H₂O₂-induced mitochondrial fragmentation in C₂C₁₂ myocytes

Abstract

In skeletal muscle and many other cell types, mitochondria exist as an elaborate and dynamic network in which “individual” mitochondria exist only transiently even under non-stimulated conditions. The balance of continuous mitochondrial fission and fusion defines the morphology of the mitochondrial reticulum. Environmental stimuli, such as oxidative stress, can influence fusion and fission rates, resulting in a transformation of the network's connectivity. Using confocal laser scanning microscopy of C₂C₁₂ mouse myocytes, we show that acute exposure to the reactive oxygen species (ROS) hydrogen peroxide (H₂O₂) induces a slow fragmentation of the mitochondrial reticulum that is reversible over 24 hours. Although H₂O₂ decomposes rapidly in culture medium, the full extent of fragmentation occurs 5-6 hours post-treatment, suggesting that H₂O₂ affects mitochondrial morphology by modulating cellular physiology. Supraphysiological (>1 mM) concentrations of H₂O₂ are cytotoxic, but lower concentrations (250 μM) sufficient to induce transient fragmentation do not lower cell viability. H₂O₂-induced mitochondrial fragmentation is preceded by decreases in inner mitochondrial membrane potential and maximal respiratory rate, suggesting a possible mechanism. Because H₂O₂ is produced in contracting muscle, our results raise the possibility that ROS generation may contribute to exercise-induced changes in mitochondrial morphology in vivo.

Introduction

In a variety of models ranging from yeast to mammalian [1, 2] and human skeletal muscle [3], mitochondria form interconnected networks that undergo continuous morphological rearrangements [4-6]. Each “mitochondrion” exists only transiently before fusing into the network (reticulum) or dividing into smaller organelles (Fig. S1). A suite of conserved dynamin-like GTPases and their interacting partners regulate mitochondrial morphology. Of these, dynamin-related protein 1 (Drp1) is the mediator of mitochondrial fission. Numerous published micrographs show the tangled, web-like mitochondrial networks that result from Drp1 knockdown [7-12]. Shaping the network's morphology in opposition to Drp1 are mitofusin-1 and mitofusin-2 (Mfn1/2). Present as transmembrane proteins in the mitochondrial outer membrane, the mitofusins form homotypic and heterotypic complexes in trans across the apposed outer membranes of the mitochondria that are to fuse, and mediate outer membrane fusion [13]. Another GTPase, optic atrophy 1 (OPA1), is thought to mediate inner membrane fusion. Homozygous loss-of-function of Mfn1, Mfn2, or OPA1 is embryonic lethal in mice [14, 15]; mutations in humans lead to neurological diseases.

Although the phenomenon of mitochondrial dynamics appears to be ubiquitous in mammalian tissues, skeletal muscle fibers present an especially interesting environment in which the dynamic nature of mitochondria remains to be understood. From previous work done in our laboratory as well as in others, mitochondria have been shown to exist as a network in skeletal muscle [1-3]. The specialization of muscle for force production would appear to impose geometric and bioenergetic constraints upon the mitochondrial network and cast doubt on the importance of mitochondrial dynamics in this tissue. However, Cribbs and Strack showed increased Drp1 phosphorylation in rat cardiac muscle after a 15-minute forced swim, suggesting altered mitochondrial dynamics in striated muscle during exercise in vivo [12]. Chen et al. showed that skeletal-muscle-specific deletion of both mitofusins in a conditional knockout mouse leads to muscle atrophy, altered fiber type composition, mitochondrial dysfunction, mtDNA copy loss, and increased accumulation of mtDNA mutations [16]. Clearly mitochondrial dynamics play a critical role in the normal function of muscle.

Reactive oxygen species (ROS) have been shown to induce changes in mitochondrial network morphology in endothelial cells and neurons [17, 18]. Because contracting skeletal muscle is known to release ROS including hydrogen peroxide [19-22], we sought to determine whether H₂O₂ could also alter mitochondrial dynamics in a C₂C₁₂ mouse myocyte model. We found that an acute, non-cytotoxic dose of H₂O₂ causes a delayed fragmentation of the mitochondrial reticulum in myocytes, and that this is preceded by depression of the mitochondrial membrane potential and maximal respiratory rate. Striated muscle mitochondria are known to undergo morphological changes as the result of chronic exercise [23-25]. As well, acute exercise has been observed to affect muscle mitochondrial morphology [26]. Now we propose that oxidative stress may be a key link connecting exercise to these changes.

Experimental procedures

Materials

High-glucose (25 mM) Dulbecco's Modified Eagle's Medium (DMEM; with or without phenol red), penicillin-streptomycin, fetal bovine serum (FBS), horse serum, L-glutamine, and phosphate-buffered saline (PBS) were purchased from Invitrogen/GIBCO (Carlsbad, CA).

Low-density culture of C₂C₁₂ skeletal myocytes

C₂C₁₂ mouse myoblasts were grown at 37°C and 5% CO₂ in DMEM with 10% FBS, 5 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. Myoblasts were seeded in 35-mm dishes at 3000 cells per dish. The medium was replaced with “medium II” (DMEM with 2% horse serum, 5 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin) after 2 days, and experiments were performed 5 days after the switch to medium II. By the day of the experiment, low-density-seeded C₂C₁₂ cells had usually proliferated to 70-80% confluence.

H₂O₂ treatment and staining of mitochondria with MitoTracker Deep Red 633

C₂C₁₂ cells were treated with H₂O₂ (Acros Organics, Fair Lawn, NJ) diluted to the desired concentration in medium II. After treatment, the cells were washed with PBS and incubated in 300 nM MitoTracker Deep Red 633 (Invitrogen/Molecular Probes, Eugene, OR) in medium II without phenol red at 37°C for 30 minutes. The cells were then washed and fresh medium II without phenol red was replaced for live cell imaging.

Live cell imaging using confocal laser scanning microscopy (CLSM)

Single images, typically corresponding to optical sections <1.8 μm thick, were taken of live MitoTracker-stained cells using a Zeiss LSM 510 META confocal scan head mounted on an upright Axioplan 2 microscope with Achroplan IR 63×/0.90 NA or Achroplan 40×/0.80 NA water dipping objectives. Movies of mitochondrial dynamics were made by taking serial images, with a delay of 5 s between scans.

Manual scoring of mitochondrial morphology

3-5 images of random fields of view were taken. All in-focus cells captured in those fields were manually classified as “fragmented” (>50% of mitochondrial area consisting of punctiform mitochondria), “intermediate” (10-50% punctiform), or “reticular” (<10% punctiform). Any myotubes or dead cells present were not scored.

Peroxide assay

The culture supernatant was aspirated off cells grown in 35-mm dishes and replaced with fresh medium II without phenol red, with or without 250 μM H₂O₂. 100 μL of supernatant was removed 0, 5, 10, 15, 30, and 60 minutes after addition of H₂O₂ and immediately added to the assay reagents (QuantiChrom Peroxide Assay Kit; BioAssay Systems, Hayward, CA) in a 96-well plate. Plates were incubated for 30 minutes and read on a SpectraMax 340PC³⁸⁴ microplate spectrophotometer at 585 nm. A standard curve using known concentrations of H₂O₂ was processed at the same time and used to back-calculate the concentrations of H₂O₂ at different time points. The experiment was repeated for 250 μM H₂O₂ in medium II in the absence of cells.

WST-8 cell viability assay

Myocytes were seeded in 96-well plates at low density and grown as usual. After incubation in 0, 125, 250, 500, 1000, or 2000 μM H₂O₂ in medium II for 6 hours, the wells were washed and fresh medium II without phenol red was replaced. WST-8 (Cell Counting Kit-8, Dojindo Labs, Gaithersburg, MD) was added to each well, incubated for 1 hour, and read on the spectrophotometer at 440 nm; the measured absorbance is proportional to the number of viable cells. A second plate was allowed to recover in medium II for 24 hours, after which viability was determined.

