Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A series of recent publications indicate that sepsis results in profound reductions in respiratory muscle force-generating capacity (1). This previous work also indicates that several free-radical species (nitric oxide [NO], superoxide, hydroxyl radicals) play a pathogenic role in the induction of sepsis-associated skeletal muscle dysfunction (1). NO synthase (NOS) activity of the respiratory muscles is increased in sepsis, and generation of reactive oxygen species (e.g. superoxide) by the contracting diaphragm is enhanced during sepsis (4, 6). In addition, administration of either inhibitors of NOS or free-radical scavengers (i.e. superoxide dismutase [SOD], catalase, dimethyl sulfoxide [DMSO], N-acetylcysteine [NAC]) has been shown to prevent sepsis-related skeletal muscle dysfunction (1, 2, 4).

The precise mechanism by which free radicals alter muscle performance in sepsis, however, is not known. One possibility is that free-radical species react directly with skeletal muscle contractile proteins during the induction of sepsis, altering actin-myosin interactions and reducing contractile protein force-generating capacity. Alternatively, it is possible that free radicals either alter mitochondrial function or damage the muscle structures involved in excitation- contraction coupling. Such latter forms of muscle damage could impair sarcoplasmic reticulum calcium handling during contraction, thereby limiting contractile protein activation and force generation during contraction.

The purpose of the present study was to investigate this issue by testing the hypothesis that free radical-induced skeletal muscle contractile protein dysfunction occurs during the development of sepsis and is of sufficient magnitude to significantly impair muscle force-generating capacity. For these experiments, contractile protein function was assessed by measuring the force generated by single Triton-skinned rat diaphragm muscle fibers in response to immersion in calcium-containing solutions (the Triton skinning procedure used for these studies removes the sarcolemmal, sarcoplasmic reticulum, and mitochondrial membranes, resulting in exposure of the contractile proteins and permitting direct activation by calcium solutions) (8). Comparison was made of force-versus-log [Ca²⁺] (pCa) relationships between fibers removed from control animals, from endo-toxin-treated animals, from groups of animals given both endotoxin and agents that inhibit/scavenge free radicals (i.e., polyethylene glycol [PEG]-SOD], a superoxide scavenger; and N-nitro-L-arginine methylester [L-NAME], a NOS inhibitor), and from appropriate control groups. We postulated that endotoxin administration would evoke significant reductions in the force-generating capacity of the contractile proteins, and that administration of SOD or a NOS inhibitor would preserve contractile protein function.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Solutions

The composition of all solutions used for skinned-fiber force determinations was calculated using a computer program (Borland International, Scotts Valley, CA) that takes into account stability constants and stock solutions to produce final solutions of the correct ionic strength and pCa (9). The skinning solution used was composed of (in mM): 50.0 potassium methane sulfonate, 15.0 phosphocreatine, 10.0 ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid (EGTA), 1.0 Mg²⁺, 2.0 magnesium adenosine triphosphate (MgATP), 20.0 imidazole, ionic strength 150.0, pCa > 8.5, and pH 7.0 at 22°C. When needed for storage of muscles, a cytidine-5-triphosphate (CTP) relaxing solution consisting of 110.0 mM potassium methane sulfonate, 5.0 mM EGTA, 1.0 mM Mg²⁺, 2.0 mM MgCTP, 20.0 mM imidazole, ionic strength 150.0, pCa > 8.5, and pH 7.0 was used. To prevent phosphorylation of the myosin light chains in these stored muscles, CTP was used in place of ATP in the storage solution because CTP is not a substrate for myosin light chain kinase. All solutions contained protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM leupeptin, 10 µM aprotinin, and 1.0 mM benzamidine) to prevent muscle breakdown during experimental manipulation.

Experimental Procedures

Studies were performed using single muscle fibers from the diaphragm of 45 adult male Sprague-Dawley rats. Three groups of studies were performed: (1) experiments in which we compared the force generation of skinned fibers from control and endotoxin-treated animals and also examined the effects of a superoxide scavenger and a NOS inhibitor on endotoxin-mediated contractile dysfunction; (2) experiments in which we determined whether it was possible to reverse endotoxin-mediated reductions in contractile protein force generation in vitro by incubation of skinned fibers with reducing agents, free-radical scavengers, or a NOS inhibitor; and (3) experiments in which we determined whether endotoxin administration altered the sodium dodecyl sulfate polyacrylamide gel eletrophoresis (SDS-PAGE) protein electrophoretic pattern for skinned fibers.

