Superoxide release from contracting skeletal muscle in pulmonary TNF-α overexpression mice

Abstract

Chronic obstructive pulmonary disease (COPD) often results in increased levels of tumor necrosis factor-α (TNF-α), a proinflammatory cytokine, which circulates in the blood. However, it is not clear whether pulmonary TNF-α overexpression (a COPD mimic) induces excessive reactive oxygen species (ROS) formation in skeletal muscle and thereby may contribute to the muscle impairment often seen in COPD. We hypothesized that ROS generation in contracting skeletal muscle is elevated when there is TNF-α overproduction in the lung and that this can induce muscle dysfunction. Cytochrome c (cyt c) in the perfusate was used to assay superoxide (O₂^·−) release from isolated contracting soleus muscles from transgenic mice of pulmonary TNF-α overexpression (Tg⁺) and wild-type (WT) mice. Our results showed that Tg⁺ muscle released significantly higher levels of O₂^·− than WT during a period of intense contractile activity (in nmol/mg wt; 17.5 ± 2.3 vs. 4.4 ± 1.3, respectively; n = 5; P < 0.05). In addition, the soleus muscle demonstrated a significantly reduced fatigue resistance in Tg⁺ mice compared with WT mice. Perfusion of the contracting soleus muscle with superoxide dismutase, which specifically scavenges O₂^·− in the perfusate, resulted in significantly less cyt c reduction, thereby indicating that the type of ROS released from the Tg⁺ muscles is O₂^·−. Our results demonstrate that pulmonary TNF-α overexpression leads to a greater O₂^·− release from contracting soleus muscle in Tg⁺ compared with WT and that the excessive formation of O₂^·− in the contracting muscle of Tg⁺ mice leads to earlier fatigue.

reactive oxygen species (ROS) are actively involved in physiological systems and are believed to play a key role in normal cell signaling events (8, 15). Specifically, ROS generation is associated with the biological response to conditions such as ischemia-reperfusion injury, infection, muscle fatigue, heavy metal poisoning, and ethanol toxicity (13, 36, 43). Previous studies have indicated that ROS formation in skeletal muscle increases during contractions (9, 12, 46, 61, 62). For instance, in vitro rat diaphragm releases moderate levels of ROS (46) at rest, and ROS generation increases significantly with repetitive muscle contractions (24, 46). Excessive ROS formation during strenuous exercise may overwhelm natural intracellular antioxidant defenses, initiate further muscular damage, and accelerate the fatigue process (3, 50).

Tumor necrosis factor-α (TNF-α) is a proinflammatory cytokine and may initiate muscular dysfunction as during chronic obstructive pulmonary disease (COPD) (18, 40). Previous research shows that an overexpression of TNF-α in the lungs of mice results in muscle wasting and impaired muscle regeneration (27). Clinically, patients with COPD experience systemic inflammation, in part as a result of elevated levels of TNF-α (18, 49), and it is postulated that skeletal muscle dysfunction and wasting may correlate with this inflammatory process (47). In addition, ROS are also produced during hypoxic stress that is associated with COPD progression (58, 59). However, a direct correlation between pulmonary TNF-α overexpression and ROS generation in a contracting skeletal muscle has not been well established. The understanding of a potential interaction between TNF-α and ROS is critical for developing innovative treatments for pulmonary disorders such as anti-TNF-α therapy in COPD (2, 4). Therefore, it is important to evaluate skeletal muscle ROS generation in a model of pulmonary TNF-α overexpression. In the current study, we used an in vitro soleus muscle model from transgenic (Tg⁺, pulmonary TNF-α overexpression) mice, which mimics many of the pathophysiological conditions of COPD (16, 60). We tested the hypothesis that pulmonary TNF-α overexpression would increase ROS release from skeletal muscle and decrease fatigue resistance. Additionally, we used the antioxidant superoxide dismutase (SOD) to test the hypothesis that one major species of ROS, superoxide (O₂^·−), is a primary ROS released from Tg⁺ skeletal muscle during muscle contractions.

METHODS

Animal care and whole muscle isolation.