Qualitative flow cytometry assay for mitochondrial depolarization

5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1, Sigma-Aldrich) is a cationic, lipophilic dye that accumulates in mitochondria and exhibits green (525 nm) fluorescence in its monomeric state. In energized mitochondria, JC-1 monomers form orange (575 nm) “J-aggregates” [27]. JC-1 was dissolved in DMSO and medium II for a final concentration of 2 μM (0.4% DMSO), which was found to be the minimal concentration at which the majority of cells showed positive 575 nm fluorescence by flow cytometry (Fig. S2). After 30 minutes of incubation at 37°C, cells were washed and trypsinized. Following resuspension in fresh medium II without phenol red, 25,000 events were collected from each sample using a Beckman Coulter EPICS XL flow cytometer. Fluorescence compensation was done such that cells stained with 2 μM JC-1 were positive for both 525 and 575 nm fluorescence, and treatment with 10 μM CCCP increased 525 nm fluorescence while decreasing 575 nm fluorescence. Data was analyzed using FlowJo software (TreeStar). Similar plots using the same JC-1 concentration have been published by Cassart et al. [28].

Polarographic measurement of oxygen consumption

Oxygen consumption rates of intact C₂C₁₂ myocytes were measured using a Clark-type electrode with a digital controller (Rank Brothers), connected to a computer running LabVIEW software (National Instruments). At the start of each trial, 4 mL of medium II was allowed to equilibrate in the sample chamber (maintained at 37°C) until the trace reached a flat line, which was defined as 100% O₂ saturation. C₂C₁₂ myocytes grown in 10-cm dishes were trypsinized, washed, and resuspended in 1 mL of equilibrated medium. One 10-cm dish of cells was used for each trial. After the cells were added to the chamber, the chamber was stoppered and endogenous respiration was recorded for 15 minutes. Then, CCCP (final concentration 10 μM) was injected, and uncoupled respiration was recorded for an additional 10 minutes. After the experiment was stopped, the cells were counted by hemacytometer. The linear least squares method was used to calculate endogenous and uncoupled slopes.

Image processing

Microscope and gel images have been linearly contrast-enhanced in ImageJ (Wayne Rasband, NIH).

Statistical analysis

Experiments were typically done with 3-6 replicates. When applicable, Student's t-test was used to test for significant differences between groups at P < 0.05.

Results

Low-density seeding of C₂C₁₂ cells permits observation of mitochondrial dynamics

Because our interest is to study mitochondrial dynamics in mammalian skeletal muscle fibers, we initially seeded C₂C₁₂ myoblasts at high density, permitting the cells to fuse into multinucleate, elongated myotubes (Fig. S3C). However, we found the MitoTracker-stained mitochondria of such myotubes to be too dense to resolve satisfactorily. Undifferentiated myoblasts (Fig. S3A) were also deemed to be an unsuitable model: MitoTracker staining was diffuse and revealed a mitochondrial “network” that, if not a staining artifact, appeared to be already fragmented before any treatment. Accordingly, others have shown that myoblasts express less mitofusin-2 than do myotubes [29]. Therefore, we studied low-density-seeded C₂C₁₂ myocytes (Fig. S3B), which were delayed but not blocked in myotube formation. Mitochondria from such cells showed a baseline configuration that more closely approximated that which exists in skeletal muscle of human and rodent models [1-3]. While low-density-seeded myocytes retain the cellular morphology of myoblasts, their expression of several mitochondrial and metabolic protein markers was found to resemble that of myotubes (Fig. S4). In contrast, myoblasts showed much lower expression of voltage-dependent anion channel (VDAC). Following these preliminary findings, all subsequent experiments were conducted using low-density-seeded C₂C₁₂ myocytes.