For the first group of studies, rats were divided into the following experimental groups: (1) control rats (n = 5) injected intraperitoneally with saline, (2) animals injected with endotoxin (n = 5), (3) animals injected with both endotoxin and L-NAME (n = 5), (4) animals injected with both endotoxin and PEG-SOD (n = 5), (5) animals injected with L-NAME alone (n = 5), (6) animals injected with PEG-SOD alone (n = 5), and (7) animals injected with both endotoxin and denatured PEG-SOD (n = 5). The groups receiving endotoxin were injected with 8 mg/kg of Escherichia coli lipopolysaccharide (Sigma Chemicals, St. Louis, MO), controls were given 1 ml saline intraperitoneally, L-NAME was administered at 1 mg/kg intraperitioneally, and PEG-SOD was administered at 2,000 U/kg intraperitoneally. Denaturing of PEG-SOD, used in the last experimental group, was accomplished by heating the enzyme at 100°C for 30 min. At 18 h after injection of these various agents, all animals were anesthetized with pentobarbital (50 mg/ kg) and their abdomens opened. Diaphragm muscles were then dissected free, rinsed in skinning solution, and, if not studied immediately, transferred to a 50% CTP relaxing/50% glycerol solution (containing protease inhibitors) for storage at 20°C. Previous studies have indicated that fibers isolated from muscles and stored in this fashion are usable for single-fiber assessment for at least three months (10). For this first group of studies, 7 fibers/animal were isolated and the force-versus-pCa relationship for each fiber was assessed (details of force-versus-pCa curve determinations are provided later). All of these determinations were completed within 2 d of animal death.

For the second group of studies, we examined skinned muscle fibers isolated from the diaphragms of control and endotoxin-treated animals; the details of saline/endotoxin administration and fiber harvesting were as presented in the previous paragraph. For each fiber, an initial force-versus-pCa curve was obtained and fibers were then incubated for 5 min in relaxing solution (pCa 8.5) containing either no agent, SOD (2,000 U/liter), L-NAME (1 mg/liter), NAC (150 mg/liter) or dithiothreitol (DTT) (10 mM). At the end of each incubation period, a second force-versus-pCa curve was constructed.

For the third group of studies, we examined the SDS-PAGE electrophoretic patterns for skinned fibers from control, endotoxin-treated, endotoxin/PEG-SOD, and endotoxin/L-NAME animals.

Force-versus-pCa Determinations

At the time fiber characteristics were assessed, muscles were removed from the relaxing/glycerol solution, placed in a petri dish containing fresh skinning solution (including protease inhibitors), and allowed to warm to room temperature (11). Small bundles of approximately 10 fibers were then separated from the whole muscle by gently pulling on one end of the muscle with a pair of fine-tipped forceps while the other end of the muscle was held stationary with a second pair of forceps. Fiber bundles were immersed for 30 min in 0.1% Triton X-100, an ionic detergent which permeabilizes membranes, exposing the contractile proteins.

After incubation, bundles were removed from the Triton X-100 and placed back in the skinning solution, and individual fibers were teased away from bundles. The ends of fibers were carefully wound around two platinum posts that protruded into the skinning solution. One of these posts was attached to a micro force transducer (Harvard Apparatus, South Natick, MA) mounted on a movable gear assembly, while the other was fixed to a stationary metal support. Sarcomere length was set to 2.6 µm using the defraction pattern from a helium-neon laser. All of the procedures involving the manipulation of single muscle fibers were conducted with the aid of a dissecting microscope.

Force-versus-pCa curves were then constructed by immersing fibers in solutions of increasing calcium concentrations and recording tension on a strip recorder. Once peak tension was achieved in a given solution, fibers were rapidly submerged in the next solution by means of a spring-loaded Plexiglas tray. Initially, fibers were submerged in a relaxing solution containing no added calcium (pCa 8.5). Each fiber was then sequentially exposed to 13 different calcium solutions: pCa 6.0, 5.90, 5.80, 5.75, 5.70, 5.65, 5.60, 5.55, 5.50, 5.40, 5.30, 5.20, and 5.0. After exposure to pCa 5.0, fibers were transferred back to the relaxing solution. The diameter of fibers was measured using a micrometer. Fibers were stored in SDS sample buffer at 70°C and later typed using gel electrophoresis as described later.

Fiber Typing

The myosin heavy chain (MHC) isoform of each single fiber studied was determined using the SDS gel electrophoresis procedure of Talmadge and Roy (12). After boiling in SDS sample buffer for four minutes, MHCs were separated using an 8% running gel and a 6% stacking gel, both of which contained glycerol. Gels were run at 70 V for 20 h in a cold room. After separation, MHC isoforms were visualized by Coomassie blue staining. Myosin isoforms were identified by comparison with the MHC separation pattern of extracted myosin samples from normal rat diaphragm (these samples contained all MHC isoforms) and by comparison with previously published data (13, 14).

Protein Assessment using SDS-PAGE

Although the single-fiber electrophoresis described in the previous paragraph is necessary for fiber typing, this procedure does not result in loading of adequate amounts of protein for comparison of protein quantities across the various experimental groups. To provide this latter assessment, we performed a series of SDS-PAGE gels in which we examined large numbers of skinned fibers. Equal protein concentrations from approximately 50 skinned fibers obtained from the various experimental groups in which contractile protein functional abnormalities were detected were loaded in adjacent lanes (i.e., we compared electrophoretic patterns between fibers from control, endotoxin-treated, endotoxin/PEG-SOD-treated, and endotoxin/L-NAME-treated groups of animals). SDS-PAGE gels were run according to the description provided in the previous section, with the exception that 8 to 16% gradient gels were used; a protein ladder of known molecular weights was run with each gel, and computer analysis was performed to estimate the molecular weights of proteins found to be altered across the various experimental groups (using SigmaGel software).

Calculations

Maximum force was normalized to a cross-sectional area; for this calculation, fibers were assumed to be cylindrical and fiber diameter was used to calculate the cross-sectional area. The maximum force-generating capacity (Fmax) of fibers was expressed in kiloPascals (kPa). Using computer software (SigmaPlot; SPSS Inc., San Rafael, CA), force data were fit to the Hill equation:

%Fmax = 100[Ca²⁺]ⁿ/{[Ca²⁺]ⁿ + (Ca₅₀)ⁿ}

where Ca₅₀ is the Ca²⁺ concentration required to produce half-maximal activation and n is the Hill coefficient (15). The Ca₅₀ was used as an index of calcium sensitivity and the Hill coefficient was used as a measure of the myofilament cooperativity.

Statistical Analysis

Comparisons of measurements made on a single parameter (e.g., maximum fiber force generation) between two or more experimental groups were made using analysis of variance (ANOVA), with post hoc testing (Student-Newman-Keuls) to determine statistical differences between individual groups. Comparisons of force-versus-pCa curves from different groups were made using a repeated measures ANOVA. A P value of less than 0.05 was taken as indicating statistical significance. Parameter values are presented as means ± 1 standard error of the mean (SEM).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of a Superoxide Scavenger (PEG-SOD) on Endotoxin-Mediated Alterations of Single Triton-Skinned Diaphragm Muscle Fiber Force-versus-pCa Curves

Endotoxin administration resulted in a large reduction in the force-generating capacity of the contractile proteins of single Triton-skinned diaphragm muscle fibers. This alteration was characterized by a reduction in both the average Fmax of Triton-skinned diaphragm fibers from endotoxin-treated animals (71.58 ± 2.47 kPa) as compared with Triton-skinned fibers from control animals (117.22 ± 3.64 kPa; P < 0.001 for comparison of this parameter between the two groups) and a reduction in calcium sensitivity for skinned fibers from endotoxin-treated animals (i.e., mean Ca₅₀ was 1.912 ± 0.05 × 10⁶ M for this group of fibers) when compared with the control group of skinned fibers (mean Ca₅₀ was 1.744 ± 0.05 × 10⁶ M for this group; P < 0.02 for comparison with the calcium sensitivity for the endotoxin group).

PEG-SOD administration to endotoxin-treated animals prevented endotoxin-induced alterations in diaphragm skinned muscle fiber force-versus-pCa curves, preventing reductions in both Fmax and calcium sensitivity. Data demonstrating this protective effect are shown in Figures 1-3. As shown in Figure 1, the force-versus-pCa relationship for skinned muscle fibers from endotoxin-treated animals was shifted downward and to the right of the force-versus-pCa curve for fibers from control animals. Fibers taken from animals given both endotoxin and PEG-SOD had a force-versus-pCa relationship similar to fibers from control animals, i.e., shifted up and to the left of the curve for fibers from animals given endotoxin alone. In addition, force-versus-pCa curves for fibers from animals given endotoxin and denatured PEG-SOD were similar to curves for fibers from animals given endotoxin alone, whereas control animals given only active PEG-SOD had force-versus-pCa curves similar to fibers from control animals given neither endotoxin nor PEG-SOD. As shown in Figure 2, Fmax for groups of skinned fibers from animals given endotoxin, with or without denatured PEG-SOD, were significantly lower than Fmax for fibers from either control animals given no agents, control animals given PEG-SOD, or animals given both endotoxin and PEG-SOD. This same pattern was seen for calcium sensitivity data (Figure 3), with an increase in Ca₅₀ (i.e., reduced calcium sensitivity) observed for groups of skinned fibers from animals given endotoxin, with or without denatured PEG-SOD, as compared with groups of fibers from either control animals given no agents, control animals given PEG-SOD, or animals given both PEG-SOD and endotoxin.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Mean force-versus-pCa relationships for single fibers from control (open circles), endotoxin (filled circles), PEG-SOD/endotoxin (open triangles), PEG-SOD alone (open squares), and denatured PEG-SOD/endotoxin (filled triangles) groups. Data points represent mean findings for all fibers within a given experimental group; error bars represent ± 1 SEM. The force-versus-pCa curves for the endotoxin and denatured PEG-SOD/endotoxin groups are statistically different from the curve for the control group.

View larger version (49K): [in this window] [in a new window]   Figure 2.   Mean Fmax (kPa) for single fibers taken (from left to right) from control, endotoxin-treated, PEG-SOD/endotoxin, L-NAME/endotoxin, PEG-SOD, L-NAME, and denatured PEG-SOD/endotoxin groups. *Value statistically different from the value for the control group.

View larger version (41K): [in this window] [in a new window]   Figure 3.   Mean Ca50 for single fibers taken (from left to right) from control, endotoxin-treated, PEG-SOD/endotoxin, L-NAME/ endotoxin, PEG-SOD, L-NAME, and denatured PEG-SOD/endotoxin groups. *Value statistically different from the value for the control group.

The distribution of fiber types examined for each experimental condition is shown in Table 1. The effects of endotoxin to reduce Fmax and PEG-SOD to prevent this endotoxin-mediated effect, as a function of fiber type, are shown in Table 2. Fmax did not vary significantly as a function of fiber type for a given experimental condition, e.g., Fmax was similar for Type IIa, IIx, IIb, and I fibers taken from the control group of animals. For all fiber types (IIa, IIx, IIb, and I), endotoxin group Fmax was lower than the Fmax for the corresponding fiber type from the control group (P < 0.05). PEG-SOD-treated animals had fiber type Fmax values similar to controls. The variance of measurements for Ca₅₀ determinations was sufficiently high that valid statistical comparisons as a function of fiber type subgroup could not be made.

Effect of a NOS Inhibitor (L-NAME) on Endotoxin-Mediated Alterations of Single Triton-Skinned Diaphragm Muscle Fiber Force-versus-pCa Curves

L-NAME also prevented endotoxin-mediated reductions in Triton-skinned diaphragm fiber force-generating capacity, as shown in Figures 2-4. Specifically, the force-versus-pCa relationship (Figure 4) for Triton-skinned single muscle fibers from animals given both endotoxin and L-NAME was similar to that observed for fibers from control animals given either no agent or L-NAME alone, and fibers from all three groups generated higher forces as a function of pCa level when compared with fibers from the endotoxin-treated group of animals. This same effectiveness of L-NAME in preserving force generation is seen when examining Fmax and Ca₅₀ data. L-NAME prevented reductions in both Fmax (Figure 2) and calcium sensitivity (i.e., prevented increases in Ca₅₀, Figure 3). As for PEG-SOD, L-NAME was effective in preventing endotoxin-mediated reductions in Fmax for all fiber types examined, as shown in Table 2.

View larger version (19K): [in this window] [in a new window]   Figure 4.   Mean force-versus-pCa relationships for single fibers from control (open circles), endotoxin (filled circles), L-NAME/ endotoxin (open triangles), and L-NAME alone (open squares). The force-versus-pCa curve for the endotoxin group is statistically different from the curve for the control group.

Effect of In Vitro Administration of Superoxide Scavenger, NOS Inhibitor, and Reducing Agents on Force Generation by Triton-Skinned Fibers from Endotoxin-Treated Animals

We were interested in determining whether the reductions in force-generating capacity of skinned fibers from endotoxin-treated animals could be reversed in vitro by incubation with various reducing agents, SOD, or L-NAME. Data for Fmax of force-versus-pCa curves before and 5 min after in vitro administration of these agents are shown in Figure 5; for comparison, we also examined the effect of these agents on fibers from control animals. We found that neither NAC (150 mg/liter), DTT (10 mM), SOD (2,000 U/liter), nor L-NAME (1 mg/liter) improved force generation by fibers from endotoxin-treated animals. These agents also had no effect on force generation by fibers from non- endotoxin treated, control animals.

View larger version (28K): [in this window] [in a new window]   Figure 5.   Examination of the effect of incubation of a series of pharmacologic agents in vitro with skinned muscle fibers from control (Cont, open bars) and endotoxin-treated (Endo, filled bars) animals. Agents examined included reducing agents (DTT and NAC), SOD, and L-NAME. The first of each pair of bars represents the Fmax of the force-versus-pCa relationship determined before administration of an agent, and the second bar in each pair represents the Fmax of a force-versus-pCa relationship determined after administration of the agent. From left to right are displayed results for in-time control fibers with no agent added, endotoxin fibers with no agent added, control fibers treated with DTT, endotoxin fibers treated with DTT, control fibers treated with NAC, endotoxin fibers treated with NAC, control fibers treated with SOD, endotoxin fibers treated with SOD, control fibers treated with L-NAME, and endotoxin fibers treated with L-NAME. None of the agents tested significantly altered Fmax for fibers from either control or endotoxin-treated animals.

SDS-PAGE Protein Patterns

SDS-PAGE gel electrophoresis with Coomassie blue protein staining of Triton-skinned muscle fibers revealed depletion of several protein bands for fibers taken from endotoxin-treated animals as compared with controls (see Figures 6 and 7, lanes B and A, respectively). There was, in particular, loss of a protein band staining at a molecular weight of approximately 30 kD, and diminution of bands staining at approximately 140 and 100 kD. In contrast, the molecular weight protein band corresponding to -actin (40 kD) stained similarly for skinned muscle fiber samples from control and endotoxin-treated animals. Coadministration of either PEG-SOD or L-NAME with endotoxin prevented alterations in these protein bands, with electrophoretic patterns for fibers taken from these groups of animals (Figures 6 and 7, lanes C and D) staining similarly to the pattern observed for fibers from control animals. Each of the previously described protein bands (i.e., those at 140, 100, and 30 kD) was severely depleted in muscle fibers taken from endotoxin-administered animals when band density was expressed as a percentage of the density of the corresponding band from control fibers. Specifically, the band at 140 kD was 58 ± 4%, the 100-kD band was 67.5 ± 5.5%, and the 30-kD band was 52 ± 6% of the corresponding protein bands from control muscle fibers. For each of the proteins, PEG-SOD administration prevented most of this depletion (i.e., the 140-kD band was 80 ± 5%, the 100-kD band was 93 ± 1%, and the 30-kD band was 96 ± 3% of control). Similarly, when L-NAME was given, the 140-kD band was 73.5 ± 0.5%, the 100-kD band was 87 ± 3%, and the 30-kD band was 86.5 ± 1.5%.

View larger version (42K): [in this window] [in a new window]   Figure 6.   Coomassie-stained SDS-PAGE protein electrophoretic patterns for skinned fibers. A portion of the gel containing high molecular weight proteins is shown. Lanes, from left to right, contain: (A) control group fibers, (B) endotoxin group fibers, (C) PEG-SOD/endotoxin fibers, (D) L-NAME/endotoxin fibers, and (E ) molecular weight markers. The arrows indicate protein bands with diminished intensity for the endotoxin-treated lane (B) as compared with the other fiber groups.

View larger version (60K): [in this window] [in a new window]   Figure 7.   Coomassie-stained SDS-PAGE protein electrophoretic patterns for skinned fibers. A portion of the gel containing lower molecular weight proteins is shown. Lanes, from left to right, contain: (A) control group fibers, (B) endotoxin group fibers, (C) PEG-SOD/endotoxin fibers, (D) L-NAME/endotoxin fibers, and (E) molecular weight markers. The arrow indicates a protein band with diminished intensity for the endotoxin-treated lane (B) as compared with the other fiber groups.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The major findings of this study were: (1) endotoxin administration to rats resulted in an alteration in the functional properties of the contractile proteins of the diaphragm, reducing force-generating capacity; (2) administration of either a free-radical scavenger, PEG-SOD, or a NOS inhibitor, L-NAME, to rats prevented the endotoxin-mediated reduction in contractile protein force generation; (3) in vitro application of a variety of reducing agents and antioxidants to fibers from endotoxin-treated animals failed to restore fiber force generation; and (4) SDS-PAGE of skinned fibers from endotoxin-treated animals revealed a selective depletion of several proteinsadministration of L-NAME or PEG-SOD to endotoxin-treated animals prevented this protein depletion, paralleling the effect of these two agents to prevent loss of contractile protein force-generating capacity.

Methodologic Considerations

When performing these experiments, single fibers were pulled from fiber bundles and then mounted on rigid supports attached to a force transducer; identification of fibers was subsequently achieved by gel electrophoresis. As a result, we were unable to prospectively control the fiber type selected for an individual study. Larger fibers tend to be somewhat easier to mount in this procedure, and as a result our fiber selection approach was biased toward preferentially sampling larger fibers (i.e., Types IIa, IIb, and IIx) and undersampling small fibers (i.e., Type I). As a consequence, the proportion of different fiber types examined in this study (5% Type I, 16% Type IIa, 32% Type IIb, and 47% Type IIx) was skewed toward larger fibers when compared with the rat diaphragm fiber type distribution (22% Type I, 28% Type IIa, 15% Type IIb, 35% Type IIx) as determined by immunohistochemistry (16).

We do not believe, however, that this factor radically alters the interpretation of our results. Endotoxin administration clearly resulted in a reduction in the Fmax of all fibers tested in this experiment, and L-NAME and PEG-SOD clearly prevented this endotoxin-related reduction for all fibers tested (see Table 2). It is possible that some minor fiber type-related susceptibility to the effects of endotoxin or to the protective effects of L-NAME or PEG-SOD may have been missed due to underrepresentation of Type I fibers in the present study. The qualitative interpretation of our findings, however, would remain exactly the same even if such a fiber type-susceptibility pattern existed. Moreover, in a recent study (7) we examined the effect of the same regimen of endotoxin administration used in the current study on rat soleus muscle (rat soleus fibers are almost exclusively Type I; the effects of scavenger administration were not assessed in this earlier study). In this earlier study, endotoxin elicited a reduction in soleus Type I fiber force generation of 36%, a value similar to reductions observed in the present study for diaphragm fibers (endotoxin elicited 30, 42, 43, and 38% reductions in force, respectively, for Type I, IIa, IIb, and IIx fibers in the current study). These data again argue that endotoxin administration leads to substantial reductions in the force-generating capacity of the contractile proteins of all skeletal muscle fiber types.

Comparison with Previous Studies

The argument that free radicals play an important role in mediating sepsis-induced skeletal muscle dysfunction is based largely on the results of a series of studies examining the effects of free-radical scavengers and NOS inhibitors on muscle function in animal models of sepsis (1). In these studies, administration of endotoxin to animals was shown to induce a downward shift of the force-frequency relationship of various limb and respiratory (both diaphragm and intercostal) skeletal muscle bundles, reducing muscle force generation for electrical stimulation frequencies ranging from 1 to 100 Hz. Concomitant administration of endotoxin and either NAC (this substance both scavenges free radicals and provides substrate for augmentation of cellular glutathione levels), SOD (a scavenger of superoxide ions), catalase (this degrades hydrogen peroxide), DMSO (a hydroxyl radical scavenger), or L-NMMA (an inhibitor of NOS) prevented muscle dysfunction in these past studies, preserving normal levels of force generation at all stimulation frequencies (1, 2, 4, 5).

These early inhibitor studies suggest that a variety of radical species are involved in the induction of muscle dysfunction during sepsis. Subsequent work has identified several biochemical pathways by which free radicals are generated in skeletal muscle during the induction of sepsis. Inducible NOS enzyme levels increase in the diaphragm during sepsis, providing a potential source of NO (4, 5). In addition, NO generation by constitutive NOS enzymes appears to be increased in a variety of muscles during sepsis, providing a second source of NO generation (17). Because NO is freely diffusible, NO levels in all muscle subcellular compartments would be expected to rise as a result of enhanced NOS activity in sepsis. Sepsis also induces an increase in the generation of superoxide-derived free-radical species (i.e., superoxide anions, hydrogen peroxide, hydroxyl radicals) by skeletal muscle, with a doubling of contraction-related free-radical generation by skeletal muscles from septic animals when compared with controls (6). Two reports suggest that the mitochondria are the source of enhanced skeletal muscle superoxide- derived free-radical generation in sepsis, and several mechanisms appear to upregulate superoxide generation by the electron transport chain (18, 19). Superoxide generated within mitochondria undergoes dismutation (i.e., by peroxidases) to form hydrogen peroxide, which can readily cross mitochondrial membranes, thereby potentially leading to increased cytosolic concentrations of this molecular species. Moreover, one report suggests superoxide anions per se can be released from mitochondria, with the implication that increased mitochondrial generation of superoxide in sepsis can lead to superoxide release into other cellular muscle compartments (20). Free radicals generated by white cells infiltrating skeletal muscle also appear to contribute to free-radical generation in sepsis, inasmuch as administration of apocynin, an inhibitor of the nicotinamide adenine dinucleotide phosphate-dependent free radical-generating complex on the surface of white cells, also blunts endotoxin-induced skeletal muscle dysfunction (3).

The present study describes a potential mechanism by which free-radical species, generated in and around skeletal muscle myocytes, alter skeletal muscle force-generating capacity during sepsis, providing a link between those studies demonstrating an enhancement of free-radical generation and those reports indicating that free-radical scavengers impact force-generating capacity. The force-generating ability of the contractile proteins is the limiting factor determining the Fmax of muscle fibers, and the large downward shift in the contractile protein force-versus-pCa relationship evoked by sepsis in the present study provides a reasonable explanation for the downward shift in the intact muscle bundle force-frequency curve evoked by sepsis. The fact that both a superoxide anion scavenger and a NOS inhibitor prevented this sepsis-induced alteration in contractile protein function indicates that one or more radical species derived from both superoxide and NO interact to reduce contractile protein force-generating capacity in this condition.

There are several conceivable mechanisms by which NO and superoxide, or the derivatives of these molecules, might interact to alter contractile protein function. For one thing, reaction of NO with either superoxide or its dismutation product, hydrogen peroxide, generates peroxynitrite, an extremely reactive, highly toxic, molecular species (21, 22). Peroxynitrite has recently been shown to alter skeletal muscle contractile protein function in vitro, markedly reducing Fmax (23). The most straightforward explanation for the current findings, therefore, is that sepsis-related increases in cytosolic levels of NO and either superoxide or hydrogen peroxide lead to peroxynitrite formation; and peroxynitrite, in turn, chemically modifies the contractile proteins, reducing-force generating capacity. Our finding that SDS-PAGE demonstrated alterations in the contractile protein bands for endotoxin-treated animals would also be consistent with such a scenario. The observed alterations in protein bands indicate either selective depletion of some proteins or radical chemical modification and altered electrophoretic protein migration. Peroxynitrite reaction with proteins is known to produce several biochemical modifications (i.e., amino acid nitrosylation, sulfhydryl oxidation) that can either alter protein mobility or predispose proteins to endogenous degradation by proteolytic systems (24). The fact that SOD and NOS inhibitor treatment prevented contractile protein band alterations after endotoxin treatment would also be consistent with this theory, inasmuch as either SOD or NOS inhibitor treatment would be expected to diminish peroxynitrite-mediated protein alterations.

Several alternative mechanisms could, however, theoretically account for the findings of the current study. For example, it is conceivable that all the endotoxin-induced contractile dysfunction observed in the present study could be entirely due to the effects of NO. If so, the effectiveness of SOD in preventing endotoxin-induced contractile protein alterations would imply that superoxide generation was a critical regulator of NO formation in skeletal muscle in sepsis (25, 26). The ability of NO to produce contractile protein alterations similar to those observed in the present study is controversial, however. One in vitro study has suggested that NO can induce large reductions in contractile protein function (27), but several others have failed to demonstrate physiologically significant effects of NO per se on contractile protein force generation (28, 29). Conversely, it is possible that all the contractile protein alterations observed in the present study could have been caused by superoxide-derived radical species (except peroxynitrite), and NO may serve simply to regulate superoxide generation after endotoxin administration. In support of such a possibility, both superoxide anions and hydroxyl radicals have been shown to directly reduce contractile protein force-generating capacity in vitro in experiments using reduced preparations (30, 31).

One might also postulate that a variety of NO- and superoxide-derived radical species may be acting independently to chemically modify the contractile proteins, with the total effects observed representing the "sum" of the effects produced by the contributing reactions. In such a case, however, one would expect that SOD or L-NAME would each have been only partially successful in preventing contractile alterations. Instead, we found that SOD prevented 83% and L-NAME prevented 88% of the sepsis-induced reductions in contractile protein function, indicating an interactive, synergistic interplay between superoxide- and NO-generating systems in inducing contractile dysfunction in the diaphragm. A final possibility is that the findings observed in the present study do not represent direct effects of free radicals on the contractile protein system. For example, free radical-mediated alterations in mitochondrial function could alter the intracellular milieu, indirectly affecting contractile protein function. Mitochondrial dysfunction, however, would be expected to primarily affect the contractile proteins by altering hydrogen and inorganic phosphate concentrations in the cytosol. In the present study, such alterations would be expected to be readily reversible in vitro, however, because skinned fibers were incubated in highly buffered, low-phosphate solutions. No such reversal of the force generation impairment was observed.

Although we think our findings point toward an effect of peroxynitrite to alter contractile protein function after endotoxin administration, additional work will be required to evaluate this possibility completely. In any case, the present findings clearly indicate that free-radical species derived from both NO and superoxide interact to induce the contractile protein alterations seen in skeletal muscle in sepsis, and inhibition of either pathway prevents these changes.

Implications

Several studies have indicated that infections can precipitate respiratory muscle dysfunction of such severity as to result in respiratory failure (1, 4, 32). In one animal study, death resulted by this mechanism if animals were not mechanically ventilated (32). In a more recent human study, minor infection was shown to induce respiratory failure (i.e., reductions in respiratory muscle strength of sufficient severity to result in hypercapnia) in patients with pre-existing respiratory impairment (33). The present work provides a potential explanation for these previous observations, demonstrating for the first time that a significant free radical-induced reduction in respiratory muscle contractile protein force-generating capacity results from endotoxin administration. In theory, therapeutic measures that prevent the development of this abnormality (e.g., administration of agents that inhibit free-radical synthesis) may be of benefit in preventing the development of respiratory muscle dysfunction and respiratory failure in selected patients with infections.