Male SP-C/TNF-α Tg⁺ (provided by Dr. Charles G. Irvin from University of Vermont) were crossed with C57BL/6 mice and screened by PCR analysis (27). All procedures were approved by the University of California San Diego Institutional Animal Care and Use Committee (IACUC). Male mice aged 7–12 mo were used in this study. Before each experiment, each mouse was anesthetized by intraperitoneal administration of ketamine (70 mg/kg) and xylazine (10 mg/kg). The soleus muscle was carefully removed from both hind limbs and then mounted in an experimental chamber (model 800MS; Danish Myo Technology, Aarhus, Denmark). The chamber was filled with oxygenated Tyrode solution (in mM: 121 NaCl, 5 KCl, 0.4 NaH₂PO₄, 1.8 CaCl₂, 0.5 MgCl₂, 24 NaHCO₃, 5.5 glucose, 0.1 EGTA; pH 7.4, room temperature) containing 20 μM cytochrome c (cyt c) as an extracellular O₂^·− probe (Sigma, St. Louis, MO) to examine whether O₂^·− is released from tissue. The oxygenated solution has no observable effect on cyt c reduction in our model, which is consistent with former studies (24, 46, 57). For each soleus muscle, the tendon located at one end of the muscle was connected to a mobile lever arm designed to adjust muscle length. The other tendon at the opposite end of the muscle was attached to a stationary force transducer (force range between 0 and 1,600 mN; Danish Myo Technology).

Experimental procedure.

After mounting was completed, the muscle optimal length (L_o) was set and the muscle was at rest for 20 min at room temperature for equilibration. To simulate physiological temperature for the soleus muscle, the temperature was increased to 34–35°C, before a 12-min contraction timeframe and a subsequent 9-min rest period. Thus we had an optimum time window to examine the possible O₂^·− release in both contracting and resting states. The soleus muscle was stimulated electrically (S48 stimulator; Grass Technologies, West Warwick, RI) using square-wave pulses. Our contractile protocol consisted of a series of repeated tetanic contractions with increasing train frequencies (0.1, 0.125, 0.166, 0.25 contractions/s; 40 V, 500-ms train duration, 1-ms pulse duration, 80 Hz), each of which lasted for 3 min. An A-D converter (model MP100WSW; Biopac Systems, Santa Barbara, CA) converted analog data to digital data, and results were analyzed using Acknowledge III 3.2.6 software (Biopac Systems, Santa Barbara, CA).

Extracellular ROS study.

Although it is difficult to detect O₂^·− produced in muscle because of its short half-life, our laboratory developed methods for identifying O₂^·− formation in skeletal muscle (56, 57). In the current study, the cyt c assay was used to measure O₂^·− release from the soleus muscle, following the combined methods described by Margoliash and Frohwirt (32), Reid et al. (46), Kolbeck et al. (24), as well as our former studies (56, 57). Through a one-electron transfer reaction, O₂^·− reduces cyt c, resulting in an increase of absorbance at 550 nm with an extinction coefficient of 18.5 × 10 M⁻³·cm⁻¹ (24, 26, 33, 57). SOD (Sigma) is an effective O₂^·− scavenger that determines the specificity of the cyt c assay for O₂^·− release. Both WT and Tg⁺ muscles were incubated with cyt c (20 μM) in the presence or absence of SOD (1,500 U/ml). This concentration of SOD is within an optimal level to scavenge O₂^·− released from the tissue and is consistent with previous studies in this type of muscle (56, 57). The measurement of cyt c reduction was obtained by taking the difference between the peak at wavelength of 550 nm and the averaged baseline of 540 and 560 nm, as described by Kolbeck et al. (24). The spectrum of cyt c was recorded using a NanoDrop 2000 microvolume spectrophotometer (Thermo Scientific).

Statistical analysis.

Data were analyzed using a multiway ANOVA and expressed as means ± SE (JMP, SAS Institute, NC). The animal was treated as a random variable, and the differences in mean values between treatment and nontreatment were determined by post-ANOVA contrast using SAS JMP software. P < 0.05 was considered to be significant.