250 μM H₂O₂ induces a reversible fragmentation of the mitochondrial network

We found that 250 μM H₂O₂ induced the mitochondrial network to fragment into punctiform mitochondrial vesicles in cells observed 6 hours post-treatment (Fig. 1). Quantification was independently verified by skeletonization analysis (Fig. S5). For cells incubated for 3 hours in various concentrations of H₂O₂, the extent of fragmentation was dose-dependent (Fig. 2A). However, because mitochondrial fission is a key step in apoptosis [30], and severe oxidative stress may induce apoptosis, we also examined cells 24 and 48 hours after an initial 6-hour treatment. We observed that while nearly all cells showed the fragmented phenotype after 6 hours, only a small fraction showed signs of punctiform mitochondria after 24 or 48 hours (Fig. 1). Mock treatment with medium lacking H₂O₂ did not result in any observable change. Furthermore, cytotoxicity was negligible with 250 μM H₂O₂ (Fig. 2B), but 1000 and 2000 μM H₂O₂ did reduce viability, which was confirmed by CLSM (not shown).

Figure 1. H₂O₂-induced mitochondrial network fragmentation is reversible.

C₂C₁₂ cells were treated with 250 μM H₂O₂ or 0 μM H₂O₂ (control) for 6 hours, then stained with MitoTracker Deep Red 633 for 30 minutes and visualized by CLSM immediately, or after a 24- or 48-hour incubation. For each image taken, the total number of cells and the number of cells displaying a reticular mitochondrial network were counted. A, representative images for each of the conditions. B, the percentage (mean ± SEM) of total imaged cells displaying a reticular mitochondrial network under each condition. An average of 235 cells was scored per condition. *Significantly different from all other bars, P < 0.001. Six-hour treatment with 250 μM H₂O₂ significantly decreased the percentage of cells having a reticular mitochondrial network, but the cells recovered within 24 hours.

Figure 2. Dose-dependence, cytotoxicity assessment, and time-dependence of H₂O₂-induced mitochondrial network fragmentation.

A, C₂C₁₂ cells were treated with 0, 62.5, 125, 250, and 500 μM H₂O₂ for 3 hours, stained with MitoTracker Deep Red, visualized by CLSM, and categorized by network morphology. The percent of cells in the “fragmented” category is shown for each condition. Mock-treated (0 μM H₂O₂) myocytes showed a predominantly reticular network morphology, which became increasingly fragmented as H₂O₂ dose was increased. An average of 52 cells was scored per condition. Data are presented as percents of the total number of cells scored in each treatment group. Data from 2 independent experiments are shown. B, myocytes were treated with 0, 125, 250, 500, 1000, or 2000 μM H₂O₂ for 6 hours and assayed for cell viability with WST-8 immediately (triangles), or washed and returned to H₂O₂-free medium, and assayed 24 hours later (squares). Data are presented as means ± SD of 5-6 replicates. Notably, the H₂O₂ concentration used in subsequent experiments (250 μM) did not significantly lower cell viability. Student's t-test was used to test for statistically significant differences between each condition and the respective 0 μM H₂O₂ control. NS, not significantly different from mock-treated (0 μM H₂O₂); *P < 0.05; **P = 0.001; ***P < 10^-7. C, cells were incubated in medium II initially containing 250 μM H₂O₂ for 5 minutes or 1 to 6 hours, then stained with MitoTracker for 30 minutes and visualized by CLSM. The percent of cells in the “fragmented” category is shown for each time point. The occurrence of fragmented mitochondria appears to peak at 5.5 hours. An average of 51 cells was scored per time point. Data from 2 independent experiments are shown.

