Regulatory, Integrative and Comparative Physiology

NF-κB mediates the protein loss induced by TNF-α in differentiated skeletal muscle myotubes

Yi-Ping Li, Micheal B. Reid


Nuclear factor-κB (NF-κB) regulates the transcription of a variety of genes involved in immune responses, cell growth, and cell death. However, the role of NF-κB in muscle biology is poorly understood. We recently reported that tumor necrosis factor-α (TNF-α) rapidly activates NF-κB in differentiated skeletal muscle myotubes and that TNF-α acts directly on the muscle cell to induce protein degradation. In the present study, we ask whether NF-κB mediates the protein loss induced by TNF-α. We addressed this problem by creating stable, transdominant negative muscle cell lines. C2C12 myoblasts were transfected with viral plasmid constructs that induce overexpression of mutant I-κBα proteins that are insensitive to degradation via the ubiquitin-proteasome pathway. These mutant proteins selectively inhibit NF-κB activation. We found that differentiated myotubes transfected with the empty viral vector (controls) underwent a drop in total protein content and in fast-type myosin heavy-chain content during 72 h of exposure to TNF-α. In contrast, total protein and fast-type myosin heavy-chain levels were unaltered by TNF-α in the transdominant negative cell lines. TNF-α did not induce apoptosis in any cell line, as assessed by DNA ladder and annexin V assays. These data indicate that NF-κB is an essential mediator of TNF-α-induced catabolism in differentiated muscle cells.

  • cachexia
  • cytokine
  • free radicals
  • signal transduction
  • inflammation

skeletal muscle is often impaired by diseases of other organs. Signs of muscle atrophy or wasting are frequently seen in inflammatory disorders that include cancer (27, 30), acquired immunodeficiency syndrome (33), chronic obstructive pulmonary disease (10), and congestive heart failure (2). Loss of muscle mass contributes importantly to morbidity and mortality in individuals with such diseases (2, 10, 27, 30, 33).

Among the various humoral factors that are altered in inflammatory disease, tumor necrosis factor-α (TNF-α) has been widely implicated as a possible mediator of muscle catabolism (2, 9, 10, 30,33). TNF-α is a polypeptide cytokine that is present in the serum of healthy individuals at undetectable-to-low, picogram-per-milliliter levels (24, 26, 32). Circulating TNF-α levels are elevated by inflammatory disease, reaching values as high as 2.8 ng/ml in rheumatoid arthritis (32) and 6 ng/ml in cancer (24). Such clinical values fall within the range of serum concentrations that induce muscle wasting in experimental animals (13, 25, 29, 31). Despite a long-standing association with catabolic pathology, the role of TNF-α in muscle wasting remains poorly understood and somewhat controversial (9).

In previous studies, we used cell culture techniques to evaluate TNF-α effects on differentiated skeletal muscle myotubes (20). We found that prolonged exposure to clinically relevant levels of TNF-α (1–6 ng/ml) stimulates muscle protein loss without causing significant cell death, a situation similar to muscle atrophy in vivo. We further determined that TNF-α stimulates the degradation of muscle-specific proteins, including fast-type myosin heavy chains (MHCf). Studies of [35S]methionine incorporation into MHCf showed that protein loss is not due to a reduction in MHCf synthesis. Rather, TNF-α appears to act directly on differentiated myotubes to stimulate protein degradation (20). These findings challenged previous conclusions that TNF-α acts via indirect mechanisms to stimulate muscle wasting (13) and provided evidence that the cytokine exerts direct catabolic effects on muscle cells.

Previous studies have also evaluated TNF-α signal transduction in differentiated myotubes. TNF-α binding to surface receptors stimulates a stereotypical cascade of events that results in proteasomal degradation of I-κBα (20), the protein that inhibits nuclear factor-κB (NF-κB). TNF-α thereby activates NF-κB and causes its translocation to the nucleus (20). This process appears to depend on TNF-α-induced reactive oxygen species (ROS) (20) that derive from mitochondrial electron transport and are essential for NF-κB activation (19).

NF-κB regulates the transcription of genes involved in immune responses, cell growth, and cell death (3, 18). Recent evidence indicates that NF-κB influences cellular proliferation and exiting of the cell cycle by undifferentiated myoblasts (14). However, the functional importance of NF-κB in differentiated skeletal muscle has not been evaluated.

The present study was conducted to assess the putative involvement of NF-κB in TNF-α-induced muscle wasting. Using a standardized cell culture model, we tested the hypothesis that NF-κB activation causes net protein loss in differentiated myotubes. To evaluate cause and effect, we developed transdominant negative skeletal muscle cell lines in which NF-κB signaling was selectively inhibited. Stable transfection of C2C12 myoblasts was used to induce overexpression of mutant I-κBα proteins. These I-κBα variants are insensitive to degradation via ubiquitin-proteasome activity and, therefore, inhibit NF-κB activation. We found that selective blockade of NF-κB prevented protein loss in myotubes challenged with TNF-α. These findings indicate that NF-κB mediates the catabolic response of differentiated muscle cells to TNF-α and suggest a central role for NF-κB in the regulation of cachexia.


Myogenic cell culture and transfection.

Myoblasts from the mouse muscle-derived C2C12 cell line were obtained from American Type Culture Collection (Rockville, MD). As described previously (20), undifferentiated cells were grown in DMEM supplemented with 20% newborn calf serum and gentamicin at 37°C in the presence of 5% CO2. Stable transfection was carried out using Lipofectamine reagent (GIBCO BRL, Gaithersburg, MD) according to the manufacturer's protocol. Myoblasts were transfected with plasmid constructs of I-κBαΔN (truncation of amino acids 1–36) or I-κBαS32/A36 (point mutations of Ser32and Ser36 to alanine); each mutant protein lacks Ser32 and Ser36, phosphorylation sites that are required for I-κBα degradation (5, 8). Plasmid constructs of I-κBαΔN and I-κBαS32/A36 under control of the cytomegalovirus (CMV) promoter were gifts of Dr. Dean Ballard (Vanderbilt University), as was the empty pCMV4 vector used for the control cell line. Plasmid pSVneo was cotransfected for selection by use of neomycin. The selected colonies were pooled.

Transfected myoblasts were stimulated to differentiate by replacing the growth medium with DMEM supplemented with 2% heat-inactivated horse serum. Differentiation was allowed to continue for 96 h before experimentation, with change to fresh medium at 48 h. Mouse recombinant TNF-α (Boehringer Mannheim, Indianapolis, IN) was added to differentiated myotubes at 24-h intervals.

Electrophoresis mobility shift assay.

Electrophoresis mobility shift assay was carried out as previously described (20). Briefly, the binding assay buffer contained 5 mM Tris · HCl (pH 7.5), 100 mM NaCl, 0.3 mM dithiothreitol, 5 mM MgCl2, 10% glycerol, 2 μg of BSA, 0.2% NP-40, and 1 μg of poly(dI-dC). Nuclear extracts were prepared according to Andrews and Faller (1). In each reaction, 4–5 μg of nuclear extract were combined with 1 ng (10,000–15,000 cpm) of NF-κB-binding DNA probe (5′-AGTTGAGGGGACTTTCCCAGGC-3′) labeled with [α-32P]dATP (3,000 Ci/mmol; Amersham Life Science, Arlington Heights, IL) by use of the Klenow fragment. After 30 min of incubation on ice, the reaction mixtures were resolved on 4.5% polyacrylamide gels. The optical density of bands detected on the X-ray film was quantified using commercial densitometry software (Pharmacia, Piscataway, NJ). Protein concentration of the nuclear extracts was determined with the Bio-Rad (Hercules, CA) protein assay kit.

Western blot analysis.

As described previously (20), cell lysates were prepared by boiling harvested cells in Laemmli buffer for 5 min and then were separated using SDS-PAGE and transferred to nitrocellulose membranes. Membranes were incubated in the presence of a monoclonal antibody to MHCf (Novocastra Laboratories, Newcastle, UK) or the FLAG tag (Sigma Chemical, St. Louis, MO) used to identify I-κBαΔN. Horseradish peroxidase-conjugated secondary antibodies were used to locate the primary antibodies. Antibodies were visualized by the enhanced chemiluminescence method (Amersham). Bands detected on the X-ray films were quantified using commercial software (Pharmacia). Protein concentration in the cell lysates was determined using the Bio-Rad Dc protein assay kit.

Analysis of apoptosis.

For DNA ladder detection, DNA of C2C12 myotubes or fibroblast-derived L929 cells (American Type Culture Collection) was extracted using the Quick Apoptosis DNA Ladder Detection Kit (BioVision, Palo Alto, CA) and was separated on 1% agarose gel containing ethidium bromide. For annexin V detection, C2C12 myotubes or L929 cells were grown on coverslips. C2C12 myotubes were differentiated for 96 h and then treated with TNF-α (6 ng/ml) for an additional 72 h; L929 cells were treated with TNF-α (1 ng/ml) for 24 h. The medium then was removed, the coverslips were washed twice with PBS, and the cells were analyzed using the Annexin V-FITC Apoptosis Detection Kit (BioVision). Briefly, the cells were incubated on the coverslips in 500 μl of 1× binding buffer, 5 μl of annexin V-FITC, and 5 μl of propidium iodide for 5 min at room temperature in the dark. Fluorescence microscopy was used to detect emissions from apoptotic cells that stained positive for annexin V.


Data were analyzed for differences among groups with use of commercial software (SigmaStat, Jandel Scientific, Corte Madera, CA). Concentration-dependent decrements in total protein content and MHCf levels were assessed using linear regression analysis (34). Differences in optical densities of NF-κB bands were evaluated using Student's t-test (34). Differences were considered significant at P < 0.05.


Transdominant negative cell lines.

To determine whether NF-κB mediates TNF-α-induced catabolism, we established C2C12 cell lines that overexpressed either of two dominant negative I-κBα mutants. The phosphorylation sites required for degradation of I-κBα (Ser32 and Ser36) are absent from I-κBαΔN (truncation of amino acids 1–36) and I-κBαS32/36A (point mutations of Ser32 and Ser36 to alanine), preventing ubiquitin conjugation and proteolysis of either protein (5, 8). Overexpression of mutant I-κBα was confirmed by Western blot analysis (Fig.1).

Fig. 1.

Expression of mutant I-κBα protein by dominant negative C2C12 myotubes. Right lane: Western blot analysis illustrating FLAG-tagged I-κBαΔN protein in the cytosol of transdominant negative C2C12 myotubes. Left lane: protein is not detectable in the cytosol of control myotubes transfected with the empty pCMV vector. Representative data are shown from 1 of 3 experiments.

Overexpression of I-κBαΔN or I-κBαS32/36A inhibited activation and nuclear translocation of NF-κB in response to TNF-α (Fig. 2). TNF-α activated NF-κB in control myotubes transfected with the pCMV4 vector. This response was indistinguishable from responses previously observed in C2C12 myotubes and in primary myotubes cultured from rat limb muscle (19,20). In contrast, NF-κB activation by TNF-α was largely blocked in myotubes that overexpressed I-κBαΔN (Fig. 2) or I-κBαS32/36A (data not shown). Densitometry showed that TNF-α increased NF-κB content of vector-transfected control myotubes by 13.2 ± 3.7-fold (mean ± SE) relative to untreated controls. In myotubes that overexpressed I-κBαΔN, TNF-α produced only a 2.2 ± 1.0-fold increase. On average, therefore, TNF-α activation of NF-κB was inhibited by >80% in transdominant negative myotubes (P < 0.05, n = 3/group).

Fig. 2.

Mutant I-κBα inhibits activation of nuclear factor-κB (NF-κB) by tumor necrosis factor-α (TNF-α). Electrophoretic mobility shift assay illustrates the functional efficacy of I-κBαΔN, a transdominant negative mutant of I-κBα. NF-κB is activated by TNF-α (3 ng/ml) in control C2C12 myotubes transfected with the empty vector pCMV4 (left lanes); this response is blocked in myotubes that overexpress I-κBαΔN (right lanes). Representative results are shown from 1 of 3 experiments.

NF-κB in TNF-α-induced catabolism.

Chronic exposure to TNF-α causes dose-dependent reductions in MHCf levels and total protein content of differentiated myotubes (20). The transdominant negative cell lines were used to test NF-κB mediation of this response. Control myotubes transfected with the pCMV4 vector were compared with myotubes that overexpressed either I-κBα mutant. Figure 3 shows Western blots that illustrate the effect of TNF-α on MHCf protein levels. Treatment with TNF-α for 72 h caused a dose-dependent reduction of MHCf levels in control myotubes but had no effect on myotubes that overexpressed I-κBαΔN. Figure4 depicts averaged data. TNF-α significantly diminished MHCf and total protein content of control myotubes. In contrast, myotubes that overexpressed I-κBαΔN were unaffected by TNF-α; neither MHCf levels nor total protein content was altered. Myotubes that overexpressed I-κBαS32/36A exhibited a similar insensitivity to TNF-α (data not shown). The capacity of mutant I-κBα to inhibit TNF-α-induced protein loss indicates that this response depends on NF-κB signaling.

Fig. 3.

Western blot illustrating TNF-α effects on fast-type myosin heavy chain (MHCf) in transfected C2C12 myotubes. TNF-α induced a dose-dependent drop of MHCf protein levels in C2C12 myotubes transfected with the empty pCMV4 vector (left) but had no effect on MHCf in C2C12 cells that overexpressed I-κBαΔN (right). Myotubes were treated with TNF-α (1–6 ng/ml) for 72 h; MHCf was detected with monoclonal antibody (Novocastra, Newcastle, UK).

Fig. 4.

NF-κB is required for TNF-α-induced protein loss. Mean data document the effects of TNF-α (1–6 ng/ml) on MHCf level (●; n = 3/data point) and total protein content (○; n = 4/data point) in C2C12 myotubes transfected with the empty pCMV4 vector (A) and those that overexpressed I-κBαΔN (B). Compared with time-matched controls, dose-dependent decrements in MHCf and total protein were observed in C2C12 myotubes transfected with the empty pCMV4 vector (A); no changes were observed in C2C12 myotubes that overexpressed I-κBαΔN (B). Values are means ± SE. Data were analyzed by linear regression; P values depict significance of differences from zero slope.

We previously determined that TNF-α does not induce apoptosis in C2C12 myotubes under the present experimental conditions (20). Similarly, we obtained no evidence that TNF-α stimulates apoptosis in transdominant negative myotubes by use of assays for DNA laddering (Fig. 5) or annexin V (data not shown). In contrast, apoptotic changes were detectable in positive control studies. DNA laddering was evident after TNF-α treatment of L929 fibroblasts (Fig. 5), a nonmuscle cell line with known sensitivity to TNF-α (4). Also, apoptotic changes were observed in differentiated C2C12 myotubes treated with a more potent stimulus (staurosporine plus cycloheximide; data not shown).

Fig. 5.

NF-κB blockade does not induce apoptosis in C2C12 myotubes. DNA ladder assay compares commercial DNA size markers (lane 1) with cellular extracts from control myotubes transfected with the pCMV4 vector (lanes 2 and3), transdominant negative myotubes that overexpressed I-κBαΔN (lanes 4 and 5), and fibroblast-derived L929 cells that are sensitive to TNF-α-induced apoptosis (lanes 6 and 7). DNA laddering was not detected in myotubes treated with TNF-α at 6 ng/ml for 72 h (+) or time-matched control myotubes (−); in contrast, DNA laddering was evident in L929 cells treated with TNF-α at 1 ng/ml for 24 h (+) compared with untreated L929 cells (−). DNA was isolated using DNAzol reagent (GIBCO BRL) and separated on 1% agarose gel.


The present study shows that activation of NF-κB is required for the loss of skeletal muscle protein induced by TNF-α. NF-κB has been studied extensively because of its involvement in biological processes that include immune and inflammatory responses, regulation of cell growth, and apoptosis (3, 18). The present study provides the first direct evidence that NF-κB regulates adaptive responses of differentiated skeletal muscle cells.

TNF-α and muscle wasting.

Muscle wasting and negative nitrogen balance are the hallmarks of inflammatory diseases that range from cancer (27, 30) to emphysema (10) and from congestive heart failure (2) to acquired immunodeficiency syndrome (33). Loss of muscle mass contributes importantly to the mortality and morbidity associated with these disease states. Accordingly, the pathological mechanisms responsible for such losses are of major clinical interest. A primary factor thought to mediate inflammatory catabolism in these and other conditions is TNF-α, which was originally termed “cachectin” because of the strong association between this cytokine and cachexia (9). Circulating TNF-α levels are elevated in inflammatory disease, with serum levels as high as 3–6 ng/ml reported in humans (24, 32). Animal studies clearly demonstrate that exogenous TNF-α stimulates loss of muscle mass and contractile function (13, 25, 29,31). It previously was believed that systemic administration of TNF-α stimulated muscle catabolism via indirect humoral or behavioral effects (13). However, the present observations and those in our previous studies (20) indicate that TNF-α acts directly on differentiated muscle cells to stimulate net protein loss. Studies of MHCf metabolism indicate that TNF-α does not alter protein synthesis (20), suggesting that the cytokine accelerates protein degradation. These observations bolster the biological relevance of TNF-α effects on skeletal muscle and the signaling mechanisms that regulate such effects.

TNF-α/NF-κB signaling.

Studies in nonmuscle cell types have identified three major pathways that transduce the TNF-α signal (16, 21). Briefly, these include a proapoptotic pathway regulated by interaction of the TNF-α-receptor complex with the Fas-associated protein with death domain. A second pathway activates the transcription factor activator protein-1 via Jun-NH2-terminal kinases. The third pathway leads to activation of NF-κB. This last pathway represents a major mechanism of transcriptional control by TNF-α (21) and has been a primary focus of research for our laboratory over the past few years.

As reviewed by several authoritative sources (3, 18), NF-κB is constitutively expressed and exists in the cytosol as part of a heterotrimeric complex. This complex typically comprises the DNA-binding proteins p50 and p65 plus the inhibitory protein I-κBα. Activation of NF-κB requires phosphorylation of I-κBα at Ser32 and Ser36, followed by ubiquitin conjugation and proteolysis of I-κBα by the 26S proteasome. The activated NF-κB dimer is then translocated to the cell nucleus, where it regulates gene expression in a manner that is cell type specific.

TNF-α rapidly activates NF-κB in skeletal muscle cells, including differentiated myotubes (19, 20) and undifferentiated myoblasts (14, 28). Much of the cascade that transduces the TNF-α signal in differentiated muscle has recently been elucidated. Events are triggered by TNF-α binding to sarcolemmal receptors, with the type 1 TNF-α receptor being most likely to regulate protein loss (17). Receptor activation stimulates mitochondrial ROS production, an event that appears to be essential for NF-κB activation in differentiated muscle (19). TNF-α stimulation increases the activity of redox-sensitive kinases, including protein kinase C (19), and causes rapid conjugation of ubiquitin to muscle proteins (20). These events result in proteasomal degradation of I-κBα and translocation of activated NF-κB to the nucleus within 15 min of TNF-α exposure (20).

NF-κB undoubtedly stimulates protein loss via effects on muscle gene expression. The most likely targets are genes that regulate the ubiquitin-proteasome pathway. Animal studies indicate that TNF-α increases ubiquitin mRNA and ubiquitin protein levels in intact skeletal muscle tissue (11, 12). Ubiquitin mRNA also is increased in excised muscle after 3 h of incubation with TNF-α in vitro (22), indicating a direct effect of the cytokine on differentiated muscle fibers. The specific genes that respond to NF-κB and the proteins that regulate the ubiquitin-proteasome pathway under these conditions have not been determined.

Muscle cell death.

Inflammatory cytokines such as TNF-α can be strongly proapoptotic, and NF-κB is known to regulate apoptosis in nonmuscle cells (18, 21). However, after 72 h of TNF-α exposure, we found no evidence that TNF-α induced apoptosis in transdominant negative cell lines or in control myotubes transfected with the empty pCMV vector. Resistance to apoptosis may have been conferred on the transdominant negative myotubes by the leak in NF-κB signaling (∼17% control) that persisted in these cells. The present data are consistent with our previous findings that TNF-α fails to induce apoptosis in differentiated C2C12 myotubes or in myotubes from rat primary cultures (20); nor do TNF-α concentrations <10 ng/ml stimulate necrotic cell death (20). Data in this study and in our previous reports have been obtained using a standardized 72-h protocol and TNF-α concentrations in the clinical range. We cannot rule out the possibility that apoptosis is induced by higher TNF-α concentrations or longer exposure times.

NF-κB and oxidative stress.

NF-κB activation by TNF-α appears to be a redox-sensitive process (28) that involves mitochondrial ROS generation as an essential intermediate step (19). In the absence of TNF-α, exogenous ROS can activate NF-κB directly (20). The present findings suggest that NF-κB activation is a novel mechanism by which ROS may stimulate muscle wasting. This model is consistent with observations that antioxidants inhibit TNF-α-induced muscle wasting in vivo (6) and provides a mechanism whereby oxidative stress may diminish muscle mass in the absence of overt cell death. For example, ionizing radiation activates NF-κB (15), and radiation therapy causes oxidative stress and muscular weakness (17). Perhaps therapeutic levels of ionizing radiation generate sufficient ROS within muscle fibers to activate NF-κB and thereby stimulate muscular atrophy.


This is the first report of a specific transcription factor that regulates loss of skeletal muscle protein. NF-κB activation is only one component of the postreceptor signaling cascade triggered by TNF-α, but it appears to be essential for the catabolic effect of this cytokine on differentiated muscle cells. These findings highlight the importance of determining NF-κB effects on skeletal muscle gene expression.


We thank Jim Agan and Juan Chen for technical assistance and Melanie Moody for assistance with graphics.


  • This study was supported by National Heart, Lung, and Blood Institute Grant HL-59878.

  • Address for reprint requests and other correspondence: M. B. Reid, Pulmonary Medicine, Suite 520B, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030 (E-mail:reid{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract