INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Numerous investigators have documented that ventilatory and limb muscle contractile performance is attenuated in animals injected with bacterial lipopolysaccharide (LPS), live bacteria, and in animals with severe infection or peritonitis (1- 3). Although the exact mechanisms behind the decline in muscle contractility in these conditions remain under investigation, several groups have recently implicated enhanced nitric oxide (NO) production (4).

NO, a second messenger with diverse biological functions, is synthesized from L-arginine by three distinct nitric oxide synthases (NOS). The neuronal (nNOS) and endothelial (eNOS) isoforms, which were first identified in neurons and endothelial cells, respectively, are now known to have widespread tissue distribution including normal skeletal muscle fibers (7, 8). The third isoform, the inducible NOS (iNOS), is present in many cell types in response to bacterial endotoxin and inflammatory cytokines such as tumor necrosis factor, interleukins, and interferon gamma (7).

Several investigators emphasized in the past few years the role of NO in mediating vascular and contractile failure of the ventilatory muscles. Hussain (9) was the first to report that NO release is enhanced in diaphragmatic vasculature of endotoxemic dogs. Subsequent studies confirmed that NO production is increased and iNOS induction develops in diaphragm and other skeletal muscles in animals injected with Escherichia coli bacterial LPS (4). Despite the confirmation of the involvement of NO in LPS-mediated muscle contractile dysfunction (4, 10), the exact source of enhanced NO synthesis and the mechanisms through which NO inhibits muscle contractility remain unclear. Many investigators proposed iNOS as the source of enhanced NO synthesis in skeletal muscles of septic animals (4, 5). Our group (6, 10), on the other hand, documented that enhanced NO synthesis also originates from the eNOS and nNOS isoforms, which are upregulated at later stages of septic shock than iNOS. The controversy regarding the role of different NOS isoforms in sepsis-induced muscle dysfunction could be attributed to the lack of isoform-specific NOS inhibitors. Most of the available iNOS inhibitors inhibit the activity of not only other NOS isoforms but other enzymes such as those mediating prostaglandin synthesis (11).

In this study, we assessed the role of iNOS in the pathogenesis of LPS-induced diaphragmatic contractile dysfunction by using mice in which iNOS gene has been disrupted (knockout). Several groups have recently used these mice to document the role of iNOS in LPS-induced multiple organ failure and vascular dysfunction (12). Our results indicate that LPS-induced impairment of diaphragmatic force generation during high-frequency stimulations was more severe in iNOS knockout mice than in wild-type mice.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents

Materials for NOS activity and E. coli endotoxin were obtained from Sigma (St. Louis, MO). L-[2,3³H]arginine was obtained from Dupont, Inc. (Mississauga, ON, Canada). Immunoblotting apparatus and reagents were obtained from Novex Inc. (San Diego, CA). Polyclonal anti-iNOS, eNOS, and nNOS antibodies were obtained from Transduction Laboratories (Lexington, KY). Enhanced chemiluminescence (ECL) kit was obtained from Amersham Inc. (Oakville, ON, Canada).

Animal Preparation

All experiments were approved by the Animal Care Committee of McGill University. F₂ B6/129 hybrid iNOS knockout mice were generated as mentioned previously (13). Wild-type B6/129 hybrid mice were purchased from Jackson Laboratories (Bar Harbor, ME) and bred to obtain an F₂ generation to serve as experimental controls. Four groups of animals (n = 6 animals for each group) were studied. Two groups of iNOS knockout mice were injected intraperitoneally with either 25 mg/kg E. coli lipopolysaccharides (KO-LPS) (serotype 055:B5) or an equivalent volume of sodium chloride (KO-control) and killed 12 h later. Similarly, two groups (WT-LPS and WT-control) of wild-type B6/129 hybrid mice were injected with either E. coli LPS or sodium chloride and killed after 12 h. This time period was chosen because our previous results indicate that iNOS induction and peroxynitrite formation in the diaphragm and intercostal muscles peak after 12 h of LPS injection (10). The animals were killed with an overdose of sodium pentobarbital and the diaphragm was quickly excised, cleaned of connective tissue, and frozen under 80° in liquid nitrogen. LPS and saline-injected mice were matched for age, gender, and weight.

Muscle Sample Preparation

Frozen diaphragmatic samples were homogenized in 6 volumes (wt/ vol) of homogenization buffer (pH 7.4, 10 mM HEPES buffer, 0.1 mM ethylenediaminetetraacetic acid [EDTA], 1 mM dithiothreitol, 1 mg/ ml phenylmethylsulfonyl fluoride [PMSF], 0.32 mM sucrose, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml pepstatin A). The crude homogenates were centrifuged at 4° C for 15 min at 10,000 rpm and the supernatants were collected.

NOS Activity

Diaphragmatic samples (50 µl) were added to 10-ml prewarmed (37° C) tubes containing 100 µl of reaction buffer of the following composition: 50 mM KH₂PO₄, 60 mM valine, 1.5 mM NADPH, 10 mM flavin adenine dinucleotide (FAD), 1.2 mM MgCl₂, 2 mM CaCl₂, 1 mg/ml bovine serum albumin, 1 µg/ml calmodulin, 10 µM tetrahydrobiopterin, and 25 µl of 120 µM stock L-[2,3³H]arginine (150 to 200 cpm/ pM). The samples were incubated for 30 min at 37° C and the reaction was terminated by the addition of cold (4° C) stop buffer (pH 5.5, 100 mM HEPES, 12 mM EDTA). To obtain free L-[³H]citrulline for the determination of enzyme activity, 2 ml of Dowex 50w resin (8% cross-linked, Na+ form) (Sigma) were added to eliminate excess L-[2,3³H]arginine. The supernatant was assayed for L-[³H]citrulline by using liquid scintillation counting. Enzyme activity was expressed in picomoles of L-citrulline produced/min/mg total protein. Protein concentration was measured by the Bradford technique with bovine serum albumin as standard. NOS activity was also measured in the presence of 1.5 mM each of ethyleneglycol-bis-(-aminoethyl ether)-N, N'-tetraacetic acid (EGTA) and EDTA, which replaced CaCl₂ and calmodulin in the reaction buffer, and in the presence of 1 mM of N^G-nitro-L-arginine methyl ester (NOS inhibitor). Ca⁺⁺/calmodulin-dependent NOS activity was calculated as the difference between that measured in the presence of CaCl₂ and that measured in EDTA/EGTA buffer. Ca⁺⁺/calmodulin-independent NOS activity was calculated as the difference between samples assayed in the presence of EGTA/EDTA and that measured in the presence of N^G-nitro-L-arginine methyl ester.

Immunoblotting

Crude muscle homogenate proteins (80 µg) were heated for 15 min at 90° C and then loaded on gradient (4 to 12%) sodium dodecyl sulfate- tris glycine polyacrylamide gels. The proteins were transferred electrophoretically onto polyvinylidene difluoride membranes and were blocked with 5% nonfat dry milk and subsequently incubated overnight at 4° C with primary antibodies. Lysates of cytokine-activated macrophages, endothelial cells, and pituitary cells were used as positive controls for iNOS, eNOS, and nNOS proteins, respectively (provided by Transduction Laboratories Inc., Lexington, KY). Specific proteins were detected using horseradish peroxidase-conjugated anti-mouse secondary antibodies and ECL reagents (Amersham Corp., Oakville, ON, Canada). The blots were scanned with an imaging densitometer (Model GS700; Bio-Rad Inc., 12-bit precision and 42-µm resolution) and optical densities of the protein bands were quantified with SigmaGel software (Jandel Scientific, San Rafael, CA). Predetermined molecular weight standards (Novex Inc., San Diego, CA) were used as markers. The membranes were stained with the silver stain and scanned to verify that equal amounts of proteins were loaded among different lanes.

Diaphragmatic Strip Preparation

Diaphragms from the four groups of mice were surgically excised with ribs and central tendon attached and placed in an equilibrated (95% O₂-5% CO₂; pH 7.38) Krebs solution chilled at 4° C that had the following composition (in mM): 118.0 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 1 KH₂PO₄, 25 NaHCO₃, and 11.0 glucose. From the central tendon to the rib, a muscle strip (2 mm wide) was dissected free from the lateral costal portion of the diaphragm. The rib was left attached to the strip and was used to secure the diaphragm strip in the custom-built Plexiglas muscle chamber. The strip was mounted in the muscle chamber, which was placed vertically in a double jacket gut bath (Kent Scientific Instruments, Litchfield, CT). A 4.0 silk thread was used to secure the central tendon to the isometric force transducer (Kent Scientific Instruments). Muscle strips were stimulated electrically at constant currents via platinum electrodes mounted in the muscle chamber and connected to a square wave pulse stimulator (Model S48; Grass Instruments, West Warwick, RI). After an equilibration period of 30 min (temperature of 22 to 25° C), the organ bath temperature was increased to 35° C and the maximal current necessary to elicit maximal force during 120 Hz stimulation frequency (600 ms duration) was identified. Muscle length was then gradually adjusted with a micrometer to the optimal value at which maximal isometric muscle force was generated in response to supramaximal stimulation (current of 300 to 350 mA, 120 Hz frequency). Muscle contractility was evaluated by stimulating the muscle at 10, 20, 30, 50, 100, and 120 Hz while maintaining constant supramaximal current and stimulation duration (600 ms). Tetanic contractions were digitized at a frequency of 1 KHz with a personal computer and stored on the hard disk for later analysis. At the end of the experiment, the strip was blotted dry and weighed. Muscle length (cm) and weight (g) were measured and used to calculate the cross-sectional area. Isometric forces were normalized for muscle cross-sectional area estimated by using the value of 1.056 g/ cm³ for muscle density (10). Peak muscle force in N/cm² was measured for each contraction within the force frequency curve. Force generation by the diaphragm during repetitive contractions was measured by assessing the decline in peak muscle force during tetanic stimulation (100 Hz, 300 ms duration) for 3 min. The changes in the ability of the diaphragm to generate force over time were quantified in two terms, the time taken for the tension to decline to 50% of initial value (T50%) and the area under the force versus time relationship.

Data Analysis

The data on NOS activity, optical densities of protein bands, and muscle force are presented as means ± SEM. Differences in NOS activity, muscle contractility, and T50% between various groups were compared by two-way analysis of variance for repeated measures. Any differences detected were evaluated post hoc by the Student-Neuman-Keuls test. A value of p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 illustrates the changes in total and Ca⁺⁺/calmodulin-independent NOS activity of the diaphragm in the four groups of animals. LPS injection in WT mice elicited a significant rise in both total and Ca⁺⁺/calmodulin-independent NOS activity (Figure 1). Total NOS activity in the diaphragm of iNOS KO-control animals was similar to that of the wild-type mice; however, there was no detectable Ca⁺⁺/calmodulin-independent NOS activity in the iNOS KO mice. Injection of LPS in iNOS KO mice elicited a substantial rise in total NOS activity with no detectable Ca⁺⁺/calmodulin-independent NOS activity (Figure 1).

View larger version (18K): [in this window] [in a new window]   Figure 1.   Diaphragmatic NOS activity in the four groups of animals. *, **p < 0.05 and 0.01 compared with corresponding control animals. ++p < 0.01 compared with wild-type-LPS animals. Notice the significant rise in total NOS activity in iNOS KO injected with LPS.

Figure 2 illustrates a representative immunoblot of diaphragmatic protein samples probed with anti-iNOS antibody. Injection of LPS in WT mice elicited induction of iNOS protein, whereas no detectable iNOS protein was observed in the control and LPS-injected iNOS KO mice (Figure 2). Representative changes in the expression of nNOS protein in the four groups of animals are shown in Figure 3. In the WT mice, the expression of nNOS protein declined in response to LPS injection and averaged 40% of control values (mean value of three independent tissue samples). By comparison, injection of LPS in the iNOS KO mice elicited an increase in nNOS protein expression (mean value of 209% of control values, Figure 3). Very weak eNOS protein was found among the four groups of animals with no apparent changes in response to LPS injection (results are not shown).

View larger version (41K): [in this window] [in a new window]   Figure 2.   A representative immunoblot of diaphragmatic protein samples probed with anti-iNOS antibody. Notice that iNOS induction was evident after LPS injection in wild-type animals, whereas iNOS protein was absent in iNOS KO animals.

View larger version (39K): [in this window] [in a new window]   Figure 3.   Changes in nNOS protein expression in the four groups of animals. Diaphragmatic protein samples were probed with anti-nNOS antibody. Note the decline in nNOS expression in response to LPS injection in the wild-type animals, whereas a significant upregulation of nNOS protein was evident after LPS injection in iNOS knockout animals.

Figure 4 shows the changes in diaphragmatic force-frequency relationship in the four groups of animals. Diaphragmatic force-frequency relationship was similar among control WT and iNOS KO animals (Figure 4). Diaphragmatic force tended to be lower in WT animals injected with LPS compared with control WT animals with statistical significance reached at frequencies of 50 Hz and above. A similar trend was observed after LPS injection in iNOS KO animals; however, at 100 and 120 Hz of stimulation, diaphragmatic force in LPS-injected iNOS KO was even lower than that observed in the LPS-injected WT animals (p < 0.05, Figure 4). The maximal relaxation rate (MRR) of the tetani produced at the highest stimulation rate (120 Hz) for both WT and iNOS KO mice was similar under control conditions (556.3 ± 107.9 and 478.8 ± 90.6 N/cm²/s, respectively). The injection of LPS did not affect the MRR in WT (505.0 ± 67.6 N/cm²/s), whereas in iNOS KO mice the MRR was significantly reduced (325.7 ± 29.18 N/ cm²/s, p < 0.05).

View larger version (20K): [in this window] [in a new window]   Figure 4.   Alterations in diaphragmatic force-frequency relationships in response to saline or LPS injection in the four groups of animals. *, **p < 0.05 and 0.01 compared with corresponding control groups. +p < 0.05 compared with wild-type-LPS group. Note the significant decline in diaphragmatic force in the iNOS KO-LPS group compared with that in the WT-LPS group.

The changes in diaphragmatic force production over time during repetitive stimulations are shown in Figure 5. The rates of force loss during repetitive tetanic stimulations in the control WT and control iNOS KO were similar with T50% averaging 57 ± 1.2 s and 58 ± 1.3 s, respectively. Similarly, the area under the force versus time relationship in control iNOS KO group averaged about 95% of that measured in the control WT group. By comparison, T50% of LPS-injected iNOS KO animals was significantly longer (81 ± 1.3, p < 0.05) than that of WT animals injected with LPS (74 ± 1.5 s). However, at any given time during the stimulation protocol, the absolute force generated by the LPS-iNOS KO group was lower than that measured in the LPS-WT animals. Moreover, the area under the force versus time relationship in the LPS-iNOS KO group averaged about 77% of that measured in the LPS-WT group (p < 0.01).

View larger version (19K): [in this window] [in a new window]   Figure 5.   Endurance of diaphragmatic muscle strips during repetitive stimulations. Diaphragmatic strips were stimulated at 100 Hz, 300 ms duration for 3 min. Note the differences in initial force between control animals and those injected with LPS.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings of this study are: (1) LPS injection elicted a mild rise in diaphragmatic NOS activity in WT animals which was associated with iNOS induction and a reduction in nNOS expression; (2) in the absence of iNOS induction in iNOS KO animals, LPS injection elicited an upregulation of diaphragmatic nNOS expression and substantial increase in muscle NOS activity; (3) LPS-induced impairment of diaphragmatic contractility was more severe in iNOS KO mice than in WT mice.

Sources of NO Synthesis in Skeletal Muscles

Under normal conditions, NO synthesis inside skeletal muscle fibers is accomplished by the activity of both the nNOS isoform, which is localized in close proximity to the sarcolemma of type II fibers, and the eNOS isoform, which is diffusely localized in type I muscle fibers (8, 15). Unlike these constitutive NOS isoforms, the iNOS isoform is not detectable in normal skeletal muscles of most species. Several investigators, on the other hand, confirmed that in vivo injection of LPS leads to the induction of iNOS in different tissues including skeletal muscles (4, 10). The observation that iNOS is expressed in the diaphragm of LPS-injected WT mice (Figure 2) is, therefore, in agreement with previous results.

In addition to the increase in Ca⁺⁺/calmodulin-independent NOS activity, our results indicate that LPS injection in WT mice was also associated with a rise in diaphragmatic Ca⁺⁺/calmodulin-dependent NOS activity (Figure 1). This observation is in accordance with our previous findings in endotoxemic rats (6, 10). Interestingly, this increase in NOS activity occurred despite a decline in nNOS protein expression (Figure 3). The exact mechanism responsible for the rise in Ca⁺⁺/ calmodulin-dependent NOS activity after LPS injection in WT animals remains unknown. We speculate that muscle constitutive NOS activity might have been enhanced by LPS exposure in a fashion similar to that observed in the eNOS isoform. Huang and coworkers (16) showed that eNOS activity is augmented after LPS exposure due to the activation of tyrosine kinases. Another possible cause of increased constitutive NOS activity in LPS-WT animals is increased availability of cofactors such as tetrahydrobiopterin (BH₄). Rosenkranz-Weiss and coworkers (17) reported that in cultured human endothelial cells, proinflammatory cytokines reduced eNOS protein expression but increased eNOS activity as a result of augmented BH₄ synthesis. It is likely that a similar induction of local BH₄ synthesis might have occurred in the diaphragm and other muscles in LPS-WT animals. This notion is supported by our recent observations in rats in which messenger RNA (mRNA) of guanosine triphosphate (GTP) cyclohydrolase I, the rate-limiting enzyme of BH₄ synthesis, was induced in limb and ventilatory muscles in LPS-injected rats (6).

Regulation of nNOS Expression

An interesting observation in our study is that LPS injection in WT animals elicited a significant decline in nNOS expression, whereas the expression and the activity of nNOS rose substantially in response to LPS injection in iNOS KO animals. These results suggest that transcriptional and post-transcriptional regulation of nNOS in response to LPS injection may be different in these two groups of mice. Little is known about the influence of LPS and/or cytokines on the expression of nNOS. Two reports described a significant upregulation of nNOS expression in skeletal muscles and the hypothalamus after LPS injection in rats (10, 18). One possible factor involved in the differential regulation of nNOS expression in WT and iNOS KO mice is that LPS may elicit different immune and cytokine responses in these two genotypes which, in turn, may influence nNOS expression differently. However, this is unlikely because LPS injection in iNOS KO and WT mice elicited similar changes in plasma levels of proinflammatory cytokines (14). Alternatively, it is possible that the presence of iNOS protein may negatively modulate nNOS expression and activity. This notion is supported by recent evidence of an interaction between the nNOS and iNOS isoforms at the transcriptional and post-transcriptional levels (19, 20). Despite these findings, the question of whether iNOS regulates nNOS transcription, mRNA stability, and protein levels in skeletal muscle fibers remains to be explored.

Role of iNOS in LPS-mediated Contractile Dysfunction

Several researchers have recently proposed iNOS as an important factor in LPS-induced skeletal muscle contractile dysfunction (4, 5, 10, 21). This proposal is based on the observations that inhibition of endogenous NO synthesis in vitro or in vivo leads to restoration of muscle force. However, because nonisoform selective NOS inhibitors were employed in all of the above-mentioned studies, one cannot easily identify the separate role of iNOS from the activity of the other NOS isoforms already present in skeletal muscle fibers.

In an attempt to define the role of iNOS in LPS-induced vascular failure and organ dysfunction, three groups of investigators developed iNOS knockout mice and studied the sensitivity of these mice to LPS-induced death (13, 14, 22). Whereas MacMicking and colleagues (14) and Wei and colleagues (22) reported that iNOS KO mice are more resistant to the influence of LPS than WT mice, Laubach and coworkers (13) found no significant differences in LPS-induced mortality between iNOS KO and WT mice. Our results indicate that LPS-induced impairment of diaphragmatic contractility was more severe in iNOS KO mice than in WT mice. These findings suggest that iNOS may play an important and a protective role against the deleterious influence of LPS. One likely pathway through which iNOS may protect muscle fibers is by attenuating leukocyte infiltration. In endothelial cells, NO release inhibits leukocyte adhesion and rolling, thereby attenuating leukocyte recruitment and minimizing tissue damage (23). Leukocyte infiltration and the release of reactive oxygen species (ROS) by leukocytes have been shown to play an important role in LPS-induced muscle contractile dysfunction (24). The protective role of iNOS in minimizing endothelial-leukocyte adhesion has recently been confirmed by Hickey and coworkers (25) who reported that LPS-induced leukocyte adhesion and rolling in cremaster muscles of iNOS KO mice were significantly higher than that of WT mice. Another protective role played by iNOS is scavenging of ROS. Indeed, inhibition of NO synthesis in endotoxemic animals has recently been shown to cause enhanced ROS-mediated hepatic injury (26). These results suggest that induction of iNOS in the endothelial cells of the diaphragm may protect the diaphragm from the deleterious influence of leukocyte infiltration. Finally, iNOS may be beneficial for muscle function through its role in upregulating the expression of the glucose transporter Glut 1 and augmenting glucose uptake (27). Thus, iNOS induction in response to LPS injection may help in overcoming the impairment of insulin-induced glucose uptake in skeletal muscle fibers.

Our study indicates that the impairment of diaphragmatic contractility in WT and iNOS KO mice after LPS injection was evident at high stimulation frequencies (50 Hz and above; Figure 4). Force generation in response to 100 Hz repetitive stimulations was also reduced in both LPS-WT and LPS-iNOS KO mice compared with their respective control groups (Figure 5). These observations are similar to those described in the diaphragm of endotoxemic rats (28). The reduction in high-frequency force after LPS injection has been attributed in part to neuromuscular transmission failure (28). In addition, LPS injection leads to the impairment of excitation along skeletal muscle membrane as a result of abnormalities in membrane ion transport and resting membrane potential (29). Other defects that are elicited by LPS injection and are likely to result in poor muscle contractile performance include inhibition of excitation-contraction coupling, contractile proteins, and mitochondrial respiration (30). The current study indicates that iNOS protein is not likely to be an important mediator of LPS-induced diaphragmatic hypocontractility because the capacity of the diaphragm to generate force at rest and during repetitive contractions was significantly higher in LPS-WT mice than in LPS-iNOS KO mice (Figures 4 and 5). However, our study does not rule out the possibility that enhanced NO release by the constitutive NOS isoforms participates in the reduction of diaphragmatic contractility after LPS injection. This is because the inhibitory effects of NO on muscle function are similar to those reported in LPS-injected animals. These effects include the attenuation of neuromuscular transmission, excitation-contraction coupling, and myofibrillar Ca⁺⁺ sensitivity (8, 15, 31). In addition, NO inhibits the activity of several important enzymes such as sarcoplasmic reticulum Ca⁺⁺-ATPase, creatine kinase, actomyosin ATPase, and cytochrome oxidase (8, 15, 32). Clearly more studies are needed to elucidate whether enhanced NO release by the constitutive NOS isoforms plays a significant role in LPS-induced muscle dysfunction.

In summary, our results indicate that LPS exerts different influences on diaphragmatic NOS expression and contractile performance in WT and iNOS KO mice. Diaphragmatic NOS activity was higher but maximal force was significantly lower after LPS injection in iNOS KO mice compared with wild-type animals. These results suggest that iNOS may play a protective role in ameliorating the negative influence of LPS on ventilatory muscle contractile function.