Journal of Applied Physiology

TNF-α acts via TNFR1 and muscle-derived oxidants to depress myofibrillar force in murine skeletal muscle

Brian J. Hardin, Kenneth S. Campbell, Jeffrey D. Smith, Sandrine Arbogast, Jacqueline Smith, Jennifer S. Moylan, Michael B. Reid


Tumor necrosis factor-α (TNF) diminishes specific force of skeletal muscle. To address the mechanism of this response, we tested the hypothesis that TNF acts via the type 1 (TNFR1) receptor subtype to increase oxidant activity and thereby depress myofibrillar function. Experiments showed that a single intraperitoneal dose of TNF (100 μg/kg) increased cytosolic oxidant activity (P < 0.05) and depressed maximal force of male ICR mouse diaphragm by ∼25% within 1 h, a deficit that persisted for 48 h. Pretreating animals with the antioxidant Trolox (10 mg/kg) lessened oxidant activity (P < 0.05) and abolished contractile losses in TNF-treated muscle (P < 0.05). Genetic TNFR1 deficiency prevented the rise in oxidant activity and fall in force stimulated by TNF; type 2 TNF receptor deficiency did not. TNF effects on muscle function were evident at the myofibrillar level. Chemically permeabilized muscle fibers from TNF-treated animals had lower maximal Ca2+-activated force (P < 0.02) with no change in Ca2+ sensitivity or shortening velocity. We conclude that TNF acts via TNFR1 to stimulate oxidant activity and depress specific force. TNF effects on force are caused, at least in part, by decrements in function of calcium-activated myofibrillar proteins.

  • cytokine
  • respiratory muscle
  • oxidative stress
  • weakness
  • diaphragm

tumor necrosis factor-α (TNF) is a proinflammatory cytokine that promotes muscle weakness and is elevated in the circulation of individuals with chronic disease (34). Exogenous TNF administered either in vitro (1, 24, 32) or in vivo (33) depresses specific force, i.e., force per cross-sectional area, in respiratory and limb skeletal muscles. This response develops within hours (1, 32, 33), persists with prolonged exposure (18), and is induced by TNF levels that are too low to cause muscle atrophy (18). The fall in specific force is accompanied by an increase in muscle-derived oxidants and is opposed by antioxidant administration (18).

The present study addresses the cellular mechanism by which TNF depresses specific force. We developed a standardized murine model in which skeletal muscle (diaphragm) was exposed to exogenous TNF in vivo. Changes in muscle function were subsequently measured ex vivo. Developmental studies defined the time course of specific force loss, confirmed that TNF increases oxidant activity in muscle fibers, and established that TNF effects can be inhibited by antioxidant (Trolox) pretreatment. This experimental model was subsequently used to test two hypotheses:

Hypothesis 1: TNF depresses specific force via the TNF type 1 receptor subtype.

Skeletal muscle fibers constitutively coexpress the TNF type 1 receptor (TNFR1) and type 2 receptor (TNFR2) subtypes. We previously proposed that TNFR1 binding stimulates oxidant production in muscle cells (25). In support of this, a growing literature indicates that TNF acts via TNFR1 to stimulate oxidant production in nonmuscle cell types (27). We used genetically engineered mice deficient in either TNFR1 (TNFR1−/−) or TNFR2 (TNFR2−/−) to test the role of each receptor subtype in skeletal muscle.

Hypothesis 2: TNF depresses myofibrillar function in skeletal muscle.

TNF exposure diminishes tetanic force of intact muscle fibers without changing tetanic calcium transients (24). This finding suggests that TNF alters events downstream of the calcium signal and depresses force via effects on myofibrillar proteins. We tested this postulate by measuring calcium-activated force in permeabilized muscle fibers isolated from muscles of TNF-treated animals.


Experimental animals.

This study conformed to the “Guiding Principles on Care and Use of Laboratory Animals” of the American Physiological Society and the National Institutes of Health. All experiments were approved by the institutional review committee for animal care. Mice were kept on a 12:12-h light-dark cycle. Food and water were provided ad libitum.