RESULTS

The soleus muscle weight as well as the body weight was substantially decreased in Tg⁺ mice compared with WT mice (in mg, 8.50 ± 0.25 vs. 10.3 ± 0.20 for soleus, n = 6, P < 0.01; in g, 30.2 ± 1.33 vs. 39.8 ± 1.48 for body weight, n = 6, P < 0.01; Table 1). As shown in Fig. 1, lung overexpression of TNF-α resulted in an increase in cyt c reduction (indicative of O₂^·− release from the muscle) as early as 6 min after the start of contractions while no marked increase was observed in WT at this time point (n = 5; P < 0.05). The amount of cyt c reduction reached a maximum at ∼9 min from the start of the contractions. The difference between Tg⁺ and WT muscles continued until the end of the contraction period at 12 min (in nmol/mg wt; 17.5 ± 2.3 vs. 4.4 ± 1.3 respectively; n = 5; P < 0.05) and remained significantly different up to the end of the subsequent rest period at 21 min. It is worth noting that O₂^·− levels still increase in WT mice during contractions, but the increase is not as marked as in Tg⁺ muscle. For instance, in WT mice at 0 min during the contraction, the cyt c reduction is 0 nmol/mg wt, but at 9 min during the contraction, the mean cyt c reduction is 5.24 nmol/mg wt, which is substantially lower than 17.2 nmol/mg wt in the Tg⁺ group (Fig. 1). The prevention of cyt c reduction in the Tg⁺ muscle treated with SOD was different from Tg⁺ muscle not treated with SOD, confirming that O₂^·− was the major ROS released during contractions in our model (n = 5–7; P < 0.05). Specifically, at the end of the 21-min protocol, 1.8 ± 1.6 nmol cyt c/mg wt were reduced in the Tg⁺ group treated with SOD compared with 17.8 ± 2.3 nmol cyt c/mg wt in Tg⁺ muscles not treated with SOD (n = 5–7; P < 0.05). In addition, SOD in the perfusate had no effect on cyt c reduction in WT group as there was little O₂^·− release from WT muscles (n = 7).

Table 1.

Mouse whole body and soleus muscle weight

	Wild Type (n = 6)		Tg+ (n = 6)
	Soleus weight, mg	Body weight, g	Soleus weight, mg	Body weight, g
	10.1	41.5	9.0	26.8
	11.0	38.2	8.2	28.8
	10.3	45.6	7.5	31.3
	10.8	40.9	8.4	28.5
	9.7	37.4	9.2	29.6
	10.0	35.4	8.7	36.1
Average ± SE	10.3 ± 0.20	39.8 ± 1.48	8.5 ± 0.25*	30.2 ± 1.33*

n = number of mice.

Tg⁺, trangenic overexpression.

Weights were recorded at 11 mo of age.

^*Significant difference compared with wild type is identified: P < 0.01.

Fig. 1.

Cytochrome c (cyt c) reduction during soleus muscle contractions in wild-type (WT) and transgenic (Tg⁺) mice with and without superoxide dismutase (SOD) treatment. # and *Results in Tg⁺ mice which were significantly different compared with WT with and without SOD and Tg⁺ with SOD, respectively.

Figure 2 demonstrates two typical function curves, each of which represents the percentage of force decline of both Tg⁺ and WT muscles. Note that the duration of contractions was the same for both animal groups. Relative force (percentage of initial force) was recorded at different time points of the 12-min contractile period. Force declined more rapidly in Tg⁺ compared with WT muscles. The time (s) to reach 70% of initial force (T_70%) was shorter in Tg⁺ than WT mice. As shown in grouped data in Fig. 3, T_70% was significantly reduced by ∼35% in Tg⁺ muscle compared with WT muscle (in s, 357.4 ± 44.9 vs. 562.5 ± 27.3 respectively; n = 7–8, P < 0.01). In comparison, our colleagues in the same lab have used a different contractile protocol demonstrating that T_50% was significantly shortened in Tg⁺ compared with the WT mice, which is consistent with our T_70% results as shown in Fig. 3 (53). Moreover, our results are supported by a representative figure (Fig. 2) displaying the difference in contraction curves for each group and a function figure (Fig. 5) using initial contraction and ending contraction values for statistical comparison of muscle activity.

Fig. 2.

Representative data demonstrating the percentage of initial force from mouse soleus muscle during the 12-min contractile period. T_70%, time to 70% of the initial force. A: WT. B: Tg⁺.

Fig. 3.

Grouped data of T_70% in WT and Tg⁺ mice. *Results in Tg⁺ mice were significantly different compared with WT.

Fig. 5.

Grouped data of percentage of initial force in WT and Tg⁺ mice. *Each end contraction was significantly different compared with that of the first contraction. #Last contraction of Tg⁺ was significantly different from the last contraction of WT.

Furthermore, we tested the rate of O₂^·− release (nmol·mg wt⁻¹·min⁻¹) from soleus muscle as illustrated in Fig. 4. Tg⁺ muscle demonstrated a significantly higher rate of O₂^·− release than WT at 6 and 9 min, respectively (n = 5; P < 0.05, Fig. 4A). Correspondingly, the rate of force decline (% of initial force/min) in Tg⁺ muscle was also significantly greater than in WT at 6 and 9 min, respectively (n = 7–8; P < 0.05, Fig. 4B).

Fig. 4.

A: grouped data demonstrating the cyt c reduction rate in WT and Tg⁺ mice (nmol·mg wt⁻¹·min⁻¹) at 3, 6, and 9 min during the contraction. *Results in Tg⁺ mice were significantly different compared with WT. B: grouped data showing the rates of force decline from the normalized initial force (%min⁻¹) in WT and Tg⁺ mice at 3, 6, and 9 min. *Results in Tg⁺ mice were significantly different compared with WT.

Grouped fatigue data are shown in Fig. 5. Contractile force at the end of contractions was significantly reduced compared with the first contraction in each respective muscle group (n = 6–8; P < 0.05). The percentage of initial force at the end contraction in Tg⁺ group was significantly lower than that of WT (n = 7–8; P < 0.05). Our data in Fig. 5 show that Tg⁺ muscles have significantly more fatigue at the end of contractions than WT muscles. In addition, Tg⁺ muscle also produced higher levels of O₂^·− than WT, as shown by cyt c reduction (Fig. 1). The muscle force starts to decline rapidly after 3 min from the initial contraction, consistent with the timeline of O₂^·− release rate shown in the Tg⁺ muscle (Fig. 4).

Although Figs. 1 and 4A are somewhat similar, they are two different notions: total cyt c reduction in Fig. 1 and cyt c reduction rate in Fig. 4A. Cyt c reduction refers to the net accumulation of reduced cyt c over time, indicating the overall O₂^·− released in a certain time period. The cyt c reduction rate refers to the amount of cyt c that is reduced per unit time at a specific time point. For example, at 9 min in the WT curve, the value for the cyt c reduction is cumulative and substantial (5.2 nmol/mg wt, shown in Fig. 1), but the cyt c reduction rate, at this same time point, is much smaller (0.28 nmol·mg wt⁻¹·min⁻¹, shown in Fig. 4A). Therefore, both Figs. 1 and 4A are addressing two different concepts.

DISCUSSION

Our results demonstrate that skeletal muscle contractile activity in lung-only TNF-α overexpressing mice leads to an increased O₂^·− release from soleus muscle and subsequent reduced resistance to fatigue.

Previous research has shown that the development of COPD can induce muscle atrophy (41). This causes peripheral muscles of COPD patients such as the quadriceps to fatigue rapidly (17). TNF-α, a proinflammatory cytokine, which is substantially elevated in COPD subjects (18, 49), has been associated with muscle wasting in COPD patients (11, 49, 52). Since pulmonary inflammation is one of the major symptoms in COPD, TNF-α could affect the function and redox condition of peripheral muscle through circulation (5, 28). It appears that this cytokine may be a key factor in reducing contractile function of the heart, diaphragm, and limb muscles (28, 45, 59). Specifically, pulmonary TNF-α overexpression has been shown to induce peripheral skeletal muscle dysfunction (60) and impair muscle regeneration (27). However, the mechanism by which this occurs has not been fully elucidated. Moreover, former studies have shown that ROS play an important role in skeletal muscle function (24, 46). Therefore, we have investigated the potential link between TNF-α and ROS formation in skeletal muscle. We used a pulmonary TNF-α overexpression (Tg⁺) mouse to model COPD conditions, as previously described (16, 60), and to delineate the role of ROS in a contracting peripheral skeletal muscle (soleus) affected by TNF-α overexpression in the lung.

Recently, Wagner's group (53) has shown that the TNF-α levels in the lung and serum of Tg⁺ mice were markedly greater than WT mice. They also demonstrated that the TNF-α level in the soleus of Tg⁺ mice was significantly greater compared with WT mice (53). In addition, the level of TNF-α mRNA was significantly elevated in the Tg⁺ than in WT mice (53), which is consistent with a study conducted by Langen et al. (27) who showed similar findings using the same Tg⁺ model as the present study. The researchers concluded that the TNF-α protein levels in the lung bronchoalveolar lavage (BAL) fluid of the Tg⁺ mice were approximately sixfold higher compared with WT. This study also determined that serum TNF-α protein levels were significantly greater in the Tg⁺ mice compared with WT mice. Furthermore, they found that mRNA values for TNF-α in soleus muscle were substantially increased compared with WT mice, which is a similar finding to Wagner's recent study (53). These results provide evidence that the Tg⁺ mice used in our experiments have substantially elevated TNF-α protein levels in the lungs, soleus muscle, and serum (or circulation).

It is worth noting that Tg⁺ mice had a ∼10% decrease in soleus weight and ∼25% decrease in body weight compared with WT mice (Table 1), which is consistent with previous studies (27, 53), suggesting the potential damage to the skeletal muscle and other pathophysiological changes in the body initiated by TNF-α. Yet, neither of these studies determined TNF-α receptor protein abundance in the skeletal muscle. Thus, this subject of interest requires further investigation.

Earlier research from the Reid laboratory demonstrated that cardiac overexpression of TNF-α causes oxidative stress and skeletal muscle dysfunction (28). In our pulmonary TNF-α overexpression model, the result is consistent with previous research which showed that increased cyt c reduction in the supernatant is attributed to O₂^·− release from stimulated muscle cells (34, 35, 39). In studies by both Kolbeck et al. (24) and Reid et al. (46), the cyt c assay was used as a sensitive tool to identify O₂^·− released from resting muscle and during muscle contractile activity. An advantage of using this cyt c assay is that it is an effective extracellular O₂^·− probe and its relatively large size (12.4 kDa) prevents its transport across the sarcolemma into intracellular compartments (46). This method has also proven reliable in other physiological models, such as brain injury under hypoxic conditions (14) and dopamine oxidation in the nervous system (23). However, cyt c has a potential disadvantage because it may react with hydroxyl radicals and certain species such as ascorbate, which can markedly reduce cyt c, thereby yielding measurements in which the cause of the cyt c reduction is uncertain. Thus, we used SOD in the muscle perfusate during several experiments to ensure that the cyt c assay precisely measured O₂^·− released from the myocytes.

There are several proposed cellular sources for O₂^·− release in skeletal muscle. One likely source is the mitochondrion, since it has been reported that elevated levels of TNF-α not only cause mitochondrial dysfunction but also stimulate mitochondria to produce ROS (48). The basal levels of ROS are believed to be produced in the mitochondria under normal metabolic conditions (7). Under certain pathological conditions, mitochondrial O₂^·− formation may increase (6), likely exiting via anion channels, and eventually released into surrounding extracellular spaces (31). In our study, we speculate that TNF-α produced in the lung induces O₂^·− release by damaging mitochondria in skeletal muscle via the circulation path as previously described (48). However, the exit of O₂^·− from mitochondria to extracellular matrix is limited by several factors: 1) O₂^·− is a highly unstable free radical with a very short half-life; O₂^·− exit across a cellular membrane may require a longer time; 2) The mitochondrial SOD as well as other intracellular antioxidants/enzymes may scavenge O₂^·− before its exit; 3) O₂^·− is a charged anion that limits its exit via the lipid membranes, including both the mitochondrial and cell membrane; and 4) O₂^·− most likely requires the activation of anion channels to facilitate its transport across the membrane; however, the anion channels may not always be open for O₂^·−. Still, it is possible that mitochondrial O₂^·− may exit via anion channels and eventually be released into surrounding extracellular spaces (31) via an unknown mechanism that requires additional studies. Moreover, using intense muscle stimulation protocols, other studies have shown that mitochondria are not the only source of O₂^·− release (34, 35, 39). For example, the flavoprotein oxidoreductase system located in the plasma membrane may directly account for the O₂^·− released in cultured skeletal muscle cells (39). Therefore, it is still uncertain whether the O₂^·− released from contracting muscle was generated primarily from mitochondria.

Additional research has demonstrated that endothelial cells are prospective sources of O₂^·− production via NADH and/or NADPH oxidases (NOXs) on the membrane (10, 37). However, this is arguable since there is no blood flow-related endothelium shear stress (SH), and thus potential SH-induced ROS would not occur in our isolated muscle preparations (30). NOX is also expressed in skeletal muscle cells (22), and this could be considered as another possible pathway for ROS generation. Moreover, other ROS sources may be involved. For example, the release of O₂^·− can also be linked to arachidonic acid (AA) in the cell membrane. Metabolism of AA, through lipoxygenase (LOX) activity, is responsible for O₂^·− release during oxidative stress in skeletal muscle (56). Through a mechanism that is not fully understood, these AA enzymes may likely transfer electrons to extracellular oxygen or pass O₂^·− directly across the sarcolemmal membrane via anion channels (20, 31, 56). Therefore, further study is needed to delineate the source of O₂^·− release in our Tg⁺ model.

Investigations to identify the potential intracellular sources of ROS with permeable antioxidant treatments, such as N-acetylcysteine, have revealed that oxidative stress may be a causative factor in the fatigue process (38, 43, 51). Reid et al. showed that diaphragm myocytes release O₂^·− into the interstitium and extracellular space and this rate rapidly increases in fatiguing muscles (46). Their finding is consistent with O₂^·− decreasing muscle function during fatigue, which was also seen in our protocol. However, our data showed that the application of extracellular SOD treatment did not mitigate fatigue of either WT or Tg⁺ muscle (data not shown). This suggests, as we expected, that SOD present in the perfusate had no influence on intracellular ROS formation due to the fact that SOD is a large protein and cannot access the intracellular compartment from the extracellular perfusate solution (24, 57). Accordingly, we speculate that scavenging of O₂^·− released from soleus muscle did not alter intracellular ROS levels, indicating that intra- and extracellular ROS formation mechanisms are somehow independent. Yet, the use of SOD in the perfusate allowed for the identification of O₂^·− as the primary ROS generated in contracting soleus muscle. Previous research has shown that SOD (at the similar level of concentration used in the current study) has no marked interaction with cyt c, suggesting minimal effect on cyt c reduction from the enzymatic action of SOD (57). Therefore, there is no need to use denatured SOD to confirm the cyt c signal (57).

Former studies have provided evidence of a substantial increase of ROS during muscle contractions (12, 21, 24, 25, 46). In response, endogenous antioxidant defense systems modulate ROS levels via enzymes such as SOD and catalase (1, 29, 42). These protective enzymes work synergistically to convert O₂^·− to O₂ and water, and thus reduce oxidative stress (42). Previous data from Patwell et al. (39) demonstrated that SOD decreases the contraction-induced reduction of cyt c in muscle cells, thus confirming that O₂^·− is responsible for this proportion of cyt c reduction, which is consistent with our current result. We also observed that SOD in the muscle perfusate had no significant effect on preventing muscle fatigue. It is likely that during contractions, the activity of exogenous SOD in the perfusate is exclusively functioning outside the cell (44, 46). SOD, as mentioned previously, cannot readily access any intracellular or interstitial compartments (24), thereby limiting its ability to abolish any possible intracellular O₂^·− at the site where it may affect intramuscular proteins, as proposed by Kolbeck et al. (24). On the contrary, we previously observed beneficial effects using the SOD mimic Tiron, which has a much smaller size than SOD and is membrane permeable. Accordingly, our previous data showed that Tiron reduces oxidative stress because it scavenges O₂^·− directly from intracellular sources (57). Therefore, the production of intracellular O₂^·−, likely unaffected by perfusion with SOD, resulted in the lack of any significant improvement from SOD in both WT and Tg⁺ muscle contractions (data not shown).

We did not observe any marked reduction of cyt c except for a small increase from background in the absorbance spectra before any muscle contraction. However, this increase can be easily corrected through a calculation that has been used previously (24, 57), particularly designed for skeletal muscle experiments. The measurement of cyt c reduction was obtained by taking the difference between the peak value at the wavelength of 550 nm and the averaged baseline values at 540 and 560 nm. Thus, the background noise is minimized. However, some researchers use the absolute value of cyt c reduction at 550 nm to monitor O₂^·−. The advantage of this second method is that it provides accurate measurements when the background solution is relatively clean. When performing muscle tissue experiments, the background absorbance may be substantially increased, possibly due to the multiple substances released from the muscle, causing cyt c reduction artifact. Therefore, the first method is preferred in our study.

It is noted that, by the time the mice were studied (older than 6 mo of age), the muscles were overexpressing TNF-α. This is supported by a recent study from Wagner's group, which has shown that the level of TNF-α in the mouse soleus is significantly higher in the Tg⁺ compared with WT mice (53). The study further showed significantly increased TNF-α levels as well as increased mRNA levels for TNF-α in other types of skeletal muscles such as extensor digitorum longus. Wagner's group concluded that the increased circulating TNF-α (released by the lung) may play a role in stimulating the expression of TNF-α in the skeletal muscle (53).

In the present study, we seek to determine how much O₂^·− is released from the muscle as a way of indirectly measuring superoxide generated inside the muscle. It is technically difficult to measure the intracellular O₂^·− in a contracting soleus muscle due to the large motion artifact. To confirm that the signal is O₂^·−, we used cell-impermeable SOD and expected that the addition of SOD would have no marked effect on the intracellular ROS formation, since it cannot penetrate the membrane to access the myocytes. On the contrary, Tiron, another O₂^·− scavenger, can penetrate the membrane and remove the intracellular O₂^·− as shown in our previous study (57). However, Tiron may chemically react with cyt c to cause an inaccurate measurement, preventing its application in the current study (57). Moreover, it is not relevant to use cell-permeable SOD mimic for the intracellular study, since the scope of the present study is focused on extracellular O₂^·− release, not on the intracellular O₂^·− generation.

Finally, enhanced levels of TNF-α are correlated with reduced function of oxidative enzymes, decreased oxygen availability, and energy usage (54). It is likely that there is a molecular mechanism for the decrease in myocyte function by TNF-α through the downregulation of the skeletal peroxisome proliferator-activated receptor γ coactivator 1-α (PGC-1α), a transcriptional cofactor (54). It was demonstrated that amplified levels of TNF-α, induced by cigarette smoke (19), caused a decrease in PGC-1α, thereby affecting cell signaling cascades that regulate muscle function and O₂ transport. This may explain in part why smoking increases muscle atrophy and induces early fatigue via TNF-α (54). Since TNF-α level is markedly increased in Tg⁺ mice, it is possible that PGC-1α may play a similar role in our soleus muscle model. Moreover, the timeframe for rate of O₂^·− release mostly matches that for rate of the force decline in both Tg⁺ and WT muscles (Fig. 4). This suggests a possible linkage between increased O₂^·− formation/release and enhanced muscle fatigue in Tg⁺ muscle. Interestingly, as early as 3 min, the rate of force decline in Tg⁺ muscle was greater than that of WT muscle. Yet, there is a slight delay in the rate of O₂^·− release compared with the rate of force decline in Tg⁺ muscle as shown in Fig. 4, which may be due to the possibility that it may take time for muscle to initiate O₂^·− release after contraction. Since previous research also demonstrated that there is a known correlation between increased levels of ROS and increased levels of fatigue (46), it is likely that Tg⁺ muscle experienced a faster rate of fatigue because more O₂^·− was generated compared with WT muscle. Interestingly, there is no significant difference between initial absolute force of Tg⁺ muscle and that of WT muscle, which is because the muscle is not yet fatigued at the onset of contractions (data not shown).

Perspectives and Significance

This study shows that pulmonary TNF-α overexpression plays a significant role in the elevated levels of O₂^·− in the skeletal muscle of Tg⁺ mice and that this results in reduced skeletal muscle fatigue resistance. Furthermore, O₂^·− release is a potential indicator of Tg⁺ muscle fatigue, which is consistent with previous findings showing that skeletal muscle that develops more fatigue also releases more O₂^·− (46). COPD patients have diminished resistance to muscle fatigue, and in addition, some show significant skeletal muscle wasting, which has been related to the presence of inflammatory markers (55). Taken together, these findings suggest that pulmonary overexpression of TNF-α is a potential underlying contributing factor to muscle dysfunction in COPD.

GRANTS

This research was supported by National Heart, Lung, and Blood Institute Grant PPG-HL-091830, National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR040155, Oakland University General Fund G110, Oakland University Research Excellence Fund of Biomedical Research, and Ohio State University Health Rehabilitation Sciences Fund 013000.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: L.Z., P.D.W., and M.C.H. conception and design of research; L.Z. performed experiments; L.Z., A.H.H., W.J.R., P.D.W., and M.C.H. analyzed data; L.Z., A.H.H., W.J.R., P.D.W., and M.C.H. interpreted results of experiments; L.Z. and W.J.R. prepared figures; L.Z., A.H.H., and W.J.R. drafted manuscript; L.Z., A.H.H., W.J.R., P.D.W., and M.C.H. edited and revised manuscript; L.Z., A.H.H., W.J.R., P.D.W., and M.C.H. approved final version of manuscript.