Fragmentation peaks 5.5 hours after the addition of 250 μM H₂O₂

Because our initial observation of H₂O₂-induced fragmentation was made following a 6-hour treatment, we sought to better characterize the time course of mitochondrial fragmentation. By visualizing cells with CLSM after 250 μM H₂O₂ pre-treatments of different durations (from 5 minutes to 6 hours), we found that fragmentation proceeded slowly, peaking at 5.5 hours with 250 μM H₂O₂ (Fig. 2C). In consideration of previous findings of rapid fragmentation caused by various stimuli in other cell types [7, 9, 31-33], we sought to determine whether there could be “fast” fragmentation in C₂C₁₂ myocytes. We found that while 100 nM of the respiratory uncoupler CCCP had no effect even after 15 minutes, 10 μM CCCP induced cytosolic MitoTracker leakage and the appearance of numerous punctiform mitochondria within 3 minutes of its addition to live cells (Fig. S6).

250 μM H₂O₂ in medium II at 37°C decomposes within 1 hour

To monitor extracellular H₂O₂ concentration over time, we periodically withdrew supernatant from H₂O₂-incubated cells and assayed its H₂O₂ content via a colorimetric assay. We found that [H₂O₂] declined rapidly, falling to the limit of detection after 1 hour (Fig. 3). This decomposition followed first-order kinetics, with apparent half-lives of 6.3 and 9 minutes in 37°C medium II with and without myocytes, respectively (Fig. 3, inset). Although H₂O₂ itself did not persist long extracellularly, we were concerned that H₂O₂ could react with components of the medium to generate other ROS that mediate fragmentation. Ruling out this possibility, we found that the extent of fragmentation was similar whether H₂O₂-treated cells were washed (and replaced in H₂O₂-free medium) 1, 2, or 5 hours post-treatment (not shown). Thus, fission toward the fragmented state proceeded in the absence of continuing extracellular stimulation.

Figure 3. Mean hydrogen peroxide concentration in cell culture supernatant as a function of time.

At time zero, 250 μM H₂O₂ was added to dishes with (solid diamond) or without (open triangle) C₂C₁₂ cells seeded. In a mock experiment (open square), H₂O₂-free medium was added to a dish of cells. Data are presented as means ± SD of 3 replicate dishes. Inset, the same data is plotted, but with the natural logarithm of [H₂O₂] on the ordinate axis. The disappearance of H₂O₂ followed first-order kinetics (t_1/2 = 6.3 min in the presence of C₂C₁₂ cells; 9.0 min in their absence).

250 μM H₂O₂-treated cells take up less JC-1, suggesting a diminished ΔΨ_m

The morphology of the mitochondrial network is determined by the relative rates of fusion and fission. In the simplest case, the demonstrated H₂O₂-induced fragmentation can be explained by the promotion of fission, or the inhibition of fusion. Using quantitative RT-PCR, we did not detect any significant changes in the mRNA levels of Mfn1, Mfn2, OPA1, Fis1, Drp1, or cytochrome c in C₂C₁₂ cells treated for 6 hours with 250 μM H₂O₂ (Supplementary Table 1, primers given in Supplementary Table 2). Drp1 is known to translocate from the cytoplasm to mitochondria upon activation [31]. However, we also did not detect any significant change in Drp1 protein levels (normalized to VDAC) in the mitochondrial fraction of C₂C₁₂ cells after the same H₂O₂ treatment (Fig. S7). The slow kinetics of H₂O₂-induced fragmentation are consistent with the gradual accumulation of punctiform mitochondria due to continued fission in the absence or reduction of fusion. Several investigators have shown that ΔΨ_m is required for fusion [34-38] and we confirmed that CCCP causes rapid fragmentation. We hypothesized that 250 μM H₂O₂ decreases, but does not completely abolish ΔΨ_m and reduces mitochondrial fusion. Using flow cytometry to measure the loading of JC-1 in trypsinized H₂O₂-pretreated C₂C₁₂ cells (Fig. 4), we confirmed a diminished ΔΨ_m both 1 and 5 hours post-treatment (Fig. 4C). Specifically, the median fluorescence intensity at 575 nm, corresponding to “J-aggregates,” was significantly lower in 1-hour H₂O₂-pretreated cells that were then stained with JC-1 (Fig. 4B). To be sure that JC-1 staining was responsive to ΔΨ_m, JC-1-stained cells were treated with 10 μM CCCP, which dramatically decreased 575 nm fluorescence and increased 525 nm fluorescence (Fig. 4B).

Figure 4. 250 μM H₂O₂ pretreatment decreases JC-1 loading.

C₂C₁₂ cells were pretreated with 250 μM H₂O₂ or 0 μM H₂O₂ (“mock”) for 1 or 5 hours, or were not pretreated (“untreated”). Subsequently, cells were stained in 2 μM JC-1 for 30 minutes, trypsinized, and analyzed by flow cytometry. Some “untreated” cells were treated with 10 μM CCCP before flow cytometry (denoted “CCCP”). A, two-parameter (log scale) pseudo-colored dot plots for fluorescence detected at 525 nm and 575 nm. B, the data from A shown as superimposed single-channel histograms. Notably, H₂O₂ pretreatment (whether for 1 hour or 5 hours) causes a marked shift toward lower 575 nm fluorescence. In A and B, data from 1 representative experiment of 3 are shown. C, quantification of JC-1 fluorescence at 575 nm (J-aggregates). The median fluorescence intensity of each single-channel histogram was expressed as a percentage of the “untreated” control of that experiment. Mean percentages ± SEM of 3 independent experiments are shown. *P < 0.01; **P < 0.05. One- and five-hour pretreatments with 250 μM H₂O₂ appear to have similar effects in decreasing median 575 nm fluorescence intensity.

250 μM H₂O₂ alters cellular respiration

Because a high ΔΨ_m normally checks further pumping of protons into the mitochondrial intermembrane space, a decrease in ΔΨ_m would be expected to exert an uncoupling effect and accelerate oxygen consumption; this is indeed the case when CCCP is used to dissipate ΔΨ_m (Fig. 5). We used a Clark-type electrode to measure dissolved O₂ in the medium incubating trypsinized H₂O₂- or mock-pretreated cells (Fig. 5), or H₂O₂-pretreated cells 24 hours after the end of the treatment period (i.e., 25 hours after H₂O₂ addition). We did not observe a statistically significant difference in endogenous respiration normalized to cell number, between any groups (Fig. S8A), although cells pretreated with H₂O₂ for 1 or 5 hours exhibited higher endogenous respiration on average. However, the rate of uncoupled (maximal) respiration was significantly lower in cells that had been pretreated with H₂O₂ for 1 or 5 hours (Fig. S8B). Because normalizing respiration rates by cell number compounds their respective measurement errors, a more precise metric is the ratio between uncoupled and endogenous respiration. This ratio is the multiplicative factor by which the endogenous respiration rate increases when 10 μM CCCP is added. When this is considered (Fig. 5, inset), untreated cells and cells that have been mock-treated with medium show no significant differences from each other. However, cells 1, 5, or 25 h post-H₂O₂-addition had significantly lower maximal-to-endogenous respiration ratios compared to any of the controls. Furthermore, 25 h post-H₂O₂ cells had significantly elevated ratios than 5 h post-H₂O₂ cells, suggesting recovery.

Figure 5. H₂O₂-treated cells show depressed uncoupled-to-endogenous respiration ratios 1, 5, or 25 hours post-treatment.

Representative respiration traces of intact C₂C₁₂ myocytes, measured using a Clark-type oxygen electrode, are shown. Untreated, mock-pretreated, or H₂O₂-pretreated cells were trypsinized and added (triangle) to temperature-equilibrated medium, and endogenous respiration was recorded for 15 minutes. Then, CCCP (final concentration 10 μM) was added (diamond) and uncoupled respiration was recorded for at least 10 minutes. The values of the slopes of the endogenous and uncoupled respiration traces are shown. Inset, the uncoupled respiration slope of each trace was divided by the endogenous respiration slope of that trace. Data are shown as means ± SEM of at least 6 independent experiments each. *P ≤ 0.01. The partial recovery seen 25 hours post-treatment suggests that the H₂O₂-induced depression was transient.

Discussion

In an established model of skeletal muscle, we confirm the existence of a myocyte mitochondrial reticulum and show that it responds reversibly to H₂O₂-induced oxidative stress. Given the enormous metabolic demand of skeletal muscle during exercise and the ROS released from contracting muscle, the phenomenon we describe may model mitochondrial changes that occur during exercise in vivo (e.g., [26]).

Our microscopy results confirm the findings of Jendrach et al., who studied the effects of H₂O₂ on the mitochondrial network in human endothelial cells [17]. Notably, these authors also described a “slow” fragmentation (i.e., on the order of hours). However, they used much higher doses of H₂O₂ for shorter durations. The ability of endothelial cells to tolerate these high doses of H₂O₂ may be cell-type-specific, or due to particular culture conditions. They also report significant recovery of the reticular morphology 24 hours post-treatment. Slow fragmentation was also observed in neurons treated with nitric oxide [18].

The H₂O₂-induced slow fragmentation we describe is kinetically inconsistent with previous reports of Ca⁺⁺-dependent fission activation [8, 31-33], most often found to be mediated by increased Drp1 activity. Instead, our observations suggest a reduction in fusion and the gradual development of fragmentation due to continuing fission. Importantly, fragmentation developed in the absence of H₂O₂ itself (Fig. 3) or H₂O₂-treated medium. We surveyed the major proteins known to be involved in mitochondrial dynamics (Mfn1, Mfn2, OPA1, Drp1, Fis1), but did not find any of them to be regulated transcriptionally after a 6-hour treatment with 250 μM H₂O₂ (Supplementary Table 1). Likewise, we did not observe any significant change in Drp1 protein levels in the mitochondrial fraction (Fig. S7). These findings indicate that H₂O₂ does not modulate mitochondrial dynamics at the level of transcription, nor does it activate mitochondrial fission through signaling.

We confirm in myocytes that depolarization of ΔΨ_m using CCCP causes rapid fragmentation of the mitochondrial reticulum [34-38]. This has been interpreted to mean that mitofusin activity depends on ΔΨ_m. It is therefore possible that H₂O₂ inhibits mitochondrial fusion by decreasing ΔΨ_m. Using flow cytometry, we observed that cells pretreated with 250 μM H₂O₂ for 1 hour showed reduced 575 nm median fluorescence (Fig. 4B). This finding suggests decreased loading of JC-1 into cells and decreased J-aggregate formation as a result of decreased ΔΨ_m. Salvioli et al. have shown that plasma membrane depolarization does not significantly affect JC-1 staining, while ΔΨ_m depolarization has a strong effect [27], so our results specifically indicate decreased ΔΨ_m. At the concentration of JC-1 used (2 μM), 525 nm fluorescence is saturated (Fig. S2), consistent with our observation that H₂O₂ pretreatment decreases 575 nm but not 525 nm fluorescence.

Overall, our findings in C₂C₁₂ cells agree with previous reports of the interplay between H₂O₂ and ΔΨ_m. Takeyama et al. showed that 500 μM H₂O₂ decreased ΔΨ_m in L929 fibroblasts by opening the mitochondrial permeability transition pore [39]. Chinopoulos et al. observed that H₂O₂ decreased ΔΨ_m, α-ketoglutarate dehydrogenase activity, and NAD(P)H levels in synaptosomes [40]. Nulton-Persson and Szweda showed that H₂O₂ inhibits succinate dehydrogenase and aconitase as well [41]. Because 250 μM H₂O₂ does not decrease viability in our cells, the hypothesis that H₂O₂ inhibits certain Krebs cycle enzymes, leading to decreased substrate input for oxidative phosphorylation and decreased ΔΨ_m, seems to be most plausible.

If ΔΨ_m decreases, we would predict consequent changes in respiration that can be measured using a Clark-type electrode. CCCP accelerates the rate of oxygen consumption by 400% within minutes (Fig. 5) but when H₂O₂ was added directly to cells, endogenous respiration did not increase in the following 15 minutes (not shown). Cells that had been pretreated with H₂O₂ for 1 or 5 hours showed higher endogenous respiration rates on average, but the difference was not statistically significant (Fig. S8A). However, H₂O₂ did decrease the CCCP-uncoupled (maximal) respiration rate (Fig. S8B). That is, H₂O₂-treated cells had lower reserve capacities for respiration (Fig. 5), a finding similar to that observed by Chen et al. [42] when both mitofusins were knocked down. Thus, the respiration of H₂O₂-treated cells resembles that of Mfn1/2-knockout cells, consistent with an abrogation of fusion.

With our flow cytometry and respiration measurements, we observed that H₂O₂-induced changes occurred within one hour (before any observable fragmentation) and persisted through the fifth hour post-treatment (contemporaneous with the full extent of fragmentation). The timing of these changes is therefore consistent with the hypothesis that decreased ΔΨ_m leads to fragmentation, rather than occurring as a consequence of fragmentation. However, other than through the use of CCCP, we were not able to directly manipulate ΔΨ_m to show that decreased ΔΨ_m is necessary and sufficient for mitochondrial network fragmentation. Because CCCP is a powerful, non-physiological stimulus, it is conceivable that it and H₂O₂ affect ΔΨ_m and/or mitochondrial morphology in fundamentally different ways. Further experiments would need to functionally assay for a fusion deficiency in H₂O₂-treated cells, perhaps using the PEG cell fusion assay described by Chen et al. [14, 42].

Before closing, it is appropriate to address the existence of a dynamic mitochondrial network in muscle and the possible role of H₂O₂ in modulating its dynamics in vivo. Despite the complicated internal structure of muscle, several studies have demonstrated the morphological plasticity of muscle mitochondria. Using electron microscopy, Green and colleagues (e.g., [43]) showed conformational changes in mitochondrial morphology in response to energy state. It is also known that the mass of the muscle mitochondrial network increases twofold in response to chronic exercise [24, 25]. We propose that ROS released from contracting muscle may be a stimulus to induce these changes. More recently, Chen et al. showed that the abolishment of mitochondrial fusion in skeletal muscle led to muscle atrophy and mtDNA instability— strong evidence for the functional significance of mitochondrial dynamics in muscle under normal conditions [16]. Additionally, mitochondrial dynamics may play an important role in the pathology of diseases involving oxidative stress, like type II diabetes [44]. It has been shown that obese and type II diabetic human subjects [45] and obese rats [29] display smaller skeletal muscle mitochondria than their lean counterparts. These authors show concomitant downregulation of Mfn2 gene expression [29, 46], but modulation of the mitochondrial network by ROS may also contribute. Now, in our report, we link H₂O₂ to transitory mitochondrial fragmentation in a muscle cell system in which mechanical forces are not an issue. This fragmentation is dose- and time-dependent, reversible, non-cytotoxic, and temporally offset from the H₂O₂ challenge. Additionally, we found that H₂O₂-induced decreases in mitochondrial membrane potential and maximal respiration rate precede mitochondrial fragmentation. We suggest that the observed changes mimic transitory changes that occur in response to exercise in vivo and provide one mechanism by which the aforementioned exercise-induced mitochondrial plasticity may occur.

List of abbreviations

H₂O₂

Hydrogen peroxide

ROS

reactive oxygen species

CLSM

confocal laser scanning microscopy

ΔΨ_m

inner mitochondrial membrane potential

Drp1

dynamin-related protein 1

Mfn1/2

mitofusin 1 and mitofusin 2

OPA1

optic atrophy 1

CCCP

carbonyl cyanide m-chlorophenyl hydrazone

JC-1

5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide