Electrical stimulation prevents doxorubicin-induced atrophy and mitochondrial loss in cultured myotubes

Abstract

Muscle contraction may protect against the effects of chemotherapy to cause skeletal muscle atrophy, but the mechanisms underlying these benefits are unclear. To address this question, we utilized in vitro modeling of contraction and mechanotransduction in C2C12 myotubes treated with doxorubicin (DOX; 0.2 μM for 3 days). Myotubes expressed contractile proteins and organized these into functional myofilaments, as electrical field stimulation (STIM) induced intracellular calcium (Ca²⁺) transients and contractions, both of which were prevented by inhibition of membrane depolarization. DOX treatment reduced myotube myosin content, protein synthesis, and Akt (S308) and forkhead box O3a (FoxO3a; S253) phosphorylation and increased muscle RING finger 1 (MuRF1) expression. STIM (1 h/day) prevented DOX-induced reductions in myotube myosin content and Akt and FoxO3a phosphorylation, as well as increases in MuRF1 expression, but did not prevent DOX-induced reductions in protein synthesis. Inhibition of myosin-actin interaction during STIM prevented contraction and the antiatrophic effects of STIM without affecting Ca²⁺ cycling, suggesting that the beneficial effect of STIM derives from mechanotransductive pathways. Further supporting this conclusion, mechanical stretch of myotubes recapitulated the effects of STIM to prevent DOX suppression of FoxO3a phosphorylation and upregulation of MuRF1. DOX also increased reactive oxygen species (ROS) production, which led to a decrease in mitochondrial content. Although STIM did not alter DOX-induced ROS production, peroxisome proliferator-activated receptor-γ coactivator-1α and antioxidant enzyme expression were upregulated, and mitochondrial loss was prevented. Our results suggest that the activation of mechanotransductive pathways that downregulate proteolysis and preserve mitochondrial content protects against the atrophic effects of chemotherapeutics.

INTRODUCTION

Cancer afflicts approximately one in two men and one in three women (31). With improved detection and treatments and increased incidence due to an aging populace (1), the number of individuals living with effects of cancer and its treatment will exceed 18 million by 2020 (12) and continue to grow. Cancer and its treatment have profound effects on skeletal muscle, including atrophy, impaired contractility, and oxidative dysfunction (70, 71, 73), that negatively impact quality of life, treatment decisions, and survival (4, 6, 13, 55) and can persist for years after treatment (44). These detrimental effects reduce quality of life and predispose to disability and cardiometabolic disease.

Exercise can mitigate the negative effects of cancer and its associated treatment and improve long-term prognosis (28, 33, 72). Part of these benefits may derive from the positive effect of exercise on skeletal muscle. Exercise induces an array of beneficial adaptations in muscle, including improved proteostasis, energy metabolism, and contractility (3, 16). Although the effects of various exercise regimens have been examined in patients (1, 37, 47), the mechanism(s) whereby exercise derives its benefits remains unknown. That is, whether the beneficial effect of exercise derives from the simple neural activation of muscle, the relative amount of stress/strain on the muscle, or some combination of these factors is not known. Knowledge of which of these fundamental mechanisms underlie the benefits of exercise could inform the development of more effective regimens that derive the greatest benefit with the least burden on patients. Therefore, we utilized in vitro modeling of repeated bouts of muscle contraction in C2C12 myotube cultures subjected to chemotherapeutic treatment, with or without electrical field stimulation (STIM), to dissect the mechanisms underlying the beneficial effects of contraction on skeletal muscle. We focused on the potential beneficial effects of muscle contraction to curtail muscle atrophy and maintain or improve mitochondrial content and function. We hypothesized that STIM promotes its beneficial effects, in part, via mechanotransductive signaling; that is, via signaling mechanisms activated by mechanical stress/strain imposed by contraction. Furthermore, we hypothesized that these signaling events upregulate protein synthesis and preserve mitochondrial content and function following administration of the chemotherapeutic doxorubicin (DOX). We chose DOX as our model chemotherapeutic in these studies because it is used in a broad range of cancers (32, 68, 69) and has well-defined myotoxicity in striated muscle (20, 22, 23) and also because studies have shown the potential of exercise to counter the myotoxic effects of DOX (41, 53, 63).

METHODS

Cell culture.

C2C12 myoblasts (CRL1772; American Type Culture Collection, Manassas, VA) were cultured in growth medium, consisting of low-glucose (1 g/L) Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS; GIBCO, Thermo Fisher Scientific, Waltham, MA) and antibiotics (50 U/mL penicillin and 50 μg/mL streptomycin), incubated at 37°C. Cells were plated (2 × 10⁴ cells/cm²) on Matrigel (60 μg/cm²; Corning, Bedford, MA) and switched to differentiation medium (DM) of low-serum (1% heat-inactivated FBS), high-glucose (4.5 g/L) DMEM when they reached 90–100% confluence, as described (45), with modifications as described (25). On day 7 postdifferentiation (d7), myotubes were treated with DOX (0.2 μM) or vehicle control (DMSO in DM) for 3 days (chronic experiments). Thirty minutes after DOX treatment was started, STIM was applied using a C-Pace pulse generator (20 V, 1 Hz, 12 ms; C-Pace 100; IonOptix, Milton, MA) for 1 h each day for 3 days. At the end of each STIM bout, myotubes were washed twice with Hanks’ balanced salt solution (HBSS), fresh DM containing either DOX or vehicle (DMSO) was added, and 23 h were allowed before the next bout of STIM or measurements. In some experiments, myotubes were treated with tetrodotoxin (TTX; 10 µM), a sodium channel inhibitor, or N-benzyl-p-toluene sulfonamide (BTS; 50 µM), an inhibitor of myosin-actin ATPase cycling (64), during the 1-h STIM bouts. Cells were lysed and collected on d10, 23 h after the last STIM bout, to eliminate the possibility that results reflect an acute effect of the STIM intervention. In acute experiments, cells were treated with vehicle, DOX alone, STIM, or STIM+DOX, as described above, to examine acute modulation of protein synthesis, expression of proteolytic mediators, and signaling molecule phosphorylation and harvested/measured after 1 or 24 h. In another set of acute experiments, cells were treated with vehicle, DOX, cisplatin (CIS; 10 µM), or paclitaxel (TAXOL; 40 nM) without STIM and harvested after 24 h.

A separate set of acute experiments examined the effects of static mechanical stretch on C2C12 myotubes differentiated on type I collagen-coated (Advanced Biomatrix, San Diego, CA) silastic membranes (Gloss/Gloss, 0.02 in.; Specialty Manufacturing, Saginaw, MI). Silastic membranes were mounted on two friction-fit C-clamps, as previously described (19). The screw nuts were set to 30 cycles from baseline when the stretching device was assembled, as described (19). C2C12 myoblasts were suspended (~1 × 10⁶ cells/mL) and plated onto Silastic membranes (~1.5–2 × 10⁵ cells per stretching device). The cells were grown to ~95% confluence and differentiated into myotubes using 2% horse serum. After 5 days of differentiation, cells were treated with DOX (0.2 μM) or vehicle control (DMSO in DM). A 5% stretch was supplied for a duration of 1 h, and in cells receiving DOX, stretch was started 30 min following the first DOX administration, identical to the STIM intervention detailed above. After 1 h of stretch, the membranes were returned to baseline, cells were washed two times with HBSS, and media were replaced with fresh media with or without DOX for a duration of 23 h, after which cells were lysed and collected, as described above.

Myotube size.

Myotube size was assessed from myotube diameter measurements. Digital images of myotube cultures were acquired at ×4 magnification. Average diameters (5 diameters per tube) of n = 5–25 myotubes per field from n = 4 random fields were measured using ImageJ software (National Institutes of Health, Bethesda, MD) by an assessor blinded to treatment status.

Immunocytochemistry.

Myofilament proteins were visualized by immunocytochemistry. Cells were grown on Matrigel-coated (60 μg/cm²), 35-mm, glass bottom imaging dishes (MatTek; Ashland, MA) or plastic, as detailed above, with the modification that the media were changed daily. Cells were fixed with 4% paraformaldehyde (Fisher Scientific, Atlanta, GA), permeabilized with 0.2% Triton X-100 (Fisher), and blocked with 5% BSA in PBS for 1 h at room temperature. Cells were incubated overnight at 4°C in fast-twitch skeletal muscle myosin antibody (1:500, MY-32; Sigma) followed by secondary antibody (1:100, anti-mouse IgG; Molecular Probes) to visualize myofilaments or 1 μM tetramethylrhodamine isothiocyanate-labeled phalloidin (Sigma) to stain actin to visualize the entire cell. Cells were imaged using a Nikon Ti-E inverted microscope with C2 confocal at ×40 for myofilament measures or an Olympus BX51 with QImaging Retiga R6 at ×10.

Measurement of contractility and Ca²⁺ cycling.

Ca²⁺ transients were recorded from d7–d10 myotubes grown on Matrigel-coated (60 μg/cm²), 35-mm, glass bottom imaging dishes (MatTek, Ashland, MA). For these experiments, cells were plated at a higher density (2.5 × 10⁴ cells/cm²), and DMEM was changed daily. C2C12 myotubes were loaded with 1 µM Fluo-2-acetoxymethyl ester (Fluo-2 AM; TefLabs, Austin, TX) for 15 min at 37°C in the dark. Cells were washed once with HBSS and placed in prewarmed DM for 10 min. The culture dish was fitted with a custom-built insert that maintained media temperature at 37°C and contained platinum electrodes to allow STIM with biphasic pulses (20 V, 1 Hz, 12 ms; Myopacer; IonOptix, Westwood, MA). The same experimental design used for performing intracellular Ca²⁺ recordings was applied to contractility measurements. Fluorescent signal and cell contractility were traced using an IonOptix system, as previously described (74). Ca²⁺ fluorescence was recorded with an inverted fluorescence microscope and galvanometer-controlled, dichroic mirror filters at 480 and 510 nm for excitation and emission, respectively (Hyperswitch; IonOptix; 58). Contractions were tracked using the edge detection feature of the IonWizard data acquisition software using visible landmarks on/within the myotube. Both contraction and Fluo-2 AM fluorescence measurements were made simultaneously from the same myotube. Experiments lasted 300 s. Transient analysis was performed using the IonWizard analysis software (IonOptix). For each test condition, data for 15–20 s of Ca²⁺ transients or contractions per myotube were averaged, using the pacing time as a common reference point, to derive an averaged monotonic Ca²⁺/contractility transient. Fractional change, which indicates the percentage of peak following STIM relative to baseline, was used to quantify contractile dynamics and Ca²⁺ transients. Images of the Ca²⁺ fluorescent signal were acquired (40 frames/s) using a spinning-disk confocal microscope (CSU-W1; Yokogawa). During these experiments, cells were stimulated (20 V, 1 Hz, 12 ms) with platinum electrodes.

An additional set of experiments were performed without the measurement of Ca²⁺ fluorescence to assess contractility. To aid in the visualization of cell movement, 2-μm latex beads (1:1,000, L-3030; Sigma) were added to the culture dish 2 h before STIM. Digital video images were collected with a down-sampled field of 512 × 512 pixels and a frame rate of 10 Hz (MyoCam-S3; IonOptix). Video images were analyzed by first establishing a hexagonal grid and tracking each local image featured within each hexagon by cross-correlation from frame to frame (Fig. 2A, left). The displacement of a local image relative to its neighboring local images allowed for calculation of local strain (i.e., change in relative displacement between neighbors normalized to the original distance between neighbors). With each local image having six neighbors, three strains oriented at 120° to each other were measured. This orientation of strains allowed for calculation of magnitude and orientation of principal strains on a Mohr’s circle similar to that based on a delta rosette strain gauge (10). Local strains are depicted as green ellipses illustrating the contraction or elongation of the local region (Fig. 2A, right). From these assessments, relative strain was calculated. This approach should improve upon the precision of contractility assessments because it measures the displacement of all optical features within a given hexagonal region relative to all six neighboring hexagons and not just that of observable edges of two cellular landmarks/features.

Fig. 2.

Contractile dynamics and intracellular Ca²⁺ cycling in day 7–10 postdifferentiation C2C12 myotubes. A: digital image of C2C12 myotube at rest with hexagonal grid overlay (scale bar = 10 µm; left). A blowup of the circled hexagon is provided for baseline (frame 0) and during the peak of contraction (frame 17) to illustrate contraction monitoring by pixel tracking (right). Yellow colored ellipses show trackable image contrast, with green ellipses tracking local pixel strains and negative displacements signifying contractions. B: negative displacements of one hexagonal grid showing several contraction cycles. C: fractional strain of n = 30 myotubes measured with edge tracking of manually selected myotube features/landmarks (○) and pixel-tracking approach described above and in methods (●). Data are means ± SE, with individual data points shown. D: average fractional change (Δ) in contraction and intracellular Ca²⁺ cycling under basal conditions (n = 14 and 17 myotubes, respectively) and following tetrodotoxin (TTX; n = 8 and 7 myotubes, respectively) or N-benzyl-p-toluene sulfonamide (BTS) administration (n = 5 and 9 myotubes, respectively). *P < 0.05 and **P < 0.01 compared with its respective basal conditions using one-way ANOVA.

Myotube mitochondrial content and reactive oxygen species production.

Mitochondrial content and reactive oxygen species (ROS) production were measured with fluorometric dyes, as described (25). Briefly, C2C12 myotubes grown in 35-mm dishes or black-walled 96-well plates were loaded with fluorescent dyes to assess mitochondrial content (1 μM MitoTracker Green FM; 490 and 516 nm for excitation and emission, respectively) and ROS production (1 μM MitoSOX Red; 510 and 580 nm for excitation and emission, respectively; both Molecular Probes, Eugene, OR) 15 min before measurement. Fluorescence was measured on a microplate reader (BioTek, Winooski, VT) for 15 min in the basal condition, and the MitoSox signal was expressed relative to MitoTracker signal to control ROS production for variation in mitochondrial content. We have shown that the MitoSox and MitoTracker signals colocalize (25).

Oxygen consumption rate.

The oxygen consumption rate (OCR) of C2C12 myotubes was measured using a Seahorse XFe96 analyzer (Seahorse Biosciences, North Billerica, MA). Briefly, 1 × 10⁴ C2C12 cells per well were seeded in a Seahorse XFe 96-well cell culture microplate, as described above, and differentiated for 7 days. Myotubes (d7) were treated with culture medium containing vehicle (control), DOX, CIS, or TAXOL at concentrations noted above for 1 day. Thereafter, culture medium was removed, cells were washed twice with HBSS, and media were replaced with Seahorse assay medium supplemented with 10 mM sodium pyruvate and 10 mM glucose, pH 7.4. This medium was used as a diluent for compounds injected into the culture dishes during measurements. Plates were equilibrated for 1 h at 37°C with no CO₂ before they were transferred into the XFe96 analyzer. After the basal oxygen consumption rate was measured, the analyzer sequentially injected the following compounds (final concentration): oligomycin (4 μM), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP; 4 μM), and rotenone/antimycin A (4 μM each; all from Sigma), with each well measured three times. Following OCR measurements, MitoTracker Green was applied to cells, as detailed above, and assessed to normalize for mitochondrial content.

Protein synthesis.

Protein synthesis was measured using the surface sensing of translation (SUnSET) technique, as described (19). Briefly, 23 h after the last bout of STIM, stretch, or control conditions, 1 μM puromycin (Sigma) was added to myotubes for 30 min before lysis and collection. Lysates were then analyzed by Western blot, as described below.

Protein expression.

Protein expression or phosphorylation was measured by Western blot analysis. Myotubes were washed with PBS, lysed [50 mM Tris, 150 mM NaCl, 10% (vol/vol) glycerol, 0.5% IGEPAL CA-630, and 1 mM EDTA, containing Protease Inhibitor Cocktail (1:100, cat no. P-8340; Sigma) and Phosphatase Inhibitor Cocktail 3 (1:100, cat no. P-0044; Sigma)], incubated on ice for 30 min, and then centrifuged at 14,000 g at 4°C for 10 min. Lysate protein contents were measured (DC Protein Assay; Bio-Rad, Hercules, CA) and diluted in gel loading buffer. Proteins were separated by SDS-PAGE (Bio-Rad), transferred to polyvinylidene difluoride membranes, and blocked with Tris-buffered saline-Tween buffer (TBST; 150 mM NaCl, 0.05% Tween 20, and 20 mM Tris·HCl, pH 7.4) containing 5% nonfatty milk or BSA. After blocking, membranes were incubated overnight with primary antibody: anti-puromycin, clone 12D10 (1:2,000, Millipore MABE-343, Research Resource Identifier RRID:AB_2566826), myosin (1:20,000, Sigma M-4276, RRID:AB_477190), E3 ubiquitin-protein ligase TRIM63 (TRIM63)/muscle RING finger 1 (MuRF1; 1:1,000, R&D Systems AF-5366, RRID:AB_2208833), phospho-p70 S6 kinase (T389; 1A5; 1:1,000, Cell Signaling Technology 9206, RRID:AB_2285392), p70 S6 kinase (1:1,000, Cell Signaling Technology no. 2708, RRID:AB_390722), phospho-AMPK-α (T172, 1:2,000, Cell Signaling Technology 2535, RRID:AB_331250), AMPK (1:2,000, Cell Signaling Technology 23A3, RRID:AB_490795), phospho-Akt (T308, 1:1000, Cell Signaling Technology 5056, RRID:AB_10695743), Akt (1:1,000, Cell Signaling Technology 4691, RRID:AB_915783), phospho-forkhead box O3a (phospho-FoxO3a; S413, 1:1,000, Cell Signaling Technology 8174, RRID:AB_10889562), phospho-FoxO3a (S253, 1:1,000, Cell Signaling Technology no. 9466, RRID:AB_2106674), FoxO3a (1:1,000, Cell Signaling Technology 2497, RRID:AB_836876), superoxide dismutase 2 (SOD2; 1:1,000, Cell Signaling Technology 13194, RRID:AB_2750869), catalase (CAT; 1:1,000, Cell Signaling Technology 14097, RRID:AB_2798391), and peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α; 1:500, Millipore AB-3242, RRID:AB_2268462). Membranes were washed for 30 min in TBST and then incubated for 1 h at room temperature in 5% milk-TBST containing secondary antibody. For puromycin measurements, the entire lane was quantified by densitometry.

RNA isolation and quantitative real-time PCR.

Total RNA was isolated from C2C12 cells using TRIzol reagent followed by a chloroform-isopropanol extraction (Invitrogen), and the concentration and quality were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher, Waltham, MA). cDNA was synthesized from 100 ng of mRNA using the qScript Supermix reagent kit per manufacturer’s instructions (Quantabio, Beverly, MA). Quantitative real-time PCR was performed using iTaq Universal SYBR Green Supermix (Bio-Rad) on a CFX96 Touch system (Bio-Rad), with the relative mRNA expression calculated using the threshold cycle (Ct; 2^−ΔΔCt) method normalized to Gapdh expression. Primer sequences are provided in Table 1.

Table 1.

Primer sequences for gene expression analysis

Gene	Forward 5′–3′	Reverse 5′–3′
Cat	CGAGGGTCACGAACTGTGTCA	GGTCACCCACGATATCACCAGATAC
Gclc	ATCTGCAAAGGCGGCAAC	ACTCCTCTGCAGCTGGCTC
Gclm	AGTTGGAGCAGCTGTATCAGTGG	TTTAGCAAAGGCAGTCAAATCTGG
Gpx1	CCTCAAGTACGTCCGACCTG	CAATGTCGTTGCGGCACACC
Sod1	ATGGGTTCCACGTCCATCAGTA	TGCCCAGGTCTCCAACA
Sod2	GAGAATCTCAGTGCTCACTCGTGTC	GGAACCCTAAATGCTGCCAGTC
Ppargc1a (Pgc-1a)	AAACCACACCCACAG GATCAG	TCTTCGCTTTATTGCTCGA
Tfam	CACCCAGATGCAAAACTTTCAG	CTGCTCTTTATACTTGCTCACAG
Keap1	TGGCCAAGCAAGAGGAGTTC	GGCTGATGAGGGTCACCAGTT
Nfe2l2 (Nrf2)	CGAGATATACGCAGGAGAGGTAAGA	GCTCGACAATGTTCTCCAGCTT
Gapdh	ACGACCCCTTCATTGACCTC	TTCACACCCATCACAAACAT

Cat, catalase; Gapdh, glyceraldehyde phosphate dehydrogenase; Gclc, glutamate-cysteine ligase catalytic subunit; Gclm, glutamate-cysteine ligase modulating subunit; Gpx1, glutathione peroxidase; Keap1, kelch-like ECH-associated protein; Nfe2l2 (Nrf2), nuclear factor erythroid 2 like 2; Ppargc1a (Pgc-1a), peroxisome proliferator-activated receptor-γ coactivator 1-α; Sod1/Sod2, superoxide dismutase 1/2; Tfam, transcription factor A mitochondrial.

Statistics.

One- or two-way analysis of variance (ANOVA) models were used to examine differences between groups. For one-way ANOVA, time, pharmacological inhibitor of contraction (e.g., BTS or TTX), or chemotherapy were utilized as between-group factors. In two-way ANOVA, chemotherapy and STIM/stretch/time were used as between-group factors. If an interaction effect was noted, pairwise comparisons (least significant difference) were performed to identify the location of group differences. Relationships between variables were determined by linear regression analysis. All data were analyzed using GraphPad Software (v7; La Jolla, CA) and are presented as means ± SE, unless otherwise stated.

RESULTS

Characterization of C2C12 myotubes and STIM.

C2C12 myotubes were differentiated for 7 days before treatment with chemotherapy and/or STIM was started. Under our culture conditions, there was a rapid growth and accumulation of myosin between d5 and d7 (Fig. 1, A and B), followed by myosin content and myotube size reaching a plateau. At d7, myotubes express myofilament proteins, such as myosin (Fig. 1, A and C), actin (Fig. 1A), and α-actinin (data not shown), and organize these proteins into myofilaments (Fig. 1C). These myofilaments are functional, as are other components of the excitation-contraction coupling system, as STIM of d7 myotubes provokes intracellular Ca²⁺ cycling (Fig. 1, D–F, and Supplemental Video S1; all Supplemental Material for this article is available at https://doi.org/10.6084/m9.figshare.8049776.v1) and contraction (Fig. 1, E and F, and Supplemental Video S2). Moreover, calcium release and reuptake were synchronized with the electric pulses and with contraction and relaxation of the myotube (Fig. 1E), respectively. Finally, the degree of fractional shortening tracked with the magnitude of the intracellular Ca²⁺ transient (Fig. 1F).

Fig. 1.

C2C12 murine myotube morphology, protein expression, Ca²⁺ cycling, and contraction. A: content and morphology of myotubes at day 3 postdifferentiation (d3), d5, d7, d8.5, and d10 are shown (scale bar = 25 µm; top) and following fixation and staining of actin with tetramethylrhodamine isothiocyanate-labeled phalloidin (scale bar = 200 µm; middle), with corresponding representative gel bands for fast-twitch myosin expression at each time point (bottom). B: average myotube myosin expression (△) and diameters (○) are shown for d3 (n = 9 wells and 107 myotubes, respectively), d5 (n = 9 wells and 116 myotubes, respectively), d7 (n = 9 wells and 177 myotubes, respectively), d8.5 (n = 9 wells and 186 myotubes, respectively), and d10 (n = 9 wells and 146 myotubes, respectively). Differences between days are denoted with different letters (all differences are P < 0.01 using 1-way ANOVA). C: myofilament formation in d7 myotube using anti-fast-twitch myosin antibody (green) in d7 myotubes (scale bar = 25 µm). D: pseudocolor images of intracellular Ca²⁺ cycling in d7 myotube measured by Fluo-2 AM in response to 20-V, 12-ms pulses at 1 Hz. Time signatures for each image are provided at the bottom of each panel (frame rate = 40/s, scale bar = 10 µm). The time sequence starts in the upper left corner image and runs from right to left, ending in the bottom right image. E: raw Fluo-2 AM fluorescence signal, showing intracellular Ca²⁺ release (positive excursion) and reuptake (negative excursion; panels at top) and contractions (negative excursion) and relaxations (positive excursion; panels at bottom) for several excitation-contraction cycles (left) and the average of 15 contraction cycles (right) in response to 20-V, 12-ms pulses at 1 Hz. F: relationship between the relative magnitude of intracellular Ca²⁺ cycling and contraction measures simultaneously in d7 myotubes (n = 8 myotubes). A.U., arbitrary units.

Myotube contractility was quantified using two techniques. The first involved manual selection and automated tracking of two cellular landmarks on myotubes using the edge-tracking feature of the IonWizard software. Data from these measurements showed that myotubes contracted with a fractional strain of 3.9 ± 2.7% (n = 13 myotubes; ○ in Fig. 2C). These measurements, however, suffer bias from manual selection of landmarks and limitations in the precision of edge tracking. To address these limitations, we developed a novel pixel-tracking image analysis approach (Fig. 2, A and B) that minimizes these biases. This approach yielded a slightly lower average fractional strain of 2.6 ± 1.0% (n = 17 myotubes; ● in Fig. 2C). Each technique has strengths and weaknesses, but their results generally agreed. Thus, we calculated an average fractional strain from all data, which was 3.2 ± 2.0% (Fig. 2C). That our new pixel-tracking approach improved the precision of measurements is suggested by the fact that the variance in the measures by edge detection was greater (P < 0.01) than that by our new pixel-tracking approach (via Levene’s test). This improvement is reflected by the fact that the coefficient of variation for our new approach is 40% lower than that of the edge-tracking approach. Of note, myotubes were not chosen for size or the density of surrounding tubes, and we cannot rule out that contractility is affected by these parameters.

Myotube contraction with STIM was sensitive to modulators of excitation-contraction coupling (ECC) and myosin-actin cycling. Application of the Na⁺ channel blocker TTX, which prevents membrane depolarization and all downstream ECC events, eliminated myotube intracellular Ca²⁺ cycling (P < 0.05) and contraction (P < 0.01; Fig. 2D and Supplemental Fig. S1A). Additionally, application of BTS, an inhibitor of myosin-actin interaction/ATPase activity, eliminated contraction (P < 0.01) without diminishing Ca²⁺ cycling (Fig. 2D and Supplemental Fig. 1B).

STIM prevents DOX-induced myosin loss.

We use myotube myosin content as an index of myotube size/atrophy, based on its concordance with myotube diameter (Fig. 1B) and our observation of similar relative reductions in both indexes in response to DOX (25). A DOX-by-STIM interaction effect (P < 0.01) was observed for myosin content, such that DOX caused a reduction in myosin content (P < 0.01), whereas 1 h of STIM each day during DOX administration prevented loss of myosin (Fig. 3A). STIM alone did not alter myotube myosin content relative to control. To confirm that STIM-induced membrane depolarization and subsequent intracellular Ca²⁺ cycling and contraction were required for this protective effect of STIM, we treated myotubes with TTX, as this eliminates intracellular Ca²⁺ cycling and contraction (Fig. 2D and Supplemental Fig. S1A). Treatment with TTX prevented the effects of STIM on myosin content (Fig. 3B; P < 0.01 DOX effect).

Fig. 3.

Effects of doxorubicin (DOX; 0.2 µM; gray bars) or vehicle [DMSO; control (CTRL); open bars] administration, with or without electrical field stimulation (±STIM; 20-V, 12-ms pulses at 1 Hz for 1 h/day), on myotube myosin content, protein synthesis, markers of proteolysis, and signaling molecule expression/phosphorylation. A: DOX administration (3 days) reduces myosin content, and STIM prevents this reduction (n = 12 per bar; P < 0.01 DOX-by-STIM interaction). B: application of tetrodotoxin (TTX; 10 µM) during 1-h STIM bouts mitigates its protective effects on myotube myosin content (n = 6 per bar; P < 0.01 DOX effect). C: representative gel image of protein synthesis measurements using incorporation of puromycin into protein, as described in methods, for a subset of replicates in 1-day DOX/STIM experiments. D: DOX administration (1 day) reduces myotube protein synthesis, and STIM fails to prevent this reduction (n = 6 per bar; P < 0.01 DOX effect and P < 0.05 STIM effect). E: DOX administration (1 day) reduces phosphorylation of p70 S6 kinase (p-p70S6K; S308), and STIM fails to prevent this reduction (n = 6 per bar; P < 0.01 DOX effect). F: DOX administration (1 day) increases muscle RING finger 1 (MuRF1) expression, and STIM prevents this increase (n = 6 per bar; P < 0.01 DOX-by-STIM interaction). G: DOX administration (1 day) decreased forkhead box O3a (FoxO3a) phosphorylation (p-FoxO3a; S253), and STIM prevents this decrease (n = 6 per bar; P < 0.01 DOX-by-STIM interaction). H: phosphorylation of Akt (pAkt; S308) was higher in STIM+DOX compared with all groups (P < 0.05), except DOX, where P < 0.01 (n = 6 per bar; P < 0.05 DOX-by-STIM interaction). Representative gel images are shown at the top of A, B, E, F, G, and H for a subset of replicates. All data are means ± SE, with individual data points shown with each bar. *P < 0.05 and **P < 0.01 DOX effects (B, D, and E), STIM effects (D), or DOX-by-STIM interaction effects with a single group differing from other groups (A, F, G, and H) using two-way ANOVA.

DOX and STIM effects on protein synthesis and proteolytic mediators.

To examine the mechanism by which STIM prevented DOX-induced myosin loss, we measured acute (1-day) modulation of protein synthesis and expression of MuRF1, an E3 ubiquitin ligase responsible for ubiquitination of myosin (9). DOX treatment reduced protein synthesis rate (Fig. 3, C and D; P < 0.01 DOX effect) and phosphorylation of p70 S6 kinase (T389; Fig. 3E; P < 0.01 DOX effect), and STIM was unable to preserve either protein synthesis or p70 S6 kinase phosphorylation in DOX-treated cells (Fig. 3, C–E, respectively).

A DOX-by-STIM interaction effect was noted for MuRF1 expression (P < 0.01), such that DOX increased MuRF1 protein (P < 0.01) and STIM prevented the DOX-induced increase in MuRF1 expression, whereas STIM alone did not alter MuRF1 (Fig. 3F). A DOX-by-STIM interaction was also noted for FoxO3a phosphorylation (S253; P < 0.01), where DOX reduced FoxO3a phosphorylation and this reduction was prevented by STIM (P < 0.01), whereas STIM alone did not alter FoxO3a phosphorylation (Fig. 3G). At this same time point (1 day), STIM reduced Akt phosphorylation (S308; Supplemental Fig. S2, P < 0.01 STIM effect), but no DOX or DOX-by-STIM interaction effects were noted. Because regulation of Akt may be transitory (59), we explored Akt phosphorylation early following the first STIM bout (1 h; Fig. 3H). At this time point, we found a DOX-by-STIM interaction effect (P < 0.05). Pairwise comparisons showed that DOX+STIM was higher than all other groups (all P < 0.05, with the exception of P < 0.01 for DOX).

Effects of other chemotherapeutics on protein synthesis and MuRF1 expression.

We explored whether other commonly used chemotherapeutics that cause myotube myosin loss (25) alter protein synthesis and MuRF1 expression similarly to DOX at 1 day posttreatment. Results showed that both CIS and TAXOL upregulate MuRF1 expression (P < 0.01 and P < 0.05; Fig. 4A) and reduce protein synthesis (P < 0.01 for both; Fig. 4, B and C), similar to DOX (P < 0.01; Fig. 4, B and C), suggesting a common effect of numerous chemotherapeutics on myotube protein metabolism.

Fig. 4.

Effects of 1 day of administration of common chemotherapeutics [doxorubicin (DOX), cisplatin (CIS), or paclitaxel (TAXOL)] or vehicle [DMSO; control (CTRL); open bars] on muscle RING finger 1 (MuRF1) expression and protein synthesis. A: all chemotherapeutics increased MuRF1 expression (n = 6 per bar). B: representative gel image of protein synthesis measurements using incorporation of puromycin into protein, as described in methods, for a subset of replicates with chemotherapy administration. C: administration of all chemotherapeutics reduced myotube protein synthesis (n = 6 per bar). All data are means ± SE, with individual data points shown with each bar. *P < 0.05 and **P < 0.01 compared with control conditions via one-way ANOVA.

Role of contraction and cell stretch in mediating effects of STIM.

To examine whether mechanical contraction and myotube shortening explain the protective effects of STIM, we treated myotubes with BTS, an inhibitor of myosin-actin interaction that prevents myotube contraction but does not disrupt intracellular Ca²⁺ cycling (Fig. 2D and Supplemental Fig. S1B), during 1-h STIM sessions. BTS blocked the effect of STIM to prevent DOX-induced myosin loss over 3 days of treatment (Fig. 5A; P < 0.01 DOX effect). To clarify whether the effects of STIM were mediated via mechanotransductive signaling, we mechanically stretched cells (5%) for 1 h during DOX or vehicle treatments. The magnitude of the stretch stimulus was chosen to match the relative strain of myotubes induced by STIM (Fig. 2C). DOX-by-stretch interaction effects were found for MuRF1 (P < 0.01) and FoxO3a (S253; P < 0.05). Pairwise comparisons showed that similar to STIM experiments (Fig. 3), DOX stimulated MuRF1expression (Fig. 5B; P < 0.01) and reduced FoxO3a phosphorylation (S253; P < 0.05; Fig. 5C), and mechanical stretch prevented these alterations (Fig. 5, B and C). Stretch alone did not alter MuRF1 expression or FoxO3a phosphorylation. In contrast to STIM, however, mechanical stretch remediated DOX-induced reductions in protein synthesis (Fig. 5, D and E; P < 0.01 DOX-by-stretch interaction). These changes in protein synthesis were generally paralleled by altered p70 S6K phosphorylation (Fig. 5F), but the interaction effect was not significant (P < 0.01 stretch effect and P < 0.01 DOX effect).

Fig. 5.

Effects of doxorubicin (DOX; 0.2 µM; gray bars) or vehicle [DMSO; control (CTRL); open bars] administration, with or without electrical field stimulation (±STIM; 20-V, 12-ms pulses at 1 Hz for 1 h/day) and application of pharmacological inhibitor of myosin-actin contraction/ATPase (1 h/day during STIM), or mechanical stretch (±Stretch; 5% for 1 h/day), on myotube myosin content, protein synthesis, and markers of proteolysis. A: application of N-benzyl-p-toluene sulfonamide (BTS; 50 µM) during STIM bouts over 3 days of DOX administration mitigates the protective effects of STIM on myotube myosin content (n = 6 per bar; P < 0.01 DOX effect). B and C: application of 5% stretch prevents increases in muscle RING finger 1 (MuRF1) expression (n = 6 per bar, except DOX, where n = 7 for all stretch experiments; P < 0.01 DOX-by-stretch interaction; B) and reductions in phosphorylation of forkhead box O3a (p-FoxO3a; S253; P < 0.05 DOX-by-stretch interaction; C) associated with 1 day of DOX administration. For C, DOX was lower than all other groups (P < 0.05), and DOX+stretch was higher than all groups at P < 0.05, except DOX, where P < 0.01. D: representative gel image of protein synthesis measurements using incorporation of puromycin into protein, as described in methods, for a subset of replicates with 1-day DOX administration with or without 5% mechanical stretch for 1 h. E: DOX administration reduces protein synthesis (P < 0.01 DOX-by-stretch interaction). F: for phosphorylation of p70 S6 kinase (p-p70S6K; T308), DOX (P < 0.01) and stretch (P < 0.01) effects were significant, but not DOX-by-stretch interaction effect. Data are means ± SE, with individual data points shown with each bar. *P < 0.05 and **P < 0.01 DOX effects (A and F), stretch effects (F), or DOX-by-stretch interaction effects (B, C, and E), with pairwise differences detailed above, as determined by two-way ANOVA.

Activation of transient receptor potential cation channel subfamily V member 1 with capsaicin does not mimic effects of STIM on myotube myosin content.

Recent work identified activation of transient receptor potential cation channel subfamily V member 1 (TRPV1) channels as integral to the muscle anabolic response to mechanical stimuli, that the anabolic response to stress was mimicked by the TRPV1 agonist capsaicin (CAP), and that CAP could prevent the catabolic effects of unloading and denervation (34). Because the effects of STIM were dependent on myotube contraction (i.e., mechanotransduction; Figs. 3B and 5A) and were mimicked by cell stretch (Fig. 5, B and C), we reasoned that CAP administration may mimic the effects of STIM. Initially, we performed a dose-response experiment to examine the effects of CAP (50 nM to 1 µM) for 3 days (d7–d10) on myotube myosin content. To our surprise, we found no effect of CAP on myosin content at any concentration. We chose a dose of 100 nM CAP because prior work showed that this was sufficient to activate anabolic pathways [e.g., p70 S6 kinase (34)]. CAP (100 nM) did not mitigate the effect of DOX administration to reduce myotube myosin protein content (Fig. 6B). To further interrogate whether a higher dose of CAP could mimic mechanotransductive signaling and protect against DOX-induced atrophy, we also treated cells with 1 μM CAP. However, this higher dose of CAP similarly did not mitigate the effects of DOX to reduce myosin content (Fig. 6C). Finally, 100 nM CAP showed a DOX-by-CAP interaction effect (P < 0.01) on ROS production, but not with directionality that would be expected to mitigate myotube myosin loss. DOX increased ROS production (P < 0.01), with an additive effect when combined with CAP (Supplemental Fig. S3A). We found an effect of DOX (P < 0.01) to reduce mitochondrial content regardless of CAP treatment (Supplemental Fig. S3B).

Fig. 6.

Effects of capsaicin (CAP) administration or doxorubicin (DOX; 0.2 µM) administration with or without CAP administration on myotube myosin content. A: dose-response effects of 3 days of CAP administration (50, 100, 250, or 500 nM or 1 µM) vs. DMSO vehicle control (CTRL) on myotube myosin content (n = 3 per bar; 1-way ANOVA). B: treatment with CAP (100 nM) was unable to prevent the effect of 3 days of DOX administration to reduce myotube myosin protein content (n = 6 per bar; P < 0.01 DOX effect; two-way ANOVA). C: similar results were found with administration of high-concentration CAP (1 μM; n = 6 per bar, except control bars, which are n = 3 per bar; P < 0.01 DOX effect; 2-way ANOVA). Representative gel images are shown at the top of A, B, and C for a subset of replicates. Note that the representative gel at the top of the panel for a subset of replicates was run on a 7.5% gel, which caused separation of myosin isoforms. Both bands were quantified. Data are means ± SE, with individual data points shown with each bar. **P < 0.01 DOX effects.

Myotube mitochondrial ROS production, content, and antioxidant gene expression.

As the TRPV1 agonist CAP did not mitigate the effects of DOX, we sought other potential mechanisms underlying the protective effects of STIM on myotube myosin content. We recently showed that DOX treatment for 3 days increased myotube mitochondrial ROS production and reduced mitochondrial content and that mitochondrial targeted antioxidant treatment prevented DOX-induced ROS production, mitochondrial loss, and reductions in myotube myosin content (25). As exercise modulates mitochondrial biology and redox balance, we examined the effects of STIM on DOX-induced changes in mitochondrial content, ROS production, and antioxidant gene expression. One day of DOX treatment increased ROS production, and these effects were maintained after 3 days of DOX treatment (Fig. 7A; P < 0.01 DOX effect). Unlike ROS production, 1 day of DOX treatment did not reduce mitochondrial content, whereas a reduction was apparent after 3 days of DOX treatment (Fig. 7B; P < 0.01 DOX-by-time interaction; P < 0.05 3-day control vs. 3-day DOX).

Fig. 7.

Effects of doxorubicin (DOX; 0.2 µM; gray bars) or vehicle [DMSO; control (CTRL); open bars] administration, with or without electrical field stimulation (±STIM; 20-V, 12-ms pulses at 1 Hz for 1 h/day) or a 5% static stretch, on mitochondrial reactive oxygen species (ROS) production, mitochondrial content, and antioxidant gene expression. A: 1 and 3 days of DOX administration increase myotube mitochondrial ROS production per unit mitochondria measured with MitoSox and MitoTracker fluorescent dyes, respectively (n = 12 per bar; P < 0.01 DOX effect). B: 3 days of DOX administration, but not 1 day, reduce myotube mitochondrial content measured by MitoTracker fluorescent dye (n = 22 per bar; P < 0.05 DOX-by-time interaction). C: DOX (P < 0.01) and STIM (both STIM and DOX+STIM; P < 0.01) increased mitochondrial ROS production during 3 days of DOX administration relative to controls (n = 6 per bar; P < 0.01 DOX-by-STIM interaction). D: DOX induced mitochondrial loss (P < 0.01), whereas STIM prevented DOX-induced mitochondrial loss (n = 6 per bar; P < 0.05 DOX-by-STIM interaction). E: 3 days of STIM upregulated peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) alone and with DOX treatment (P < 0.05 STIM effect), whereas DOX increased transcription factor A mitochondrial (TFAM; n = 6 per bar; P < 0.05 DOX effect). F: no DOX-by-STIM interaction for 1 day, but a STIM effect (P < 0.05) was noted on phosphorylation of AMPK (pAMPK; T172; n = 9 per bar). G: DOX administration (1 day) increases forkhead box O3a (FoxO3a) phosphorylation (p-FoxO3a; S413; P < 0.05), and STIM prevents this increase (n = 6 per bar; P = 0.06 DOX-by-STIM interaction). H: DOX increased FoxO3a phosphorylation (S413; P < 0.01), and mechanical stretch was unable to fully reduce this increase (n = 6 per bar, except DOX, where n = 7; P < 0.05 DOX-by-stretch interaction; P < 0.01 DOX higher than −Stretch, P < 0.05 DOX higher than +Stretch). Data are means ± SE, with individual data points shown with each bar, with the exception of gene expression data. *P < 0.05 and **P < 0.01 for DOX or DOX-by-time/STIM/stretch interaction effects, with pairwise differences detailed above using two-way ANOVA.

A STIM-by-DOX interaction effect was found on ROS production (P < 0.01), such that STIM did not prevent DOX-induced upregulation of ROS production at 3 days (Fig. 7C). In fact, both STIM groups showed increased ROS production at 3 days compared with non-STIM groups (P < 0.01; Fig. 7C). DOX alone also increased ROS production (P < 0.01). A STIM-by-DOX interaction was noted for mitochondrial content (P < 0.05), as STIM prevented DOX-induced reductions in mitochondrial content (Fig. 7D). This effect of STIM to preserve mitochondrial content was accompanied by increased expression of Pgc-1α mRNA (P < 0.05 STIM effect; Fig. 7E), although no effect of STIM was noted on transcription factor A mitochondrial (Tfam) expression. DOX upregulated Tfam expression (P < 0.01 DOX effect).

Energetic insufficiency brought about by DOX-induced mitochondrial loss could contribute to myosin loss via upregulation of AMPK and activation of FoxO3a/MuRF1 (57). In this context, the protective effects of STIM could derive from its ability to maintain mitochondrial content (Fig. 7D) and/or function. To begin to explore this possibility, we evaluated AMPK (T172) phosphorylation 1 h and 1 day following the start of treatments. We found no DOX or DOX-by-STIM interaction effects on AMPK phosphorylation after 1 h (1.5 h post-DOX/1 h post-STIM; Supplemental Fig. S4), but there was a STIM effect (P < 0.05) 1 day after the start of treatments (Fig. 7F). Thus, the pattern of AMPK phosphorylation did not match the pattern of changes in mitochondrial content (Fig. 7D) or MuRF1 expression (Fig. 3F). To clarify whether there was any effect of DOX, or other chemotherapeutics, to impair mitochondrial function in ways that would activate AMPK, we measured myotube respiration 1 day following chemotherapy administration. We chose this time point because this is when metabolic/signaling alterations (i.e., protein synthesis, FoxO, and MuRF1 expression) are apparent. Relative to control (i.e., vehicle) conditions, we found no effect of any chemotherapeutics on basal respiration (DOX, 92 ± 9%; CIS, 94 ± 10%; TAXOL, 85 ± 6%), maximal respiration (DOX, 103 ± 9%; CIS, 99 ± 9%; TAXOL, 92 ± 6%), calculated ATP production (i.e., basal respiration minus proton leak and nonmitochondrial oxygen consumption; DOX, 93 ± 9%; CIS, 98 ± 11%; TAXOL, 89 ± 6%), or calculated spare respiratory capacity (maximal respiration minus basal respiration; DOX, 110 ± 10%; CIS, 103 ± 10%; TAXOL, 98 ± 7%). Moreover, none of the other measured or calculated variables available via this analysis differed between chemotherapeutics and controls (data not shown). These results further buttress the lack of phosphorylation of AMPK at 1 day following DOX treatment and suggest that signaling of DOX and other chemotherapeutics to upregulate MuRF1 expression is likely not via impaired mitochondrial respiration.

Interestingly, however, we found that FoxO3a phosphorylation at S413, a site phosphorylated by AMPK (24), 1 day following the start of DOX treatment showed a strong trend (P = 0.06) toward a DOX-by-STIM interaction effect. Pairwise comparisons showed that there was an increase (P < 0.05) in FoxO3a phosphorylation with DOX and that STIM prevented this increase. To further examine mechanosensitive control of FoxO3a, we examined the effect of mechanical stretch on this phosphorylation site at 1 day posttreatment and found a DOX-by-stretch interaction effect (P < 0.05; Fig. 7H). DOX increased FoxO3a phosphorylation relative to control and stretch control (P < 0.01 and P < 0.05, respectively). However, stretch did not fully remediate the increase in FoxO3a phosphorylation. In fact, FoxO3a phosphorylation was increased in the DOX+stretch group compared with control (P < 0.05).

Finally, we examined whether STIM upregulated antioxidant gene expression, as recent studies have suggested this possibility (29) and because we previously showed that mitochondrial targeted antioxidant prevents myotube myosin loss with DOX administration (25). DOX (P < 0.05) and STIM (P < 0.01) effects were found for expression of the redox-sensitive transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2), whereas no effects were observed for kelch-like ECH-associated protein (KEAP1; Fig. 8A). DOX effects were observed for a number of antioxidant genes, including SOD1, SOD2, CAT, and glutamate-cysteine ligase-modulating subunit (GCLM; all P < 0.01; Fig. 8, B and C). DOX-by-STIM interaction effects (both P < 0.05) were found for glutamate-cysteine ligase catalytic subunit (GCLC) and glutathione peroxidase (GPx1; Fig. 8C). In the former, DOX increased GCLC expression relative to all other groups (P < 0.05 to P < 0.01). In the latter, DOX increased GPx1 expression relative to all other groups (P < 0.05). As PGC-1α has been shown to mediate many of the beneficial effects of muscle contraction/exercise, we further examined whether protein levels were potentiated by STIM. We found both DOX (P < 0.05) and STIM (P < 0.01) effects, but no interaction effect (Fig. 8D). To further examine whether there are any effects of STIM on protein expression of these analytes, we measured SOD2 and catalase (Fig. 8, E and F). These proteins were chosen because they showed some of the strongest potentiation by STIM (Fig. 8B) and because prior work has shown that these antioxidant systems are upregulated by exercise and may be responsible for the effects of exercise to counter DOX-induced muscle atrophy (50, 54). We found evidence for a DOX-by-STIM interaction effect on catalase (P < 0.05), with DOX (P < 0.01), STIM (P < 0.05), and combined DOX+STIM (P < 0.05) increasing catalase expression relative to controls. In SOD2, only a DOX effect (P < 0.05) was noted.

Fig. 8.

Effects of doxorubicin (DOX; 0.2 µM; gray bars) or vehicle [DMSO; control (CTRL) open bars] administration, with or without electrical field stimulation (STIM; 20-V, 12-ms pulses at 1 Hz for 1 h/day), on antioxidant gene expression. A: 3 days of STIM and DOX both increased nuclear factor erythroid 2-related factor 2 (Nrf2) expression (n = 6 per bar; P < 0.05 DOX effect and P < 0.01 STIM effect) but had no effect on kelch-like ECH-associated protein (KEAP1; n = 3 per bar). B: DOX increased superoxide dismutase 1 (SOD1), SOD2, and catalase (CAT) expression, independent of STIM (n = 6 per bar; P < 0.01 DOX effect). C: DOX upregulated glutamate-cysteine ligase modulating subunit (GCLM), independent of STIM (P < 0.01 DOX effect), and DOX alone upregulated glutamate-cysteine ligase catalytic subunit (GCLC) and glutathione peroxidase (GPx1; P < 0.05 DOX-by-STIM interaction; n = 6 per bar). D: there were DOX (P < 0.05) and STIM (P < 0.01) effects on peroxisome proliferator-activated receptor-γ coactivator 1-α (PGC-1α) expression (n = 5 per group, except for n = 6 for DOX). E: there was a significant DOX-by-STIM interaction for SOD2 (P < 0.01; n = 9 per group) with the increase with DOX treatment being significant (P < 0.01). F: both DOX and STIM upregulate catalase (P < 0.05 DOX-by-STIM interaction; n = 9 per group) with an increase with DOX (P < 0.01), STIM (P < 0.05), and DOX+STIM (P < 0.05). Data are means ± SE, with individual data points shown with each bar, with the exception of gene expression data. *P < 0.05 and **P < 0.01 for DOX or DOX-by-STIM interactions, with pairwise comparisons noted above using two-way ANOVA.

DISCUSSION

How muscle contraction protects skeletal muscle against the detrimental effects of chemotherapeutics (41, 53, 66, 67) is unclear. Using in vitro muscle contraction models, we show that the atrophic effects of DOX are prevented by daily bouts of electrically induced contraction. STIM prevented DOX-induced downregulation of Akt and FoxO3a phosphorylation and increased MuRF1 expression, but it was unable to recover protein synthesis. STIM effects were mediated via mechanotransductive signaling, as pharmacological inhibition of myosin-actin interaction/ATPase abolished these effects and mechanical stretch mimicked the effects of STIM. Finally, STIM upregulated PGC-1α and prevented the loss of mitochondrial content. Our findings suggest that muscle contraction counters the catabolic effects of chemotherapy through effects to downregulate proteolysis and preserve mitochondrial content via mechanotransductive signaling pathways.

Contrary to our original hypothesis, the most likely mechanism to explain the beneficial effects of STIM is reductions in proteolysis, as STIM downregulated DOX-induced increases in MuRF1 expression, whereas protein synthesis was not rescued. How STIM modulates MuRF1 expression is suggested by Akt and FoxO3a phosphorylation data. Prior studies show that muscle stretch and contraction upregulate Akt activation (59), and our data advance these results to show that STIM-induced myotube contraction counters the effects of catabolic mediators, such as DOX, to reduce Akt phosphorylation. FoxO3a transcriptional effectiveness is regulated, in part, by its cytoplasmic-nuclear localization, which is regulated via targeted phosphorylation, particularly by Akt (62). Correspondingly, increased Akt phosphorylation with STIM was associated with similar preservation of FoxO3a phosphorylation on the Akt site (S253). The effects of STIM to prevent DOX-induced increases in MuRF1 further suggest that these signaling events likely reduced FoxO3a transcriptional activity. Supporting the relevance of our results for in vivo skeletal muscle are data showing that electrical stimulation of skeletal muscle prevents disuse-induced atrophy, in part, via downregulation of MuRF1 expression (14).

One strength of our in vitro contraction model is that it allows us to discern whether STIM mediates these effects via mechanotransductive or Ca²⁺-chemotransductive signaling pathways. BTS application during 1-h daily STIM bouts prevented the ability of STIM to preserve myotube myosin content in the face of DOX, implicating mechanotransductive pathways in the beneficial effects of STIM. This notion was reinforced by our results from mechanical stretch experiments, which showed similar upregulation of FoxO3a phosphorylation and decreased MuRF1 expression. These findings concur with studies showing that muscle stretch upregulates Akt (2) and reduces nuclear FoxO content and DNA binding (52). Such a mechanism may be operative in vivo, as recent studies show that muscle stretch diminished atrophy associated with an experimental intensive care unit stay via downregulation of MuRF1 expression, whereas it did not alter protein synthesis (56). Collectively, these data extend the role of muscle mechanotransductive signaling to include protection against catabolic stimuli and suggest that this occurs through suppression of proteolysis, rather than preservation or upregulation of protein synthesis (75).

Building on this last point, although we attempted to match mechanical stretch to STIM by utilizing a similar relative strain, the two stimuli differed with respect to their effects on protein synthesis. Unlike STIM, stretch prevented the DOX-induced reduction in protein synthesis. The more pronounced effects of stretch on myotube protein synthesis may relate to the fact that the mechanical stretch stimulus (5%) was slightly higher than the fractional strain during STIM (~3%). We limited the stretch to 5% because the stretch device has not been validated below 5%. Additionally, STIM imparts its mechanotransductive effects through shortening of the cell, whereas stretch is an elongation stimulus. More specifically, STIM produces stress/strain on myotubes primarily via shortening along their longitudinal axis (Supplemental Video S2), in-line with the orientation of myofilaments (Fig. 1C). In contrast, the uniaxial mechanical stretch model we employed would cause multiaxial lengthening and, in turn, stress/strain, because myotubes grow in random orientations relative to the axis of stretch. This may explain differences between the two stimuli, as transverse stretch is more anabolic than uniaxial stretch (30, 43). Extrapolating from these results, one might hypothesize that a more intense mechanical stimulus than that provided by STIM may more effectively counter the deleterious effects of DOX by also preserving protein synthesis (75). Regardless of differences between STIM and stretch conditions on protein synthesis, the similarity in their ability to prevent DOX-induced modifications in Akt/FoxO3a/MuRF1 signaling underscores the importance of exercise-induced downregulation of proteolysis in prevention of chemotherapy-induced atrophy via mechanotransductive pathways.

Numerous mechanotransductive signaling pathways could mediate the effects of STIM/stretch. We chose to examine the effect of TRPV1 antagonism with CAP because recent reports suggested that activation of TRPV1 with CAP promoted muscle anabolism and was sufficient to prevent unloading- and denervation-induced atrophy (34, 35). Moreover, activation of TRPV1 channels by CAP increases mitochondrial content and function in C2C12 myotubes and mouse muscle in vivo (46). However, we found that treatment of myotubes with CAP did not prevent DOX-induced atrophy, nor did it counter DOX-induced oxidant production or mitochondrial loss. Although we used a concentration of CAP (100 nM) shown to promote muscle anabolism/anticatabolism (34, 35) and mitochondrial adaptations (46), higher concentrations (1 µM) similarly failed to prevent DOX-induced myosin loss. In fact, we found no evidence for effects of CAP on myotube myosin content throughout a range of doses. Thus, our results do not support the contention that the beneficial effects of STIM in our model are mediated via mechanotransductive signaling through activation of TRPV1 channels.

In light of the role for mitochondrial oxidant production and/or rarefaction in the deleterious effects of DOX and the protective effects of reducing oxidant stress (25), we examined whether the effects of STIM are explained by effects on mitochondrial content and ROS production. STIM preserved mitochondrial content, an effect accompanied by increased PGC-1α expression, a key mediator of mitochondrial remodeling that is upregulated in response to muscle contraction and that protects against disuse-induced mitochondrial loss and atrophy (40, 61). In keeping with studies overexpressing PGC-1α (5, 39), STIM may prevent DOX-induced atrophy, in part, through effects on mitochondrial dynamics to maintain mitochondrial content and/or function. Energetic insufficiency brought about by mitochondrial loss would be expected to upregulate AMPK, which can phosphorylate FoxO3a (24), leading to transcription of MuRF1. However, we found no evidence for an effect of DOX to impair mitochondrial oxygen consumption, and the pattern of AMPK phosphorylation (T172) did not mirror the effects of DOX or STIM on Akt/FoxO3a/MuRF1 activation/expression. Despite this, FoxO3a phosphorylation at S413, which is regulated by AMPK, was increased by DOX and reduced to control levels by STIM. In the absence of concordance between AMPK and FoxO3a phosphorylation and impaired mitochondrial function, we posit that STIM-induced upregulation of PGC-1α may contribute to inhibition of FoxO3a S413 phosphorylation and MuRF1 expression (27, 57, 61) and, in turn, myotube myosin loss. However, as PGC-1α was also upregulated by DOX, this does not solely explain the beneficial effects of STIM.

As prior work from our laboratory and others suggests that improved mitochondrial redox balance prevents the deleterious effects of DOX (21, 22, 25), the beneficial effects of STIM could also be explained, in part, by its ability to reduce ROS production and/or increase antioxidant capacity. Contrary to the former possibility, however, STIM did not diminish the DOX-induced increase in mitochondrial ROS production. In fact, in keeping with prior work (60), STIM-induced myotube contraction increased ROS production (Fig. 7C), which may partially account for the effects of STIM to upregulate Akt (26) and PGC-1α (65). We acknowledge that STIM could increase ROS production at other sites throughout the cells that we did not measure. However, work by our laboratory and others suggests the importance of mitochondrial oxidant production in promoting myotube atrophy with DOX (22, 25).

Contrary to recent reports (29), we found no effect of STIM on antioxidant gene expression, with the predominant effect being related to DOX. This may be explained by our milder STIM protocol. In our model, ROS production, which is an impetus for antioxidant gene expression, was increased ~20% over baseline (Fig. 7C), whereas in the study by Horie et al. (29), STIM increased ROS production ~100% over baseline. Measurements of mRNA, however, did not reflect the effects of STIM on protein expression, as we found effects of DOX and STIM to upregulate SOD2 and catalase, two antioxidant enzymes that are suggested to contribute to protective effects of exercise on DOX-mediated muscle adaptations (50, 54). Importantly, however, we found no additive effect of combined DOX and STIM on either SOD2 or catalase. Thus, although STIM upregulated antioxidant protein expression (e.g., catalase), it did not potentiate it to a greater extent than DOX alone in a manner that might explain STIM effects to suppress atrophy. A caveat to these measurements is that protein expression may not reflect enzyme activity or, in turn, antioxidant capacity. Thus, further studies will be needed to discern whether STIM alters cellular or organellar redox balance and whether these adaptations contribute to its antiatrophic effect.

How does STIM in myotube cultures relate to exercise in vivo? We are cautious in extrapolating our STIM conditions to represent specific types of exercise for several reasons. First, STIM may affect myotubes differently depending on their differentiation state. For example, earlier during differentiation, myotubes may not contain myofilaments or undergo contraction upon STIM (18). At this developmental stage, myotubes would experience Ca²⁺-chemotransductive signaling with STIM, but not mechanosensitive signaling. In our studies, where STIM is applied later in differentiation when cells express myofilaments and readily contract (Figs. 1 and 2), both Ca²⁺-chemotransductive signaling and mechanosensitive signaling are activated. Thus, the same STIM regimen could incite unique adaptations at different developmental stages. Second, there could be differences in adaptations to STIM depending on whether the myotubes are primary cultures or are derived from transformed cell lines (51). Third, caution is urged with evaluating the nature of STIM effects based solely on signaling molecule expression/posttranslational modification. Most studies examining STIM have evaluated the effects of a single bout of STIM on signaling pathways that mediate adaptations to different exercise regimens, without determining whether cells express the end phenotype of these adaptations. Finally, building on this last point, the pattern of phenotypic adaptations found with our STIM protocol does not match signaling pathways and or phenotypes of classical aerobic or resistive-type exercise. For instance, we observed increases in PGC-1α and antioxidant expression and maintenance of mitochondrial content in the face of DOX with STIM, adaptations that mimic the effects of aerobic exercise. However, we found no evidence for activation of AMPK, a hallmark of aerobic exercise thought to mediate at least some of its effects (42). Moreover, although our STIM protocol was sufficiently anabolic to prevent DOX-induced loss of myosin content, as might be expected with resistive-type exercise, it did not upregulate protein synthesis or associated translational signaling. For all of these reasons, we have not interpreted our STIM regimen as being emblematic of either aerobic or resistive exercise.

Despite our hesitance to draw parallels between our STIM regimen and exercise regimens, it recapitulated effects observed with classical exercise in vivo. For instance, Mijwel et al. (48) recently showed that both aerobic and resistive-type exercise in patients with breast cancer receiving chemotherapy prevented skeletal muscle atrophy and preserved citrate synthase activity, a marker of mitochondrial content. These antiatrophic effects of exercise were supported by studies in patients with testicular cancer, where resistance exercise tended to preserve fast-twitch muscle fiber size and, in turn, lean body mass (8). In contrast, in a mouse model of DOX administration, de Lima et al. (11) found that aerobic exercise was not sufficient to prevent muscle fiber atrophy. Additionally, activation of signaling pathways and proteolytic mediators found in our cell culture with STIM was not concordant with that observed in patients or animal models (11, 49). The reason for this disparity may lie in the timing of the measurements. Our measurements were conducted early following DOX treatment (1 day), with or without STIM, whereas these other studies made these assessments several days following DOX administration or after several months of chemotherapy, with or without exercise. Examination of these signaling molecules at an earlier time point is likely required to discern the proximal signaling events that contribute to DOX-induced atrophy or the antiatrophic effects of STIM/exercise. Despite these differences, the relative changes noted in our myotube cultures in response to DOX (i.e., ~20% reduction in myosin and ~15% reduction in mitochondrial content) translate well to those observed in humans (i.e., ~20% atrophy and ~25–30% reductions in mitochondrial content; 25, 48). Thus, our results generally agree with the effects of exercise in two human studies that found antiatrophic effects of muscle contraction/exercise in patients receiving chemotherapy, further reinforcing the potential translational value of our results.

Several limitations of our study should be acknowledged. First, responses in myotube cultures to STIM and DOX may not emulate skeletal muscle in vivo. However, myotubes expressed functional myofilaments, contracted in response to STIM, and were sensitive to pharmacological inhibitors of ECC and myosin-actin interaction/myosin ATPase similar to in vivo muscle. Moreover, adaptations in protein metabolism and signaling were similar to those demonstrated in vivo in response to DOX and exercise (41), and these patterns mimicked those in other atrophy models (56) and with the use of electrical stimulation to protect against atrophy in vivo (14, 15). Additionally, recent studies show that the myotoxic effects of DOX in cultured muscle cells emulate effects in vivo (7). Thus, although this model has limitations, it is likely sufficient to study the metabolic and signaling pathways whereby muscle contraction prevents the deleterious effects of chemotherapeutics. As an aside, we did not study the effects of DOX in vivo using a rodent model because the doses of DOX used in these models are sufficient to induce cardiac dysfunction and, in some cases, heart failure, which itself could have myopathic effects. Additionally, some of the pharmacological agents used in this model would be toxic to the animal. Second, static mechanical stretch may provoke cellular responses different from the cyclical contracting myotubes with STIM (36). However, similar results to prevent upregulation of proteolytic signals via FoxO/MuRF1 were found in both models, implicating mechanotransductive signaling in the beneficial effects of STIM. Third, the DOX treatment in myotubes is more acute (3 days) than DOX administration clinically, and the concentration may be higher than what muscle is exposed to in vivo. Despite these caveats, the magnitude of atrophy (~20% reduction) and mitochondrial loss (~15% reduction) observed in myotubes agrees reasonably well with what we (25) and others (48) have observed in patients with breast cancer receiving chemotherapy. Moreover, the beneficial effects of STIM on myotubes mimic the ability of exercise to mitigate these skeletal muscle maladaptations (48, 49). Fourth, our model evaluates the effects of STIM on myotubes that are mechanically quiescent, which does not mimic changes in muscle activity noted clinically. In patients, muscles may experience a reduction in use in response to chemotherapy (17), but there is muscle contraction/activity before cancer and treatment. We chose not to try to simulate this pre-DOX treatment muscle activity and subsequent disuse in myotubes because of the complexity of the timing of beginning and reducing STIM and questions regarding the amount and type of STIM that would be appropriate. As mentioned above, however, the effects of STIM noted in our model reflect the effects of exercise in patients (48, 49). Thus, although imperfect, our model reflects the beneficial effects of exercise observed in vivo.

In summary, the beneficial effects of contraction-induced, mechanotransductive signaling in C2C12 myotubes can prevent DOX-induced activation of the FoxO-MuRF1 pathway via downregulation of Akt activation. These effects of STIM may be mediated, in part, via upregulation of PGC-1α expression, which has been shown to inhibit FoxO3a-induced expression of MuRF1 (61) and upregulate mitochondrial content. These effects of STIM parallel the benefits of neuromuscular electrical stimulation (NMES) in humans to prevent atrophy (15). NMES maintains or improves muscle size and strength in clinical populations (38) and might be a useful exercise modality in patients with cancer during chemotherapy, as they are often unable to perform facility-based exercise programs because of fatigue or other logistical/clinical hurdles. Our results would suggest that the most important feature of NMES, or any other exercise-based intervention to offset muscle maladaptations to chemotherapy, should likely focus on mechanical stress/strain.

GRANTS

This study was funded by NIH Grants R01-AR-065826, R21-CA-191532, S10-OD-017969, and P30-RR-032135. B. A. Guigni was funded by Department of Defense Science Mathematics and Research for Transformation Scholarship 2016-85335.

DISCLOSURES

B. M. Palmer is an employee of IonOptix, LLC. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

B.A.G., J.A.C., and M.J.T. conceived and designed research; B.A.G., D.K.F., J.J.B., and B.M.P. performed experiments; B.A.G., D.K.F., J.J.B., B.M.P., and M.J.T. analyzed data; B.A.G., D.K.F., J.J.B., B.M.P., J.A.C., and M.J.T. interpreted results of experiments; B.A.G., D.K.F., B.M.P., and M.J.T. prepared figures; B.A.G., D.K.F., and M.J.T. drafted manuscript; B.A.G., D.K.F., J.J.B., B.M.P., J.A.C., and M.J.T. edited and revised manuscript; B.A.G., D.K.F., J.J.B., B.M.P., J.A.C., and M.J.T. approved final version of manuscript.