Cell Physiology

Interleukin-1 stimulates catabolism in C2C12 myotubes

Wei Li, Jennifer S. Moylan, Melissa A. Chambers, Jeffrey Smith, Michael B. Reid


Interleukin-1 (IL-1) is an inflammatory cytokine that has been linked to muscle catabolism, a process regulated by muscle-specific E3 proteins of the ubiquitin-proteasome pathway. To address cellular mechanism, we tested the hypothesis that IL-1 induces myofibrillar protein loss by acting directly on muscle to increase expression of two critical E3 proteins, atrogin1/muscle atrophy F-box (MAFbx) and muscle RING-finger 1 (MuRF1). Experiments were conducted using mature C2C12 myotubes to eliminate systemic cytokine effects and avoid paracrine signaling by nonmuscle cell types. Time-course protocols were used to define the sequence of cellular responses. We found that atrogin1/MAFbx mRNA and MuRF1 mRNA are elevated 60–120 min after myotube exposure to either IL-1α or IL-1β. These responses are preceded by signaling events that promote E3 expression. Both IL-1 isoforms stimulate phosphorylation of p38 mitogen-activated protein kinase and stimulate nuclear factor-κB (NF-κB) signaling; I-κB levels fall and NF-κB DNA binding activity increases. Other regulators of E3 expression are unaffected by IL-1 [cytosolic oxidant activity, Forkhead-O (Foxo) activity] or respond paradoxically (AKT). Chronic exposure of C2C12 myotubes over 48 h resulted in reduced myotube width and loss of sarcomeric actin. We conclude that IL-1α and IL-1β act via an oxidant- and AKT/Foxo-independent mechanism to activate p38 MAPK, stimulate NF-κB signaling, increase expression of atrogin1/MAFbx and MuRF1, and reduce myofibrillar protein in differentiated myotubes.

  • skeletal muscle
  • atrophy
  • cachexia
  • cytokines
  • inflammation

chronic inflammatory diseases cause loss of skeletal muscle mass (muscle atrophy), resulting in weakness and premature fatigue that exacerbate the primary illness. Inflammation-associated atrophy is attributed, in part, to elevated levels of circulating cytokines that exert endocrine effects on skeletal muscle. Among the putative mediators, clinical and preclinical data primarily implicate tumor necrosis factor or TNF (23, 60), interferon-γ or IFN (53), interleukin-6 or IL-6 (3, 72), and the two interleukin-1 isoforms IL-1α (21, 50) and IL-1β (14, 52).

TNF is the archetype in this category. Exogenous TNF stimulates loss of muscle protein in experimental animals (23) and in differentiated myotubes (48). This response is triggered by a cascade of procatabolic signaling events (44, 4648, 55) that upregulate component proteins in the ubiquitin-proteasome pathway (1, 44, 46) and increase pathway activity (50). In particular, two E3 ubiquitin ligases, atrogin1/muscle atrophy F-box (MAFbx) and muscle RING-finger 1 (MuRF1), are responsive to TNF. Atrogin1/MAFbx and MuRF1 are selectively expressed by striated muscle and are thought to impart specificity to the ubiquitin-proteasome system (57), thereby regulating protein degradation. Atrogin1/MAFbx and MuRF1 are upregulated in a multitude of atrophy models, including immobilization, denervation, fasting, diabetes, cancer, and renal failure (6, 28). Mice deficient in either atrogin1/MAFbx or MuRF1 have reduced atrophy in response to denervation, with a 56% and 36% sparing of muscle loss, respectively (6). Human data show that these same genes are elevated with preoperative fasting (67), acute quadriplegic myopathy (17), and in patients with amyotrophic lateral sclerosis (41).

Inflammatory mediators appear to activate atrogin1/MAFbx and MuRF1 through disparate signaling pathways. Atrogin1/MAFbx mRNA expression is at least in part dependent on the mitogen-activated protein kinase p38 MAPK, as the p38 MAPK inhibitor SB203580 blunts TNF upregulation of atrogin1/MAFbx mRNA (44). Both atrogin1/MAFbx and MuRF1 appear to be controlled by nuclear factor κB (NF-κB) signaling. Mice expressing a constitutively active IκB kinase (IKK) have elevated NF-κB activity and increased MuRF1 mRNA (10), whereas atrogin1/MAFbx upregulation in mechanically unloaded muscle is inhibited by a dominant negative IκB that interrupts NF-κB signaling (36). Transcription of E3 genes are also stimulated by forkhead O (Foxo) transcription factors (57). Foxo transcription factors are kept in check by active AKT. Conditions that reduce AKT activity, such as starvation or absence of growth factors, lead to release of Foxo inhibition and an increase in E3 ligase mRNA (63, 71). Consistent with this model, Foxo transcription factors are elevated in muscle of animals subjected to atrophic conditions including starvation, glucocorticoid administration, and diabetes (40, 22). In addition, overexpression of Foxo3a activates the atrogin1/MAFbx promoter while small interfering RNA (siRNA)-mediated knockdown of Foxo1 and Foxo3 reduces promoter activity (64). Humoral factors that stimulate Foxo activity (glucocorticoids, IGF) also increase MuRF1 mRNA expression (63, 71), suggesting that Foxo may also promote MuRF1 expression.

Several reports suggest atrogin1/MAFbx and MuRF1 expression can be influenced by IL-1. Bodine and associates (6) noted that mRNA levels for both genes were elevated in the gastrocnemius muscle of a mouse after 3 days of daily IL-1 injections. Also, Cannon et al. (11) observed that both plasma IL-1β and hindlimb muscle MuRF1 mRNA were elevated in tumor-bearing mice with a cachectic phenotype.

These findings may reflect direct or indirect actions of IL-1 on skeletal muscle. In intact animals, IL-1 causes anorexia (56) and elevates circulating IGFBP-1, effectively reducing IGF bioavailability (4). Thus IL-1 could promote E3 ubiquitin ligase expression indirectly via systemic effects (40, 63). Alternatively, it is possible that IL-1 acts directly on skeletal muscle IL-1R receptors to upregulate E3 ligases and increase muscle protein degradation. Consistent with this model, IL-1 activates p38 MAPK (51) and NF-κB (16, 51) in muscle: signals that favor expression of atrogin1/MAFbx and MuRF1 (10).

To address this issue, we developed an experimental model of cellular events by which IL-1 might act on muscle cells to stimulate catabolism (Fig. 1A). Based on this model, we tested the hypothesis that IL-1α and IL-1β act directly on skeletal muscle to 1) increase atrogin1/MAFbx and MuRF1 mRNA and 2) reduce muscle-specific proteins and myotube size. Experiments were conducted using mature C2C12 myotubes to eliminate systemic cytokine effects and avoid paracrine signaling by nonmuscle cell types. Chronic exposure studies defined potential catabolic activity. Endpoints included sarcomeric actin protein and myotube diameter. Time-course studies defined primary cellular responses to direct IL-1α or IL-1β exposure. Endpoints included atrogin1/MAFbx mRNA, MuRF1 mRNA, and upstream events known to regulate both genes: cytosolic oxidant activity, p38 MAPK, NF-κB, and the AKT/Foxo pathway. Inhibitor studies evaluated p38 MAPK and NF-κB as mediators of atrogin1/MAFbx and MuRF1 expression. The results of this study support our primary hypothesis. Direct IL-1 exposure upregulates both E3 ligases and stimulates myotube catabolism. However, the postreceptor signaling events that mediate these responses do not appear to involve ROS or Akt/Foxo signaling. A revised experimental model consistent with our current findings is shown in Fig. 1B.

Fig. 1.

Models of atrogin1/MAFbx and MuRF1 regulation. A: original experimental model depicts intracellular pathways by which interleukin-1 (IL-1) may alter E3 expression and promote catabolism in skeletal muscle. This model was the basis for hypotheses tested in the current project. Curved line represents sarcolemma. AKT, v-akt murine thomoma viral oncogene homolog; Foxo, Forkhead-O; p38 MAPK, p38 mitogen-activated protein kinase; NF-κB, nuclear factor-κB; ROS, reactive oxygen species. Figure adapted with permission from Glass (27). B: revised model of IL-1/E3 signaling events based on data from the current project.



Mouse recombinant IL-1α and IL-1β were purchased from R&D Systems (Minneapolis, MN). TNF was purchased from Pierce Biotechnology (Rockford, IL). Antibodies to p38 MAPK, Foxo, phospho-AKT, and phospho-Foxo were purchased from Cell Signaling Technologies (Danvers, MA). I-κBα and α-actin antibodies were purchased from Sigma-Aldrich (St. Louis, MO). Antibodies to AKT, phospho-p38 MAPK and phospho-ERK, were purchased from ECM Biosciences (Versailles, KY), NF-κB consensus oligonucleotide was purchased from Promega (Madison, WI). The p38 MAPK inhibitor SB203580 was purchased from Sigma-Aldrich. The NF-κB inhibitor wedelolactone (38, 54) was purchased from Calbiochem (La Jolla, CA). Primers for β-actin (forward and reverse, respectively: 5′-AGGCCCAGAGCAAGAGAGGTA-3′ and 5′-CCATGTCGTCCCAGTTGGTAA-3′), atrogin1/MAFbx (5′-ATGCACACTGGTGCAGAGAG-3′ and 5′-TGTAAGCACACAGGCAGGTC-3′), and MuRF1 (5′-ACGAGAAGAAGAGCGAGC-3′ and 5′-CTTGGCACTTGAGAGGAA-3′) were purchased from Invitrogen (Carlsbad, CA).

Cell culture.

C2C12 myoblasts were purchased from the American Type Culture Collection (Manassas, VA) and grown in DMEM supplemented with 10% FBS for 4 days. Upon 80% confluence, cells were serum restricted in DMEM with 2% heat-inactivated horse serum before and after treatment. On the fifth day after serum restriction, mature myotubes were treated with either PBS + 0.1% BSA (control) and 1 ng/ml IL-1α, 1 ng/ml IL-1β, or 6 ng/ml TNF. After 5–120 min, myotubes were harvested for biochemical analyses. For assessment of effects on myotube width and actin loss, starting on day 4, myotubes were treated with cytokine every 12 h for 48 h.

Total protein and Western blot analyses.

After treatment, myotubes were washed with PBS and scraped into 200 μl of 120 mM Tris, pH 7.5, 200 mM DTT, 20% glycerol, 4% SDS, and 0.002% bromphenol blue. Lysates were sonicated on ice and then heated at 98°C for 5 min. Equal amounts of protein were loaded in each lane of 4–15% Tris·HCl polyacrylamide gels and electrophoresed at 200 V for 50 min. Proteins were either stained using Simply Blue (Invitrogen) and scanned for total protein or were transferred at 200 mA for 2 h to nylon membranes for Western blot analysis. Membranes were incubated in blocking buffer (Odyssey, LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature and incubated with primary antibodies overnight in Odyssey blocking buffer mixed 1:1 with PBS-0.2% Tween, followed by four 5-min washes. Membranes were incubated with fluorescence-conjugated secondary antibodies in Odyssey-PBS 0.2% Tween plus 0.01% SDS for 45 min, followed by four 5-min washes. The membrane was then dried and blots were imaged by use of an infrared fluorescent scanner (Odyssey) to quantify differences. The intensity of each blot was adjusted to avoid saturation and demonstrate treatment-induced differences. Fluorescence intensity data were normalized for total protein and expressed as a percentage of the respective control.

Cytosolic oxidant activity.

As detailed previously (45), the fluorochrome probe 2′,7′dichlorofluorescin diacetate (DCFH-DA; Molecular Probes, Eugene, OR) was used to measure oxidant activity. The cytosol of mature C2C12 myotubes was loaded with nonfluorescent probe by incubation with DCFH-DA 10 μM for 15 min. Accumulation of the oxidized derivative dichlorofluorescein (DCF; excitation 480 nm, emission 520 nm) was measured by use of a fluorescence microscope (TE 2000S, Nikon, Melville, NY) and a charge-coupled device camera (CoolSNAP-ES, Roper Scientific Photometrics, Tucson, AZ) controlled by a computer with image acquisition software (Metamorph 6.1, Universal Imaging, Downingtown, PA). Images were acquired in real time and stored for later analysis of mean emission density from the six brightest myotubes, an index of general oxidant activity.

Electrophoretic mobility shift assay.

Myotubes were treated according to experimental protocols, and nuclear extracts were prepared following commercial nuclear and cytoplasmic extraction protocols (Pierce, Rockford, IL). NF-κB activity in nuclear extracts was determined using 32P-labeled oligonucleotides (Igκ intronic κB site: 5′-CTCAACAGAGGGGACTTTCCGAGAGGCCAT-3′). Ten micrograms of nuclear protein were incubated with the oligonucleotide followed by nondenaturing PAGE. Protein-DNA complexes were detected by autoradiography and quantified by densitometry.

RNA and reverse transcription.

Myotubes were collected and homogenized in 500 μl TRIzol reagent (Invitrogen); RNA was isolated by following the TRIzol RNA extraction protocol. RNA concentration was quantified at 260 nm by spectrophotometry. Samples were diluted to equal concentrations of RNA. Two microliters of RNA were used to synthesize cDNA at 37°C for 90 min with a thermocycler using 40 units of Moloney murine leukemia virus (M-MLV) reverse transcriptase, 4 μl of 5× M-MLV reaction buffer, 0.2 μl of 10 mM dNTP, 0.5 μg of random primer, and 12.4 μl of diethyl pyrocarbonate (DEPC)-treated H2O for a final volume of 20 μl (reagents from Promega).

Real-time PCR.

mRNA was quantified using the ABI 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). cDNA was obtained from the reverse transcription reaction of mRNA as described above. Real-time PCR analyses were performed in duplicate using 96-well plates with SYBR green master mix (Applied Biosystems). β-Actin was used as an endogenous control. Five microliters of cDNA were used as a template for real-time PCR; 1 μl forward primer 50 μM and 1 μl reverse primer 50 μM were added to 25 μl SYBR green master mix and 18 μl DEPC-treated water. Reactions were performed in a 50-μl reaction volume under the following conditions: initial step at 50°C for 2 min followed by 95°C for 10 min and 40 cycles of denaturation at 95°C for 15 s and hybridization and elongation at 60°C for 1 min. Primer sequences were as described in materials. Threshold cycles for both the endogenous control (β-actin) and treatment (atrogin/MAFbx, MuRF1) cDNA for each reaction were determined by Applied Biosystems Sequence Detection Software 1.3. The abundance of target mRNA relative to β-actin mRNA was determined using the comparative cycle threshold method (26, 49).

Myotube width measures.

To measure myotube width, 4–5 brightfield images per dish were captured, randomized, and coded. Measurements were made without knowledge of treatment using Metamorph image acquisition software. Two diagonal lines were drawn across each image. Width was measured where the diagonal lines transected myotubes.

Statistical analyses.

Data are expressed as means ± SE. Differences were assessed using Student's t-test for the comparison of two means from normally distributed data sets or one-way ANOVA for comparisons of multiple means. When ANOVA detected a significant overall difference, Tukey's post hoc test for multiple comparisons was performed. P values <0.05 were considered significant.


Effects of chronic IL-1 exposure.

TNF is a catabolic cytokine that stimulates protein loss in skeletal muscle (74). We demonstrate that IL-1 also has catabolic effects that are illustrated in Fig. 2. Repeated doses of IL-1α or IL-1β over a 48-h period reduced myotube width by 13% and 10%, respectively. Parallel doses of TNF served as a positive control and reduced myotube width 11%. In these same cells, sarcomeric actin was also reduced 13–14% by either IL-1 isoform or by TNF.

Fig. 2.

Prolonged IL-1 exposure reduces myotube width and sarcomeric actin content. A: myotube morphology. Panels depict representative images of myotubes following repeated 12-h exposures to 1 ng/ml IL-1α, 1 ng/ml IL-1β, 6 ng/ml tumor necrosis factor (TNF), or an equal volume of cytokine-free vehicle (control). Images were captured after 48 h of repeated exposure. B: sarcomeric actin. Panels depict original blots. C: relationship between averaged myotube width and averaged integrated intensities from Western blot analyses of sarcomeric actin; paired measurements of both variables were obtained from each myotube culture dish. Means shown ± SE, n = 3 dishes (IL-1α) or 7 dishes (IL-1β and TNF) per group; *P < 0.05 vs. control.

IL-1 effects on atrogin1/MAFbx and MuRF1 mRNA.

E3 ubiquitin ligases atrogin1/MAFbx and MuRF1 mediate skeletal muscle catabolism (6, 28). Direct exposure of C2C12 myotubes to IL-1α or IL-1β caused a progressive, time-dependent increase in atrogin1/MAFbx mRNA (Fig. 3A). MuRF1 mRNA was also increased by either IL-1 isoform (Fig. 3B). The response to IL-1α was evident sooner and was larger in magnitude than the response to IL-1β. After 120 min, both IL-1 isoforms stimulated larger increases in atrogin1/MAFbx mRNA than MuRF1 mRNA. Subsequent studies assessed signaling events that regulate atrogin1/MAFbx and MuRF1 expression.

Fig. 3.

Time course of IL-1-stimulated increases in atrogin1/MAFbx and MuRF1 mRNA. Data depict atrogin1/muscle atrophy F-box (MAFbx) (A) and muscle RING finger 1 (MuRF1) (B) mRNA levels in myotubes collected 30, 60, or 120 min after exposure to IL-1α 1 ng/ml (filled circles) or IL-1β 1 ng/ml (open circles). Means shown ± SE; atrogin1/MAFbx, n = 3 per group at all time points; MuRF1, n = 3 at time points 30 and 60 min, at 120 min n = 6 (IL-1α) or n = 9 (IL-1β); *P < 0.02 vs. control.

Cytosolic oxidant activity.

Oxidative stress increases atrogin1/MAFbx and MuRF1 mRNA levels in muscle (45) and is a common response of many cell types to proinflammatory cytokines (33, 48, 66). However, neither IL-1 isoform altered cytosolic oxidant activity in our myotubes (Fig. 4). IL-1α did not affect mean activity, and the modest (∼5%) increase after IL-1β was not statistically resolvable. In contrast, TNF was included as a positive control in this experiment and increased oxidant activity as expected (48).

Fig. 4.

Cytosolic oxidant activity is unaffected by IL-1. Panels depict representative fluorescence micrographs (left) and cytosolic oxidant activities (right) in live, intact myotubes following 30 min exposure to 1 ng/ml IL-1α, 1 ng/ml IL-1β, 6 ng/ml TNF, or an equal volume of cytokine-free vehicle (con). Bars depict means ± SE; n = 12/group; *P < 0.05 vs. time-matched controls.

p38 MAPK signaling.

p38 MAPK participates in TNF-induced regulation of atrogin1/MAFbx expression (44). p38 MAPK is activated by IL-1 in nonmuscle cell types (24, 35), is suggested to mediate IL-1β signaling in C2C12 myotubes (51), and is a positive regulator of atrogin1/MAFbx expression (44). We found that either IL-1α or IL-1β stimulated rapid phosphorylation of p38 MAPK (Fig. 5A). The response was transient, peaking within 5–15 min, and largely reversing by 60 min. The timing of this response differed between cytokines. IL-1α induced a more rapid increase, with phosphorylation peaking at 5 min. IL-1β elicited a slower but more persistent response. Inhibition of p38 activity with SB203580, a compound that specifically competes for the ATP binding site of 38α and 38β (81), reduced IL-1 induction of atrogin1/MAFbx mRNA by ∼40% without altering MuRF1 expression (Fig. 5B).

Fig. 5.

IL-1 stimulation of p38 mitogen-activated protein kinase (MAPK) activity and dependence of atrogin1/MAFbx expression. A: p38 MAPK phosphorylation. Panels depict original blots (top) and averaged integrated intensities (bottom) from Western blot analyses of p38 MAPK phosphorylation in myotubes collected 5, 10, 15, 30, or 60 min after exposure to 1 ng/ml IL-1α (filled circles) or 1 ng/ml IL-1β (open circles). B: effect of p38 inhibition on E3 expression. Bars depict atrogin1/MAFbx and MuRF1 mRNA levels in myotubes collected after a 30-min pretreatment with either vehicle (0.1% DMSO) or SB203580 5 μM plus a 2-h exposure to 1 ng/ml IL-1α or 1 ng/ml IL-1β. Means shown ± SE; n = 3/group at each time point; *P < 0.05 vs. time-matched control cells.

NF-κB signaling.

NF-κB promotes expression of MuRF1 (10) and is activated by IL-1β in differentiated muscle cells (16, 51). Our data suggest the response of NF-κB to IL-1 is largely isoform independent. Figure 6A illustrates transient decrements in myotube content of I-κBα, the NF-κB inhibitor protein, following IL-1α or IL-1β exposure. I-κBα levels were diminished within 10–15 min and remained depressed for at least 30 min. Consistent with canonical regulatory mechanisms (25), the DNA binding activity of NF-κB subsequently increased (Fig. 6B). DNA binding activity was elevated 30–60 min after IL-1α or IL-1β exposure, returning to control levels by 120 min. As predicted, inhibition of NF-κB activation with wedelolactone, a selective irreversible inhibitor of IKKα and IKKβ, (38), blunted IL-1-induced increases in both atrogin1/MAFbx and MuRF1 mRNA (Fig. 6C).

Fig. 6.

IL-1 stimulation of NF-κB activity and NF-κB dependence of MuRF1 expression. A: I-κBα degradation. Panels depict original blots (top) and averaged integrated intensities (bottom) from Western blot analyses of I-κBα in myotubes collected 5, 10, 15, 30, or 60 min after exposure to IL-1α 1 ng/ml (filled circles) or IL-1β 1 ng/ml (open circles). B: NF-κB DNA binding. Panels depict original blots (top) and averaged optical densities (bottom) from EMSA analyses of NF-κB DNA binding activity in nuclear extracts from myotubes collected 30, 60, or 120 min after exposure to IL-1α 1 ng/ml (filled circles) or IL-1β 1 ng/ml (open circles). C: effect of NF-κB inhibition on E3 expression. Bars depict atrogin1/MAFbx and MuRF1 mRNA levels in myotubes collected after 30 min pretreatment with either vehicle (0.05% DMSO) or wedelolactone 12.5 μM plus a 2-h exposure to IL-1α 1 ng/ml or IL-1β 1 ng/ml. Means shown ± SE; n = 3/group at each time point; *P < 0.05 vs. time-matched control cells.

AKT/Foxo signaling.

Atrogin1/MAFbx and MuRF1 mRNA levels are positively regulated by the Foxo family of transcription factors (64, 71). Foxo activity is limited by AKT signaling. Activated AKT phosphorylates Foxo, inhibiting Foxo translocation to the nucleus. The observed increases in atrogin1/MAFbx and MuRF1 mRNA levels (above) suggested IL-1 would decrease AKT phosphorylation and increase Foxo activity. Neither occurred. Instead, AKT phosphorylation was paradoxically increased by IL-1α and IL-1β (Fig. 7A). This response was delayed relative to other signaling events, requiring 60 min to resolve. Neither IL-1 isoform nor TNF altered the DNA binding activity of Foxo (Fig. 7B). As a positive control, we used wortmannin to inhibit phosphoinositol-3 kinase, an upstream regulator of AKT, thereby lessoning AKT activity and inducing Foxo DNA binding (59).

Fig. 7.

IL-1 stimulates AKT phosphorylation without promoting Foxo DNA binding. A: AKT phosphorylation. Panels depict original blots (top) and averaged integrated intensities (bottom) from Western blot analyses of AKT phosphorylation in myotubes collected 60 or 120 min after exposure to 1 ng/ml IL-1α (filled circles) or 1 ng/ml IL-1β (open circles). B: Foxo DNA binding. Panels depict original blots (left) and averaged optical densities (right) from EMSA analyses of Foxo DNA binding activity in nuclear extracts from myotubes collected 120 min after exposure to 1 ng/ml IL-1α, 1 ng/ml IL-1β, 6 ng/ml TNF, or 100 nm wortmannin. Means shown ± SE; n = 3/group; *P < 0.05 vs. time-matched, vehicle-treated control cells.


Our data show that IL-1α and IL-1β act directly on differentiated C2C12 myotubes to decrease myotube width and sarcomeric actin. These events are preceded by increases in atrogin1/MAFbx and MuRF1 mRNA levels and a series of time-dependent signaling events that stimulate expression of these two E3 proteins: p38 MAPK phosphorylation, I-κBα degradation, and NF-κB activation (Fig. 8). Other signals that promote atrogin1/MAFbx and MuRF1 expression under different circumstances appear unaffected by IL-1 (cytosolic oxidant activity, Foxo activity) or respond to IL-1 paradoxically (AKT phosphorylation). These findings provide new information on the cellular response to IL-1 and broaden our understanding of E3 regulation in muscle.

Fig. 8.

Sequence of IL-1 effects on E3 mRNA and upstream regulators. Integrative model depicts changes in atrogin1/MAFbx (atrogin) and MuRF1 mRNA as dashed lines. Shaded areas depict changes in p38 MAPK phosphorylation (p38), I-κBα protein content (I-κBα), and NF-κB DNA binding activity (NF-κB). Data are time-matched averages of the individual responses to IL-1α and IL-1β (Figs. 3, 5, and 6).

IL-1 and its receptors.

As reviewed by Allan et al. (2), the IL-1α and IL-1β isoforms are products of two distinct genes on chromosome 2 and have high sequence homology. Both are synthesized as large precursor proteins that are cleaved to their active form by cytosolic caspase-1 before being released into the extracellular environment. Both isoforms exert biological effects by binding the membrane-associated type I IL-1 receptor (IL1-R1), which associates with the IL-1 receptor accessory protein (IL-1RAcP) to form a transmembrane signaling complex. The type II IL-1 receptor (IL-1R2) lacks a signaling domain and therefore functions as a decoy receptor. Skeletal muscle constitutively coexpresses both IL-1R1 and IL-1R2, which localize to the sarcolemma and nuclei of muscle fibers (29). Skeletal muscle also expresses both isoforms of IL-1. In murine muscle, IL-1β protein levels exceed IL-1α levels, a difference that increases with age (31).

In the blood, IL-1 circulates at low levels of 4–6 pg/ml (5, 18) that become elevated with disease. Children with insulin-dependent diabetes have a moderate increase (11–13 pg/ml) in circulating IL-1 (18). Patients with more severe inflammatory disease such as chronic obstructive pulmonary disease have levels that range from 15 to 118 pg/ml (34, 39). Local tissue concentrations can be much higher. For example, Borge et al. (7) measured interstitial cytokine levels in muscle during experimental sepsis. IL-1β levels peaked at 1 ng/ml 180 min after intravenous injection of LPS. We used this same concentration in our cell culture system, demonstrating that elevated local cytokine levels possess strong potential to induce muscle catabolism.

Interleukins and muscle catabolism.

IL-1α and IL-1β have long been considered putative mediators of inflammation-associated muscle atrophy (14, 21, 50, 52). In humans, elevated serum levels correlate with wasting-associated diseases such as cancer (58), sporadic inclusion body myositis (20), chronic obstructive pulmonary disease (65), and HIV infection (61). In rats, Fong and colleagues (21) demonstrated that a twice-daily administration of IL-1α over 7 days produced anorexia, weight loss, and a decrease in skeletal muscle protein. These earlier studies demonstrate a procatabolic role for IL-1 in vivo but cannot discriminate between indirect, systemic effects versus a direct response of muscle to IL-1. Cell culture studies to defined direct responses have shown that IL-1β can impair myogenesis in C2C12 myoblasts by blocking the actions of IGF (8). To our knowledge, the present study is the first to demonstrate that chronic exposure of cultured myotubes to IL-1 can reduce myofibrillar protein and myotube diameter. These results add weight to the idea that IL-1α and IL-1β can act directly on skeletal muscle to trigger catabolic programs that lead to degradation of myofibrillar proteins.

Interleukins and E3 expression.

Two previous reports have suggested that IL-1 influences E3 expression in muscle. Bodine and associates (6) found that daily IL-1β administration to an adult mouse for 3 days elevated both atrogin/MAFbx mRNA and MuRF1 mRNA levels in the medial gastrocnemius muscle. Also, Cannon, et al. (11) described concurrent elevations of circulating IL-1β protein and MuRF1 mRNA in hindlimb muscles of tumor-bearing mice. The current data show that E3 upregulation is a direct response of differentiated C2C12 myotubes to either IL-1α or IL-1β. Rapid increases in atrogin1/MAFbx and MuRF1 mRNA resemble the time course of TNF effects in the same experimental system (44, 55).

Not all catabolism-associated cytokines elicit this response. As previously observed for IFN (70), we have seen that direct IL-6 exposure does not alter expression of atrogin1/MAFbx or MuRF1 (data not shown). This is consistent with the data of Haddad and colleagues (32) who infused IL-6 into tibialis anterior muscle of intact rats for 14 days. They observed no changes in mRNA for either of these E3s.

Interleukins and ROS.

Evidence suggests that TNF utilizes oxidants as second messengers to promote catabolic responses through NF-κB (48). We hypothesized that the IL-1 isoforms would behave similarly. However, IL-1 treatment did not produce the anticipated rise in oxidants. The reasons for this are unclear. The signaling complexes that form upon ligand binding to IL-1R1 and TNFR1 are very different. IL-1RI acts through IL-1 receptor-associated kinase (IRAK) and TNF receptor-associated factor 6 (TRAF6), whereas TNF-α receptor I (TNFRI) acts through TNF receptor 1-associated death domain (TRADD) and TRAF2 and TRAF5 (75). These are only a few of the many differences in receptor complex composition. However, signaling from these receptors converges on the IKK complex resulting in activation of the canonical NF-κB pathway. Our data suggest the IL-1 receptor complex may not use reactive oxygen species (ROS) as second messengers. Alternatively, the concentration or compartmentalization of ROS may limit detection by our method. To the best of our knowledge, there is no precedent for IL-1 induction of oxidant production in skeletal muscle. In colonic (13) and esophageal (12) smooth muscle, IL-1β induces oxidant production but not until 2 h postexposure. This time course of oxidant production is inconsistent with the rapid responses typical for cell signaling molecules.

p38 MAPK and atrogin1/MAFbx.

IL-1 stimulates activity of all three MAPK isoforms in a variety of nonmuscle cell types (24, 35, 43, 77). Pharmacological studies suggest that MAPKs may also mediate IL-1 effects in myotubes (51). The principal focus of our current analyses was p38 MAPK. We have identified p38 MAPK as a positive regulator of atrogin1/MAFbx expression in TNF-stimulated muscle (44). Similarly, both IL-1 isoforms caused a rapid increase in p38 MAPK phosphorylation that preceded the rise in atrogin1/MAFbx mRNA. Pharmacological inhibition of p38 MAPK reduced IL-1-induced atrogin1/MAFbx mRNA by 40%. The standard 5 μM dose of SB203580 (8 times the IC50) should be adequate to reduce p38 activity while minimizing nonspecific effects. The fact that gene expression was partially inhibited suggests that additional pathways may contribute to the atrogin1/MAFbx response to IL-1.

NF-κB regulation of E3 expression.

As recently reviewed by Sethi and co-workers (69), the I-κBα/NF-κB protein complex resides in the cytosol of mammalian cells. Stimuli that activate NF-κB trigger phosphorylation, ubiquitination, and proteasomal degradation of I-κBα. This disinhibits the NF-κB dimer which translocates to the nucleus and binds to κB sites in DNA promoter regions, stimulating transcription of selected genes that appear to include both atrogin1/MAFbx (36) and MuRF1 (10). Consistent with this canonical model, we observed I-κBα degradation and NF-κB activation in response to either IL-1 isoform. These changes occurred before any detectable increase in mRNA of either E3 protein. In addition, wedelolactone was used to pharmacologically inhibit IKK-α and -β, upstream activators of NF-κB (38, 54). Wedelolactone depressed IL-1 induction of atrogin1/MAFbx and MuRF1 mRNA, suggesting both genes are regulated by NF-κB under the current conditions.

AKT/Foxo signaling in muscle.

The Foxo family of transcription factors is phosphorylated by AKT, which excludes Foxo from the nucleus (9). Dephosphorylation enables translocation to the nucleus where Foxo binds to the promoter regions of atrogin1/MAFbx and MuRF1 and stimulates gene expression (37, 64). AKT/Foxo signaling is necessary for upregulation of both E3s caused by nutrient restriction or dexamethasone and is the mechanism by which IGF-1 inhibits E3 expression (64, 71, 76). We previously confirmed that the AKT/Foxo pathway is functional in C2C12 myotubes and that it modulates constitutive atrogin1/MAFbx expression (55). However, AKT signaling is not required for the increase in atrogin1/MAFbx mRNA stimulated by TNF (55).

IL-1 effects on AKT/Foxo signaling are largely unstudied in skeletal muscle. IL-1β is known to stimulate AKT phosphorylation in a variety of cell types, including neurons (68), hepatocytes (73), astrocytes (78), and lung carcinoma cells (15). In C2C12 myoblasts, a prior report indicates that IL-1β does not alter AKT protein levels or insulin-stimulated AKT phosphorylation (30). We found that either IL-1α or IL-1β stimulate phosphorylation of AKT in C2C12 myotubes, a response detectable within 60 min. The current study is the first to test IL-1 effects on Foxo signaling in any cell type. Neither isoform altered DNA binding activity. In combination, the AKT/Foxo pathway does not appear to mediate IL-1 effects on E3 expression. This is not unprecedented; as mentioned above, TNF regulation of atrogin1/MAFbx appears to bypass AKT/Foxo signaling (55). It has also been shown that p38 mediates upregulation of atrogin1/MAFbx mRNA in cardiac myocytes (80). In muscles of patients with chronic obstructive pulmonary disease have elevated atrogin1/MAFbx mRNA without a concurrent reduction in phosphorylated AKT (19). Thus a model is emerging that suggests inflammatory conditions activate atrogin1/MAFbx via p38, whereas metabolic stimuli act through downregulation of AKT. Alternatively, recent studies suggest that activation of AKT is not limited to promoting anabolic signaling and may contribute to catabolism. Russell et al. (62) demonstrated that AKT plays a role in protein degradation induced by proteolysis-inducing factor (PIF). Inhibition of AKT by LY294002 or dominant negative AKT prevents PIF-induced protein degradation. AKT inhibition also prevents PIF induced-protein degradation of I-κBα and nuclear translocation of NF-κB (79). Thus IL-1 activation of AKT could promote E3 expression via NF-κB.

In conclusion, integrating the available data, proinflammatory cytokines IL-1 and TNF appear to promote muscle protein loss through several common pathways that regulate E3 expression. Similar to TNF, IL-1 induces p38 MAPK phosphorylation, NF-κB DNA binding, and E3 mRNA expression in a temporally coordinated manner. In addition, both IL-1 and TNF appear to stimulate atrogin1/MAFbx and MuRF1 expression via an AKT-independent mechanism. AKT functions as a nutrient sensor and is downstream of growth factors such as insulin-like growth factor-1 (IGF-1) (42). Moylan and associates (55) have proposed that AKT-independent signaling provides a mechanism whereby inflammatory cytokines might stimulate E3 expression and muscle catabolism despite high levels of circulating nutrients or growth factors. This concept is reinforced by the current data and suggests a working model for future evaluation.


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


View Abstract