For time course experiments, age-matched adult male ICR mice (Harlan, Indianapolis, IN) were randomly assigned to receive either an intraperitoneal (ip) injection of TNF (100 μg/kg) (Pierce Biotechnology, Rockford, IL) or an equal volume of buffer (Krebs-Ringer solution; see below). The diaphragm was excised for study 1, 2, 4, 24, or 48 h after injection. For antioxidant pretreatment experiments, age-matched male ICR mice were randomly assigned to receive Trolox (10 mg/kg) (Sigma-Aldrich, St. Louis, MO) or an equal volume of vehicle (Krebs-Ringer solution) by intraperitoneal injection 1 h before TNF injection. The Trolox dosage was adapted from Betters et al. (4), who used a 20 mg/kg priming dose to block oxidative stress and contractile dysfunction of rat diaphragm during mechanical ventilation. In preliminary studies, we found that pretreatment using 10 mg/kg or 20 mg/kg were equally protective against TNF effects; we used the lower dose in formal studies. The diaphragm was excised 1 h after TNF injection. Receptor-deficient mice (TNFR1−/− B6.129-Tnfrsf1atm1Mak, TNFR2−/− B6.129S2-Tnfrsf1btm1Mwm; Jackson Labs, Bar Harbor, ME) and genetic controls (C57BL/6J; Jackson Labs) received TNF or vehicle intraperitoneally 1 h before the diaphragm was excised. Before muscle excision, mice were anesthetized with isoflurane and euthanized by cervical dislocation.

Muscle force measurements.

Muscle fiber bundles were isolated from the costal diaphragm and mounted in a temperature-controlled bath containing Krebs-Ringer solution (in mM: 137 NaCl, 5 KCl, 1 MgSO4, 1 NaH2PO4, 24 NaHCO3, 2 CaCl2) bubbled with 95% O2-5% CO2 at room temperature. Silk suture (6-0) was used to fix the tendon to a force transducer (BG Series 100g, Kulite, Leonia, NJ) mounted on a micrometer by which muscle length was adjusted. The muscle was positioned between platinum wire stimulating electrodes and stimulated to contract isometrically using electrical field stimulation (supramaximal voltage, 1.2-ms pulse duration). The output of the force transducer was recorded using an oscilloscope (546601B; Hewlett-Packard, Palo Alto, CA) and a chart recorder (BD-11E; Kipp and Zonen, Delft, The Netherlands). In each experiment, muscle length was adjusted to optimize twitch force (optimal length, Lo). The bath temperature was then increased to 37°C, and 30 min were allowed for thermoequilibration. The force-frequency relationship was determined using contractions evoked at 2-min intervals using stimulus frequencies of 1, 15, 30, 50, 80, 120, 150, 250, and 300 Hz, and a tetanic train duration of 500 ms. Maximal tetanic contractions (300 Hz) were stimulated between lower-frequency contractions to monitor contractile stability. After each experiment, Lo was measured using an electronic caliper, and the muscle was removed, blotted dry, and weighed. Cross-sectional area was determined as defined by Close (9).

Permeabilized fiber force measurements.

Single chemically permeabilized diaphragm fibers were prepared using a technique similar to that described by Campbell and Moss (6). Diaphragms were isolated from mice and immediately immersed in ice-cold relaxing solution (in mM: 100 KCl, 10 imidazole, 4 ATP, 2 EGTA, 5 MgCl2) at pH 7.0. Thin bundles were then cut from the costal diaphragm, tied to glass capillary tubes, and chemically permeabilized in relaxing solution containing 1% Triton X-100 (4 h, 4°C). Bundles were stored at −20°C in relaxing solution containing 50% glycerol for up to 7 days before use.

Segments of individual fibers (Lo 979 ± 41 μm, width 32.1 ± 1.1 μm) were attached between a force transducer (403B, Aurora Scientific, Ontario, Canada) and a motor arm (312B, Aurora) using the technique illustrated in Figure 1 of Campbell and Moss (6) and stretched to a sarcomere length of 2.59 ± 0.01 μm in pCa 9.0 solution, where pCa = −log10[Ca2+]. pCa solutions contained (in mM) 20 imidazole, 14.5 creatine phosphate, 7 EGTA, 4 MgATP, 1 free Mg2+, as well as free Ca2+ ranging from 1 nM (pCa 9.0) to 32 μM (pCa 4.5) and sufficient KCl to adjust the ionic strength to 180 mM. Experiments were performed at 22°C. The Ca2+ dependence of steady-state isometric force was assessed by activating the fiber in solutions with pCa values ranging from 6.2 to 4.5. Once force had attained its steady-state value in a given pCa solution, the fiber was rapidly (0.6 ms) shortened by 0.2 Lo, held at this length for 20 ms, and then stepped back to its original length. Rates of tension recovery (ktr) were calculated by fitting a single-exponential curve to each recovery time course. Maximum shortening velocity was determined in pCa 4.5 solution using the slack test method described by Edman et al. (12). Forces developed by permeabilized fibers were analyzed using four-parameter Hill curves {F = F0[Ca2+]n/([Ca2+]n + [Ca2+50]n), where F0 is maximal force, n is the Hill coefficient, and [Ca2+50] is the [Ca2+] required for half-maximal activation.} fitted separately to the data from individual fibers. Experimental records were acquired and analyzed using SLControl software (7).

Fig. 1.

In vivo tumor necrosis factor-α (TNF) administration alters contractile function of mouse diaphragm. A: data from diaphragm fiber bundles tested 1 h after administration of TNF or vehicle ip. Specific force was depressed by TNF across a range of maximal and near-maximal stimulus frequencies (80–300 Hz; P < 0.05). B: data obtained 24 h after TNF or vehicle administration show persistent depression of specific force (80–300 Hz; P < 0.05) in TNF-exposed muscle. C: time course of changes in maximum (300 Hz) tetanic force after TNF or vehicle administration. Relative to data from uninjected control animals, TNF depressed force at 1 h, a decrement that persisted for at least 48 h. Vehicle administration did not systematically alter maximal force. Means shown ± SE; n = 3/group. *P < 0.05 by repeated-measures ANOVA.

Cytosolic oxidant activity.

Intact diaphragms were transferred to a bath containing oxygenated Krebs-Ringer solution at 37°C. The muscle was pinned at near-optimal length as determined in prior studies of muscles from age- and sex-matched animals of the same strain. Intracellular oxidant activity was determined by use of the fluorochrome probe 2′,7′-dichlorofluoroscin diacetate (DCFH-DA; Molecular Probes, Eugene, OR). In brief, DCFH-DA dissolved in ethanol was added to buffer containing the isolated muscle (final concentration 50 μM). Intracellular esterases cleave the diacetate side chain to yield nonfluorescent DCFH, which accumulates in the cytosol and reacts with muscle-derived oxidants to yield the fluorescent derivative 2′,7′-dichlorofluorescein (DCF; 480 nm excitation, 520 nm emission). An increase in emission intensity reflects DCF accumulation and is measured as an index of net oxidant activity. Emissions were quantified using an epifluorescence microscope (TE 2000S; Nikon USA, Melville, NY) and charge-coupled device camera (CoolSNAP-ES; Roper Scientific Photometrics, Tuscon, AZ) controlled by a computer using data-acquisition software (Metamorph 6.1; Universal Imaging, Downingtown, PA). Images were obtained from a 0.27-mm2 site on the muscle and were stored for later quantification. Emissions from DCFH-free buffer or muscle fibers are not detectable under these conditions, and background correction is not required. Photooxidation artifact was controlled by conducting experiments in a darkened laboratory and by correcting the data using a previously reported protocol (3).

Statistical analyses.

Differences between force-frequency curves were analyzed using two-way, repeated-measures ANOVA (14). Paired comparisons between DCFH data were evaluated using Student's paired t-tests (37). Parameters of Hill curves fitted to data from permeabilized fibers were tested for differences between groups by use of unpaired t-tests. Statistical calculations were performed using commercial software (SigmaStat, SPSS, Chicago, IL; and Microsoft Excel). Results are reported as means ± SE.


Time-course of TNF effects in vivo.

TNF administration to mice altered contractile function of the diaphragm. Figure 1, A and B, depicts specific forces of fiber bundles isolated 1 or 24 h after TNF or vehicle injection. The TNF-associated decrement was most prominent at stimulus frequencies >100 Hz. Data sets of comparable size were collected 2, 4, and 48 h after TNF injection and showed similar changes (data not shown).

Figure 1C shows maximal tetanic force as a function of the time interval between TNF injection and muscle excision. Relative to contemporary data from naive animals (no injection), maximal force was depressed ∼25% after 1 h, a response that persisted for at least 48 h. Similarly, maximum forces of the TNF-treated group were less than forces of vehicle-treated controls over this same period.

Figure 2 illustrates the effect of TNF on intracellular oxidant activity in diaphragm muscle fibers. Oxidant activity was increased 1 h after TNF administration (6 of 6 paired comparisons, Fig. 2A) and returned to near-basal levels after 24 h (Fig. 2B). On the basis of time-course data, the diaphragm was isolated for analysis 1 h after TNF administration in all subsequent studies (below).

Fig. 2.

TNF increases intracellular oxidant activity in diaphragm muscle fibers. A: oxidant activity measured by use of 2′,7′-dichlorofluoroscin (DCFH) in fiber bundles isolated from diaphragm of mice 1 h after ip administration of vehicle or TNF. TNF increased oxidant activity in 6 of 6 individual comparisons (○) by 44% ± 20.9 (•, average response). DCF, 2′,7′-dichlorofluorescein. *P < 0.05 by paired t-test. B: effect of TNF on oxidant activity was not detectable 24 h after administration (symbols as in A). In A and in B, n = 6/group.

Antioxidant pretreatment.

TNF actions were inhibited by pretreating animals with Trolox, a hydrophilic antioxidant. Data in Fig. 3 show that Trolox pretreatment diminished intracellular oxidant activity in muscle fibers from TNF-treated animals (Fig. 3A) and abolished the decrement in specific force (Fig. 3B). Thus muscle-derived oxidants appear to be essential postreceptor mediators of TNF-induced force loss.

Fig. 3.

Trolox pretreatment opposes TNF effects. A: cytosolic oxidant activity measured 1 h after TNF administration by use of DCFH; ○, data from each pair of diaphragm fiber bundles; •, the average response. Oxidant activity was 42% ± 7.5 less in diaphragm fibers from animals pretreated with Trolox (10 mg/kg ip) (Trolox + TNF) than in fibers from animals pretreated with vehicle (vehicle + TNF); n = 4/group. *P < 0.05 by paired t-test. B: TNF depressed specific force of diaphragm from animals pretreated with vehicle (vehicle + TNF) relative to muscles of animals receiving vehicle alone (vehicle). Pretreatment with Trolox abolished TNF effects on specific force (Trolox + TNF). Values are means ± SE; n = 3/group. *P < 0.05 by repeated-measures ANOVA.

TNFR1 involvement.

We determined the receptor subtype that mediates TNF effects by comparing responses of muscles from mice deficient in either TNFR1 (TNFR1−/−) or TNFR2 (TNFR2−/−). Intracellular oxidant activity was stimulated by TNF, a response that was blunted by TNFR1 deficiency (Fig. 4A) . In 8 of 8 paired comparisons, oxidant activity was less in TNFR1−/− fiber bundles than in genetic controls (Fig. 4A, left), suggesting TNFR1 signaling stimulates oxidant production. Oxidant activity was not different between TNFR2−/− and control muscle (Fig. 4A, right).

Fig. 4.

TNF acts via TNF type 1 receptor (TNFR1) to increase oxidant activity and depress force. A: following TNF treatment, cytosolic oxidant activity in fiber bundles from muscle of mice deficient in TNFR1 (TNFR1−/−) (A, left) was lower than that of genetic controls [wild type (WT)] in 8 of 8 paired comparisons (○) by 18% ± 4.1 (•, average response); P < 0.01 by paired t-test. Activity in fiber bundles from mice deficient in TNF type 2 receptor (TNFR2−/− ) did not differ from genetic control values after TNF (A, right). B: relative to vehicle-treated controls (WT + vehicle), depression of specific force by TNF (WT + TNF) is abolished by TNFR1 deficiency (TNFR1−/− + TNF) but not by TNFR2 deficiency (TNFR2−/− + TNF). Values are means ± SE; n = 3/group. *P < 0.05 by repeated-measures ANOVA.

TNF effects on specific force also required TNFR1. TNF administration to TNFR2−/− and genetic control animals depressed specific force of isolated diaphragm fiber bundles (Fig. 4B), a response comparable to muscles of wild-type ICR mice (Fig. 1A). In contrast, TNF did not alter specific force of TNFR1−/− muscle (Fig. 4B), indicating TNFR1 is essential for TNF-induced dysfunction. Forces were not different among TNFR1−/−, TNFR2−/−, and genetic control muscles after vehicle injection (data not shown), suggesting subtype-specific receptor deletion per se does not alter contractile function.

TNF depresses myofibrillar force.

TNF effects on myofibrillar function were tested using chemically permeabilized muscle fibers from diaphragms of mice pretreated with TNF or vehicle. TNF decreased maximum specific force of calcium-activated fibers (Fig. 5A, Table 1), a response similar to that observed in electrically activated fiber bundles (Fig. 1A). TNF did not alter other aspects of myofibrillar function, including the calcium concentration required for half-maximum activation, Hill coefficient, rate of tension recovery (ktr), and maximum shortening velocity (Table 1), nor did TNF alter the relationship between activating calcium concentration and relative force (Fig. 5B).

Fig. 5.

Myofibrillar force is depressed in diaphragm of TNF-treated animals. A: maximal specific force of chemically permeabilized diaphragm fibers from TNF-treated animals (TNF) is less than maximal force of fibers from vehicle-treated controls (vehicle). Curves depict averages of Hill equation curves fitted to individual fibers. Averaged maximal forces (F0) of Hill equation curves differed between groups; other parameters did not (Table 1). Values are means ± SE; n = 15 fibers/group. *P < 0.02 by Student's unpaired t-test. B: data from A, reexpressed as relative forces; averaged Hill equation curves of the 2 groups (not shown) did not differ in any parameter. F0, maximal tetanic force.

View this table:
Table 1.

TNF effects on contractile parameters of permeabilized muscle fibers


These studies demonstrate that a single intraperitoneal injection of TNF depresses specific force of murine diaphragm by increasing intracellular oxidant activity. Data are consistent with our hypotheses that 1) TNFR1 binding mediates TNF effects on oxidant activity and force, and 2) TNF depresses the force generated by myofibrillar proteins.

TNF-induced contractile deficit.

Weakness of respiratory and limb skeletal muscle is a serious concern in many inflammatory diseases. Premature death and decreased quality of life correlate with muscle weakness in patients with chronic obstructive pulmonary disease (38), muscular dystrophy (13), cancer (26), and chronic heart failure (28). TNF is widely thought to contribute to such loss of function (2, 13, 31, 36).

TNF has catabolic properties that stimulate loss of muscle protein and muscle atrophy (11, 25). TNF also weakens muscle in the absence of atrophy. The latter mechanism appears to be mediated by TNF-stimulated oxidants. Pretreatment with the antioxidant Trolox attenuated oxidant activity and preserved force in our present study. These findings reinforce earlier observations that N-acetylcysteine, a nonspecific antioxidant and reduced thiol donor, also attenuates TNF effects on specific force (18).

In the present study, loss of specific force persisted 48 h after treatment, even though oxidant activity returned to basal levels within 24 h. These findings indicate that modifications to myofibrillar proteins persist beyond the initial insult. The present data are the first to demonstrate that TNF alters contractile function of the myofibrillar lattice, as shown by force loss in permeabilized fibers, and indicates that myofibrillar proteins are modified or damaged.

Receptor subtype specificity.

Skeletal muscle expresses two cell surface receptors for TNF: the 55-kDa TNFR1 (p55) and 75-kDa TNFR2 (p75) subtypes. TNFR1 contains an apoptotic death domain and stimulates oxidant production in nonmuscle cell types (19). TNFR2 does not contain a death domain and primarily transduces signals favoring cell survival (35). The only receptor subtype-specific response previously identified in skeletal muscle is TNFR1 involvement in protein turnover during cancer cachexia (20). The present study expands our knowledge of this biology, identifying TNFR1 as the receptor subtype by which TNF stimulates oxidant activity and force loss in muscle.

TNF depresses myofibrillar force.

In vitro incubation with TNF depresses tetanic force of intact muscle fibers without altering cytosolic Ca2+ transients (24). This observation suggested that TNF causes dysfunction downstream of the Ca2+ signal, i.e., at the myofilament level. The present data confirm this interpretation. Permeabilized fibers from TNF-treated muscle showed a reduction in calcium-activated force with no changes in Ca2+ sensitivity or cross-bridge cycling rate.

TNF depresses force by increasing intracellular oxidant activity, a previous observation (18) confirmed by the present study. Existing data do not identify the oxidants involved. Skeletal muscle fibers generate both reactive oxygen species (ROS) and nitric oxide (NO) derivatives (15, 23). The fluorochrome probe used to detect TNF-stimulated oxidants (DCFH) and the antioxidants that preserved force (N-acetylcysteine, Trolox) react with both ROS and NO derivatives. Positive results do not discriminate between these two oxidant cascades.

Prior reports have tested the direct effects of various oxidants on permeabilized fiber preparations. Exposure to a superoxide-generating system, hydroxyl radicals, or peroxynitrite, depresses maximum Ca2+-activated force without a shift in Ca2+ sensitivity (5, 29). Hydrogen peroxide (5–10 mM) depresses Ca2+-activated force in permeabilized fast fibers (17, 22), an effect not seen at lower concentrations (5, 22) or in permeabilized slow fibers (17, 22). The NO donor sodium nitroprusside depresses maximum force, alters Ca2+ sensitivity, decreases shortening velocity, and reduces ATPase activity, changes reversed by the thiol donor dithiothreitol (21). Thus both ROS and NO derivatives can depress force. Systematic studies are required to identify the oxidant cascade by which TNF acts.

The myofilament proteins affected by TNF are not known. One possible target is the troponin complex. Under hypoxia/fatigue conditions, degradation of troponin I and troponin C depresses maximal force in permeabilized fibers, a response attributed to increased ROS (10). Another possible target is the myosin heavy chain. Exposure to ROS and peroxynitrite can modify the myosin protein through sulfhydryl modification (16, 30). Tropomyosin is a candidate protein identified in cardiac muscle. Recent data show that contractile dysfunction after coronary microembolization is caused by tissue TNF production that leads to disulfide formation on tropomyosin, a process inhibited by antioxidant pretreatment (8).

Summary and potential relevance.

Reduction of specific force is a mechanism, independent of muscle atrophy, by which TNF causes weakness. The present data expand our understanding of this process, confirming a causal role for muscle-derived oxidants, identifying TNFR1 as the essential receptor subtype, and establishing myofilaments as the site of dysfunction. Other elements of cellular mechanism—postreceptor signaling events, oxidant type and source, myofibrillar protein target(s)—remain to be discovered.

Doing so is important. Muscle weakness complicates chronic inflammatory disease, increasing illness and promoting death. There is no clinical treatment for this problem, no drug or nutritional therapy that has proven effective. Continued research to define the mechanism of TNF action will identify potential targets by which muscle strength may be preserved or restored in chronic illness.


This research was supported by National Heart, Lung, and Blood Institute Grant R01-HL-59878 (M. B. Reid).


  • 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