performed at room temperature. The samples were observed and photographed using the Axio Imager microscope (Zeiss, Oberkochen, Germany). Eight random fields of view were captured for each sample. Fusion Index (FI) was estimated using ImageJ by dividing the number of nuclei inside the myotubes by the total number of nuclei for each sample.
Flow cytometry. For flow cytometry, the cells (non-treated or treated with antioxidants and TNF as described in the "TNF, antioxidants and inhibitors treatment and myoblast differentiation" of Materials and Methods) were incubated with MitoViewGreen probe (100 nM, 30 min), TMRM (200 nM, 15 min), mitoROS indicator MitoTrackerOrange CM-H2TMRos (500 nM, 30 min), or with ratiometric mitochondrial lipid peroxidation indicator MitoCLox (200 nM 4 h) [52,53]. Then, the cells were washed with PBS two times, deattached with Trypsin-EDTA (1/1) solution (PanEco, Russia), centrifuged, and resuspended in 100 uL of PBS (on ice). The flow cytometric measurements were performed on the Amnis ® FlowSight ® Imaging Flow Cytometer (Luminex, Austin, USA) kindly provided by the Moscow State University Development Program PNR5. The Amnis ® FlowSight ® Imaging Flow Cytometer was equipped with a 488 nm laser (60 mW) and an SSC laser (10 mW). Data were analyzed in Amnis ® IDEAS ® . Statistical analyses for MitoTrackerOrange CM-H2TMRos and MitoViewGreen stain results were performed for the medians of fluorescence distributions detected in the 560-595 nm or 480-560 nm channel, respectively. To analyze the oxidation of MitoCLox, the ratio of the medians of fluorescence distributions in the green (480-560 nm) and red (595-640 nm) channels was calculated for each sample. The data in the figures are presented as the means of these medians, normalized to the control.
Western blot analysis. Immunoblotting was performed as described previously [26]. The cells were lysed in buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 50 mM DTT, 0.01% bromophenol blue). Equal amounts of protein were separated onto 6% SDS polyacrylamide gels (for MYH2) or 12% SDS polyacrylamide gels (and for Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)) and then transferred to PVDF membranes (Amersham, Chicago, IL, USA). Membranes were probed with antibodies against Myosin Heavy Chain (MYH2) (R&D Systems, Minneapolis, MN, USA) and then with HRP conjugated Goat Anti-Mouse IgG (Sigma Aldrich, Burlington, MA, USA). To visualize the peroxidase reaction, the SuperSignal West Dura kit (Thermo Fisher Scientific, Waltham, MA, USA) was used in accordance with the manufacturer's protocol. Images were obtained using the ChemiDoc™ MP System (Bio-Rad, Hercules, CA, USA). The obtained images were analyzed using ImageLab software (version 5.2.1, Bio-Rad, USA). All data are normalized to the expression level of a housekeeping protein (GAPDH).
Mitophagy detection. For mitophagy analysis, cells seeded in confocal dishes were transduced with a lentiviral vector containing mitoQC construction [58]. Mito-QC consists of a tandem mCherry-GFP tag located on the outer mitochondrial membrane. The GFP signal (green) is quenched upon delivery of mitochondria to lysosomes, while the mCherry signal (red) is preserved in lysosomes, allowing monitoring and quantification of mitophagy. Transduction was performed the next day after cell plating and 12-15 h before the addition of antioxidants (see "TNF, antioxidants and inhibitors treatment and myoblast differentiation" in "Materials and Methods"). Living cell fluorescent microscopy was performed 3 or more days after transduction. At least 8 random pictures were captured for each sample, measurements were performed for 60 cells on average in each experiment; 3-5 experiments were performed for different points. Mitophagy level was evaluated for fluorescent (effectively transfected) cells by counting the % of cells that contained mitolysosomes. The analysis was performed manually on merged composite images using ImageJ. The images were captured with the same exposition for each channel in each repeat. Statistical analysis. Statistical analyses were performed in GraphPad Prism 9 using ANOVA test with Dunnet's correction for multiple comparisons or Mann-Whitney U test. Data are presented as mean ± SD. p-values less than 0.05 (*), 0.01 (**), 0.001 (***) and 0.0001 (****) were considered significant. All experiments were performed in no less than 3 biological replicates. The number of biological replicates in each experiment (n) is denoted in corresponding figure legends.

TNF Suppresses Myogenic Differentiation of MB135 Myoblasts and Stimulates Mitochondrial Oxidative Processes and Mitophagy
Pretreatment of MB135 myoblasts with 50 ng/mL TNF 24 h before induction of myogenesis resulted in a strong decrease in myoblast fusion and myotubes formation ( Figure 1A,B), as well as inhibition of myosin heavy chain protein (MYH2), a myotube marker, observed on day 3 of differentiation ( Figure 1C-E). At the same time, the addition of TNF led neither to a significant change in the number of myoblasts/myotubes nuclei per field ( Figure 1A,E) nor to the appearance of apoptotic cells. Moreover, TNF did not decrease the mitochondrial membrane potential ( Figure A1). Thus, the suppression of myogenesis could not be due to the toxic or mitotoxic effect of TNF. Data are presented as % relative to the untreated control (mean ± SD, ***-p < 0.001, ****-p < 0.0001, n ≥ 3, ns: no significant difference).
TNF additionally increased the level of mitoROS both before (day 0) and after (day 1) the induction of differentiation (Figure 2A,B). Mitochondrial lipid peroxidation ( Figure 2C,D) and the number of cells containing mitolysosomes (ML) were also increased by TNF on day 1 after induction of differentiation. TNF also induced an additional decrease in the number of mitochondria in myoblasts ( Figure 2G,H), which Myogenic differentiation of human MB135 myoblasts on day 1 after induction of differentiation was accompanied by a significant increase in the level of mitoROS, measured by MitoTrackerOrange CM-H2TMRos (Figure 2A,B), and an increase in lipid peroxidation of the mitochondrial inner membrane, measured by a MitoCLox ratiometric fluorescent probe [59,60] (Figure 2C,D). Simultaneously, an increase in mitophagy, measured using a fluorescent mito-QC construct, and a decrease in the number of mitochondria, measured using MitoViewGreen, were detected ( Figure 2E-H). All these effects have been previously described in other myoblast cultures and seem to be a part of the mitochondrial reticulum renewal program during myogenic differentiation [15][16][17]41,43,44].
TNF additionally increased the level of mitoROS both before (day 0) and after (day 1) the induction of differentiation (Figure 2A,B). Mitochondrial lipid peroxidation ( Figure 2C,D) and the number of cells containing mitolysosomes (ML) were also increased by TNF on day 1 after induction of differentiation. TNF also induced an additional decrease in the number of mitochondria in myoblasts ( Figure 2G,H), which indicated an additional stimulation of mitophagy.

Antioxidants Partially Restore Differentiation Impaired by TNF
The antioxidants Trolox (100 mkM, a water-soluble analogue of vitamin E) and SkQ1 (40 nM, the mitochondria-targeted antioxidant, which consists of plastoquinone residue conjugated to the penetrating decyltriphenylphosphonium cation [61]), added first 48 h before TNF and then again in the differentiation medium, significantly reduced both the level of mitochondrial lipid peroxidation (Figure 3A-D) and the level of mitoROS ( Figure 3E,F) in both control and TNF-treated cells.
Trolox and SkQ1 added first 48 h before TNF and then again in the differentiation medium partially restored the fusion defects ( Figure 4A,B) and MYH2 expression ( Figure 4C-E) affected by this cytokine. However, the addition of both antioxidants only at the stage of differentiation was ineffective ( Figure A2). In addition, these antioxidants did not restore myogenesis when TNF was added to the differentiation medium ( Figure A3).

Antioxidants Restore TNF-Impaired Myogenesis, in Part by Suppressing Excessive Mitophagy
The excess of cells containing ML in the MB135 cell population treated with TNF was decreased by Trolox and SkQ1 antioxidants ( Figure 5). The effect of antioxidants was statistically significant on day 1 after the induction, suggesting that restoration of TNF-impaired myogenesis by antioxidants may be at least partially related to suppression of excessive mitophagy.
To test this possibility, we analyzed the effects of autophagy inducers as well as an autophagy inhibitor on myogenic differentiation of MB135 myoblasts. Autophagy inducers rapamycin (mTORC1 inhibitor, 100 nM) and AICAR (AMPK activator, 0.5 µM) added 24 h before differentiation induction caused a decrease in myoblast fusion and MYH2 expression on day 3 of differentiation ( Figure 6A-C). On the contrary, the suppression of autophagy and mitophagy by the ULK1 inhibitor (5 µM) added 30 min before TNF led to a partial restoration of TNF-impaired myogenesis and did not affect normal differentiation ( Figure 6D-F). These data confirm a possible role of excessive mitophagy in the antimyogenic effect of TNF. It is important to note that the suppression of autophagy during differentiation (when SBI-0206965 was added to differentiation medium (MDM)) led to the suppression of MYH2 expression ( Figure 6E,F), which indicated an inhibition of differentiation. . Data are presented as % relative to the untreated control at day 0 (mean ± SD, *-p < 0.05, **-p < 0.01, ****-p < 0.0001, n ≥ 3, ns: no significant difference).

Antioxidants Partially Restore Differentiation Impaired by TNF
The antioxidants Trolox (100 mkM, a water-soluble analogue of vitamin E) and SkQ1 (40 nM, the mitochondria-targeted antioxidant, which consists of plastoquinone residue conjugated to the penetrating decyltriphenylphosphonium cation [61]), added first 48 h before TNF and then again in the differentiation medium, significantly reduced both the level of mitochondrial lipid peroxidation ( Figure 3A-D) and the level of mitoROS ( Figure 3E,F) in both control and TNF-treated cells. medium partially restored the fusion defects ( Figure 4A,B) and MYH2 expression ( Figure 4C-E) affected by this cytokine. However, the addition of both antioxidants only at the stage of differentiation was ineffective (Appendix Figure A2). In addition, these antioxidants did not restore myogenesis when TNF was added to the differentiation medium (Appendix Figure A3). (D,E) Densitometric analysis of Western blots for MYH2 expression. Data are presented as % relative to the untreated control in (B,D) or as % relative to TNF in (E) (mean ± SD, *-p < 0.05, **p < 0.01, ****-p < 0.0001, n ≥ 3, ns: no significant difference).

Antioxidants Restore TNF-Impaired Myogenesis, in Part by Suppressing Excessive Mitophagy
The excess of cells containing ML in the MB135 cell population treated with TNF was decreased by Trolox and SkQ1 antioxidants ( Figure 5). The effect of antioxidants was statistically significant on day 1 after the induction, suggesting that restoration of Data are presented as % relative to the untreated control in (B,D) or as % relative to TNF in (E) (mean ± SD, *-p < 0.05, **-p < 0.01, ****-p < 0.0001, n ≥ 3, ns: no significant difference). normal differentiation (Figure 6D-F). These data confirm a possible role of excessive mitophagy in the antimyogenic effect of TNF. It is important to note that the suppression of autophagy during differentiation (when SBI-0206965 was added to differentiation medium (MDM)) led to the suppression of MYH2 expression ( Figure 6E,F), which indicated an inhibition of differentiation.  microscopy images. Data are presented as mean ± SD, *-p < 0.05, ***-p < 0.001, n ≥ 3, ns: no significant difference. All the data are presented as mean ± SD, **-p < 0.01, ****-p < 0.0001, n ≥ 3, ns: no significant difference. Data are presented as % relative to untreated control in (C,F) or as % relative to TNF in (G).

Discussion
Excessive or chronic inflammation is a common feature of various pathologies characterized by the loss of muscle mass. Proinflammatory cytokines such as TNF initiate intracellular signaling pathways leading to protein breakdown and muscle atrophy. These cytokines can also significantly impair myogenic differentiation essential for skeletal muscle regeneration [6][7][8][9]. The mechanisms of their damaging effects are not fully understood.
In vitro experiments allow to separately analyze the effects of cytokines at the stages preceding differentiation and upon differentiation. Here, we investigated the effects of TNF on the myogenesis of human MB 135 myoblasts. TNF inhibited myogenesis both when added at the onset of differentiation and 24 h before the differentiation induction, and when removed, when the growth medium was changed to the differentiation medium. Thus, TNF-treated cells could not complete their myogenic commitment. A similar inhibition of myogenic commitment by TNF in mouse C2C12 myoblasts is associated with prevention of cell cycle exit required for further differentiation [8]. In our model, the addition of TNF to subconfluent MB135 myoblasts monolayer 24 h prior to differentiation induction blocked differentiation (myotube formation and MYH2 expression) without a significant effect on cell proliferation.
The production of mitochondrial ROS accompanies myogenic differentiation and presumably correlates with the metabolic switch from glycolysis to oxidative metabolism [62]. At the same time, mitochondrial dysfunction associated with excessive production of ROS is characteristic of various muscle pathologies [3,17,23]. Using the MitoTracker Orange CM-H2TMRos (a nonfluorescent form of MitoTracker Orange emitting fluorescence when oxidized in the mitochondrial matrix), we demonstrated that myogenic differentiation of MB135 myoblasts was accompanied by an increased production of mitoROS at early stages, which is consistent with the data obtained earlier in myoblast cell lines [18,21,22]. We also analyzed the oxidation of mitochondrial inner membrane phospholipids using MitoCLox, a novel mitochondria-targeted ratiometric fluorescent probe [59,60]. Previously, we observed a fraction of myoblasts with significantly oxidized MitoCLox in MB135 cells [59]. This fraction increased in high-density cultures, indicating that mitochondrial lipid peroxidation was associated with the commitment state of myoblasts. We have now observed a significant increase in mitochondrial lipid peroxidation 24 h after induction of differentiation. TNF slightly but statistically significantly increased both the level of mitoROS and the level of mitochondrial lipid peroxidation after differentiation induction. An additional increase in mitoROS levels induced by TNF was also observed before differentiation induction (day 0).
To study the possible role of mitochondrial ROS production in TNF-dependent myogenesis impairment, we used the mitochondria-targeted antioxidant SkQ1 [61]. SkQ1 effectively scavenged mitoROS, thereby preventing oxidation of MitoTracker Orange CM-H2TMRos and mitochondrial lipid peroxidation. The untargeted antioxidant Trolox also prevented these oxidative events but at much higher concentrations. Both antioxidants prevented the impairment of myogenic differentiation by TNF when added at the proliferation stage 48 h before TNF and then into the differentiation medium. Adding antioxidants only at the stage of differentiation was ineffective. This is consistent with the fact that antioxidants were not able to prevent the effect of TNF added to the differentiation medium. Myoblasts are thus most sensitive to antioxidant treatment at the stages of proliferation/preparation for fusion. We previously observed similar effects of SkQ1 and Trolox in a cell model of FHSD, where oxidative stress and impaired myogenesis were caused by low-level expression of DUX4 in MB135 myoblasts [31]. Both models show that the stage of preparation of myoblasts for fusion (myogenic commitment) is very sensitive to the excessive production of mitoROS.
Increased generation of mito-ROS by dysfunctional mitochondria is characteristic of muscle pathologies [15][16][17]. Excess mito-ROS can stimulate the generation of ROS by cytoplasmic systems. For example, NOX2 activation in neutrophils and endothelium depends on mito-ROS [34][35][36][37]. In turn, cytoplasmic ROS are capable of causing mitochondrial dysfunctions and excessive generation of mito-ROS [38]. The antioxidant Trolox is able to directly remove both mito-ROS and cyto-ROS, while SkQ1 is directly able to remove only mito-ROS. The fact that both antioxidants similarly prevent TNF-induced defects in differentiation points to mito-ROS as the main link in the vicious loop of excessive ROS generation leading to oxidative stress and impaired differentiation.
Antioxidant Tempol, but not SkQ1, reduces the thickness of the formed myotubes in MB135 culture [31]. We have now shown that neither SkQ1 nor Trolox affected myoblast fusion and expression of MYH2. In general, published data on the effect of antioxidants on normal myogenesis are ambiguous, which obviously reflects the complexity and inconsistency of the role of ROS in this process, as well as underestimation of the endogenous antioxidant response of differentiating myoblasts [17,19].
Mitochondrial ROS are able to stimulate selective autophagy of mitochondria (mitophagy), which is often considered as a mechanism to prevent the excessive generation of mitoROS by dysfunctional mitochondria [42,63]. We observed a significant increase in mitophagy 24 h after the induction of MB135 myoblasts differentiation. These observations are consistent with the results of earlier studies on other cultures of differentiating myoblasts [41]. We also observed a stimulation of mitophagy by TNF before (day 0) and, to a lesser extent, after (day 1) induction of differentiation in our model. Stimulation of general autophagy by TNF in mouse C2C12 myoblasts is accompanied by mitochondrial depolarization, ROS generation, and apoptosis [64]. However, in our model, depolarized mitochondria were not observed, probably due to the induction of efficient mitophagy that eliminated dysfunctional mitochondria. Trolox and SkQ1 added 24 h before TNF and then again after induction of differentiation suppressed mitophagy without affecting mitophagy in the absence of TNF, while they reduced mitoROS production and mitochondrial phospholipid peroxidation under the same conditions. Thus, an increased mitoROS production is not essential for mitophagy during normal myogenesis, while TNF-induced excessive mitophagy is largely dependent on mitoROS production in MB135 myoblasts.
Mitochondria removal during muscle injury and after intense exercise is regulated by AMPK/ULK1 signaling [55,56]. This signaling also regulates autophagy in general. Excessive stimulation of AMPK by AICAR led to disruption of myogenesis, as well as stimulation of autophagy by inhibition of mTOR by rapamycin; this is consistent with the previously published data [46][47][48]. Suppression of AMPK/ULK1 with an SBA 0206965 (added to the growth medium 30 min before TNF) partially restored the differentiation defects caused by TNF, without affecting normal myogenesis. These data are consistent with a possible role of excessive mitophagy in TNF-induced disruption of myogenic differentiation. However, it should be taken into account that the substances we used (AICAR, rapamycin, SBA 0206965) also affect the general autophagy, which is involved in the regulation of myogenesis, and also has other side effects. It should also be noted that the addition of SBI 0206965 to the differentiation medium resulted in the blocking of myogenesis, which confirmed the important role of autophagy and mitophagy in myogenesis [40,43,44].
Overall, our results indicate that impairment of myogenesis by TNF is mediated by mitoROS-dependent excessive mitophagy that prevents myoblasts from completing their myogenic commitment/preparation for fusion. The delicate balance between mitophagy induction and prevention of mitophagy overstimulation is important for normal myogenesis.
From a practical point of view, myogenesis is strongly dependent on autophagy; thus, the use of autophagy inhibitors to correct defects in myogenesis associated with excessive mitophagy seems highly questionable. At the same time, the mitochondria-targeted antioxidant SkQ1 does not affect normal mitophagy and myogenesis, but only suppresses excessive mitophagy and thus stimulates myogenesis impaired by the TNF inflammatory cytokine. SkQ1 and similar antioxidants may be potentially useful for the complex therapy of inflammatory muscle pathologies associated with impaired myogenic differentiation.