INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Severe sepsis has been reported to be increasing in incidence and entails a mortality rate of at least 20% in most studies (1). Ventilatory support is frequently required in patients with sepsis, and impaired diaphragm contractility may be an important contributing factor to the high incidence of respiratory failure encountered in this setting. In this regard, several previous studies have demonstrated a significant decrease in diaphragm force-generating capacity in different animal models of endotoxemia (2, 3). A number of mechanisms have been proposed to account for this observation, including an imbalance between energy utilization and supply (3) as well as direct cytotoxic effects of various mediators (e.g., oxygen free radicals, cytokines) released as part of the systemic inflammatory response associated with the sepsis syndrome (4, 5).

Recent studies have suggested that nitric oxide (NO) may be one of the principal mediators of diaphragm dysfunction during sepsis. Although early investigation of NO-induced diaphragm dysfunction during sepsis focused largely on its adverse effects on the regulation of diaphragmatic blood flow (6), there is increasing evidence that NO also has the ability to depress myofiber contractility directly (7, 8). Thus, Kobzik and colleagues (7) initially reported that NO chelators and blockers of NO synthase (NOS) enhanced force production by isolated rat diaphragm muscle strips in vitro, whereas the opposite effect was observed after exposure to NO donors. More recently, both Hussain and colleagues (9) and Boczkowski and coworkers (8) have documented increased expression of the inducible NOS isoform (iNOS) in the diaphragm of rats injected with lipopolysaccharide (LPS). The latter investigators also found an apparent correlation between LPS-induced diaphragm iNOS expression and reductions in force-generating capacity of diaphragm muscle strips studied in vitro (8). However, the precise mechanisms underlying NO-induced reductions in diaphragm myofiber force-generating capacity are currently unknown. Although not a strong oxidant itself, NO reacts with superoxide to form the highly reactive oxidant anion, peroxynitrite (10). Peroxynitrite formation can in turn cause cellular injury through a number of mechanisms, including lipid peroxidation, protein modification, DNA strand breakage, and inhibition of mitochondrial respiration (10). In addition, increased peroxynitrite formation contributes to the generation of another highly reactive oxidant species, hydroxyl radical (10).

Accordingly, in the present study it was hypothesized that enhanced iNOS expression and attendant increases in NO production during sepsis might trigger diaphragm myofiber injury, thereby contributing to the abnormal myofiber function observed under these conditions. Here we show that in two different animal models of sepsis, an increase in iNOS expression in the diaphragm is associated with morphologic evidence of widespread damage to the myofiber membrane, or sarcolemma. Furthermore, the loss of sarcolemmal integrity at the microscopic level is correlated with abnormal electrophysiologic properties of the membrane, which could in turn help to account for impaired myofiber contractility during sepsis. Importantly, the morphologic as well as the electrophysiologic signs of diaphragm myofiber membrane injury and dysfunction in this setting can be significantly alleviated by the initiation of NOS inhibition within a short time period after sepsis induction.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Procedures

All aspects of the study were approved by the local institutional animal care and use committee, and the animals were killed by overdose with sodium pentobarbital administered intraperitoneally at the end of the experimental protocols. Pathogen-free adult male Sprague-Dawley rats weighing 450 to 500 g were randomly assigned to one of three groups: (1) control (CTL), (2) an acute endotoxemia model induced by administration of E. coli LPS, and (3) a subacute sepsis model induced by cecal ligation and perforation (CLP). The rats were housed at room temperature with a 12:12-h light-dark cycle and provided with rat chow as well as water ad libitum. All animals were anesthetized intraperitoneally by administration of sodium pentobarbital (50 mg/kg) in preparation for surgery, which was performed under sterile conditions.

To produce the acute endotoxemia model (LPS group), the aorta was infused with LPS (20 mg/kg of serotype 055:B5; Sigma Chemical Co., St. Louis, MO). After induction of anesthesia, the carotid artery was cannulated with the catheter directed retrograde into the aortic arch, and the artery was then ligated above this level. After LPS administration, animals were maintained in the anesthetized state and continued to breath spontaneously until being killed 4 h later. Control rats for the LPS group underwent the same procedures but were injected with saline rather than LPS. In the subacute sepsis model (CLP group), a state of ongoing bacterial peritonitis (confirmed by changes in white blood cell counts and blood cultures in a subset of animals) was produced in rats according to previously described methods (14). Briefly, a midline laparotomy was performed, and the cecum was ligated distal to the ileocecal valve so that bowel continuity was preserved. The ileocecal mesentery was separated while ensuring that vessels to the cecum were ligated. Two holes were then made in the antimesenteric surface of the cecum using an 18-gauge needle. The cecum was returned to the peritoneum, and the incision was closed in two layers. Animals recovered from anesthesia and moved about freely in their cages until being killed 24 h later. Control animals for the subacute sepsis model were also subjected to laparotomy but did not undergo the CLP procedure, and the abdominal incision was closed without further manipulation.

In order to evaluate the possible role of NO in sepsis-mediated diaphragm myofiber membrane injury, a cohort of animals from both the LPS and the CLP groups also received N-monomethyl-L-arginine (LNMMA; Sigma Chemical), a potent inhibitor of NOS activity (15). This was administered by infusion into the aorta at a dose of 8 mg/kg 90 min after either the LPS injection or the CLP procedure; this dose and time point were selected on the basis of previously described salutary effects of LNMMA on rat diaphragm contractility after LPS injection (8). In addition, the specificity of LNMMA effects on diaphragm myofiber membrane properties was ascertained by treating another subset of animals from each group in the same manner with its stereoisomer, DNMMA, which has little or no inhibitory effect on NOS activity (15).

Morphologic Assessment of Sarcolemmal Injury

The level of sarcolemmal injury was evaluated, as previously described (16, 17), by perfusing muscles with a low-molecular weight (FW = 631) fluorescent tracer dye (Procion Orange; Sigma Chemical) to which the sarcolemma of normal myofibers is impermeable. The tracer dye (1% in Ringer's solution, total volume of 20 ml) was infused into the aorta immediately prior to death. After euthanasia, the diaphragm was removed en bloc and divided into its two halves. One randomly selected hemidiaphragm as well as a limb muscle (soleus) sufficiently thin to allow good dye diffusion were then immediately submerged in oxygenated 1% Procion dye/Ringer's solution for an additional 90 min at room temperature. The muscles were subsequently rinsed, snap-frozen in isopentane precooled with liquid N₂, and stored at 80° C.

Myofibers with sepsis-induced sarcolemmal damage were identified by their inability to exclude the low molecular weight tracer dye from the cytoplasm. Frozen hemidiaphragm and soleus tissues were embedded in mounting medium, and serial sections (10 µm thick) from the midbelly of the muscles were cut with a cryostat at 20° C. The sections were mounted and viewed under epifluorescence microscopy (fluorescein filter settings, magnification ×100). The images were photographed with a video camera and captured to computer with the public domain program NIH Image (version 1.49). Using computer-assisted image analysis, myofibers demonstrating clear cytoplasmic staining (i.e., fibers containing intracellular dye because of a loss of sarcolemmal integrity) were then counted by an observer blinded to the identity of the samples, and the percentage of dye-positive fibers on each muscle section was determined. Areas with sectioning artifacts (folds, tears, etc.) were avoided, and the edges of the sections were also excluded to avoid any fibers potentially damaged by the muscle dissection. A minimum of 300 myofibers from randomly selected microscopic fields were counted on each tissue section, and for each muscle at least three tissue sections were analyzed and averaged.

Detection of iNOS Expression by Western Analysis

Immunodetection of iNOS expression was performed as previously described (9). Protein samples were prepared for SDS-polyacrylamide gel electrophoresis by homogenizing tissues in a buffer containing 0.2% Nonidet P-40, 40 mM KCl, 10 mM HEPES, 3 mM MgCl₂, 1 mM DTT, 5% glycerol, 1 µg/ml aprotinin, 0.5 µg/ml leupeptin, and 0.5 mM PMSF. The homogenate was then centrifuged at 10,000 × g for 20 min and the supernatant was collected for protein quantitation using a commercial kit (Bio-Rad Protein Assay; Bio-Rad Laboratories, Hercules, CA). The samples were heated for 15 min at 90° C prior to loading (100 µg/well) on precast 4 to 12% gradient gels (Novex, Inc., San Diego, CA). Electrophoresis was performed with TRIS-glycine running buffer (pH, 8.3) in a vertical slab gel apparatus at 125 V for approximately 2 h. The separated proteins were then electrophoretically transferred onto polyvinylidene difluoride membranes (Bio-Rad Laboratories), which were subsequently blocked overnight at 4° C with 5% nonfat dry milk. The membranes were incubated with a primary monoclonal anti-iNOS antibody (1:500) raised in mouse (Transduction Laboratories, Lexington, KY), followed by incubation with a horseradish-peroxidase-conjugated antimouse secondary antibody (1:500). Specific iNOS signal on the immunoblot was detected using enhanced chemiluminescence (Amersham Canada, Oakville, ON), and lysate of cytokine-activated murine macrophages served as a positive control for iNOS expression as previously described (9).

Muscle Histochemistry

To identify macrophages, which are known to be an important potential source of increased NO production (18, 19), muscle sections were fixed in 10% buffered formalin for 10 min, rinsed, and incubated in acid phosphatase medium for 4 h at 37° C. The sections were then rinsed and treated with 1% ammonium sulphide for 30 s. Macrophages present in muscle were identified by their heavy and diffuse acid phosphatase reactivity as previously described (20).

Electrophysiologic Evaluation of Diaphragm Myofiber Membrane Properties

After the animals were killed in the acute endotoxemia model, the excised hemidiaphragm contralateral to that frozen for morphologic analysis was immediately immersed in Krebs solution (NaCl, 118 mM/L; KCl, 4.7 mM/L; MgSO₄, 1.2 mM/L; KH₂PO₄, 1.0 mM/L; glucose, 11 mM/L; NaHCO₃, 25 mM/L; CaCl₂, 2.5 mM/L) perfused with 95% O₂/5% CO₂ at room temperature (pH, 7.4). After an initial 30-min period to allow cellular recovery and thermoequilibration, the resting membrane potential (E_m) of individual diaphragm myofibers was recorded. Briefly, E_m was measured with microelectrodes (Narishige, Tokyo, Japan) prepared on a horizontal mechanical puller from borosilicate glass capillary tubes to a tip diameter of ~ 0.1 µm. Microelectrodes were filled with a solution of 3 M KCl, and the tip resistance was measured. Microelectrodes with a tip resistance of 10 to 15 M were used. After positioning its tip above the muscle with a micromanipulator, the glass microelectrode was advanced until a myofiber was penetrated, as judged by a sharp negative deflection on an oscilloscope monitor (Tektronix, Beaverton, OR) and chart recorder (Hewlett-Packard, Palo Alto, CA). The microelectrode tip was then advanced and retracted approximately 10 µm in both directions to confirm its position within the cell. The difference between the reference electrode in the medium and the steady-state recording of potential for 20 s, free of high-frequency interference, was considered to be the E_m of the fiber. The microelectrode tip was then withdrawn and moved randomly over the entire hemidiaphragm, and this process was repeated until E_m was measured in 25 to 35 individual myofibers.

Statistical Analysis

All data are reported as means ± SE and were analyzed with a statistical analysis software package (Statistix; Analytical Software, St. Paul, MN). To assess differences between experimental groups for a given muscle, one-way ANOVA was employed with post-hoc application of Fisher's least-significant-difference test. To compare values obtained for different muscles within the same animal, Student's t test for dependent samples was applied. Linear regression was performed using the method of least squares. Statistical significance was defined as p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Extent of Sepsis-induced Sarcolemmal Injury in Diaphragm and Limb Muscle

Representative micrographs illustrating the effects of sepsis on diaphragm myofiber sarcolemmal integrity in the LPS and CLP groups are shown in Figure 1. Note that although there was intense tracer dye staining of connective tissue throughout the diaphragm in nonseptic control animals, almost no dye uptake was found within individual myofibers (Figure 1a). In contrast, numerous myofibers demonstrated variable degrees of tracer dye staining within the cytoplasm in the LPS (Figure 1b) and CLP (Figure 1c) diaphragms, a finding consistent with sepsis-induced sarcolemmal damage and a consequent impairment of myofiber structural integrity.

View larger version (113K): [in this window] [in a new window]   Figure 1.   Representative micrographs illustrating the effects of sepsis on sarcolemmal integrity in the diaphragm. (a) CTL group. Very little tracer dye uptake within individal myofibers was found, although intense staining of extracellular connective tissue indicates good dye diffusion throughout the tissue. (b) LPS group. Note the presence of numerous myofibers with sarcolemmal damage, as indicated by their inability to prevent entry of the low molecular weight tracer dye. (c ) CLP group. Myofibers with sarcolemmal damage are characterized by variable degrees of intracelluar fluorescent staining, similar to the findings in the LPS group.

Group mean data for the percentage of myofibers exhibiting morphologic evidence of sarcolemmal injury in diaphragm and soleus muscles are shown in Figure 2. As can be seen, diaphragm myofibers of both LPS and CLP animals demonstrated a markedly greater level of sarcolemmal damage than did those found in nonseptic CTL rats. Although the prevalence of such damaged myofibers in the diaphragm tended to be higher in CLP than in LPS, this did not achieve statistical significance. On the other hand, the very low prevalence of soleus myofibers with sarcolemmal damage in CTL animals (< 1%) was not altered by LPS administration, whereas there was a significant increase in sarcolemmal damage within soleus muscles of the CLP group. However, the level of sepsis-induced sarcolemmal injury was significantly higher in the diaphragm than in the soleus for both the LPS and CLP models (p < 0.05).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Comparative prevalence of diaphragm and soleus myofiber membrane injury in the two sepsis models. This figure shows the percentage of myofibers in diaphragm and soleus that demonstrated an abnormal permeability to the low molecular weight tracer dye. In the diaphragm, both LPS and CLP animals exhibited a significantly greater level of membrane damage than did CTL animals. In the soleus, only the CLP animals showed significantly increased membrane damage when compared with CTL animals. Values are mean ± SE. *p < 0.05 versus CTL group;p < 0.05 versus LPS group.

Differential Expression of iNOS in Diaphragm and Limb Muscle of Septic Rats

The effects of the two sepsis models on iNOS expression in the diaphragm and soleus was determined by immunoblotting. In nonseptic CTL animals, there was no visible iNOS expression in either muscle, as has been previously reported (8, 9). In contrast, there was a detectable signal for iNOS protein consisting of a band of approximately 130 kD in the diaphragm at 4 h after acute LPS administration, whereas no such signal was found in the soleus of these same animals (see Figure 3). Furthermore, the signal for iNOS protein in the diaphragm was even stronger at 24 h after the CLP procedure, and substantial iNOS expression was also found in the soleus muscle of CLP animals (see Figure 4).

View larger version (38K): [in this window] [in a new window]   Figure 3.   Western analysis for detection of iNOS protein at 4 h after LPS administration. Macrophage lysate was used as a positive control for iNOS expression. Protein homogenates from CTL diaphragm and soleus (SOL) muscles showed no detectable iNOS expression. In contrast, diaphragm samples from the LPS group demonstrated the presence of iNOS protein, whereas no iNOS expression was found in the soleus muscle of LPS-injected animals. The data are representative of results obtained in three to four animals per group, and each lane represents separate muscle samples obtained from different animals.

View larger version (29K): [in this window] [in a new window]   Figure 4.   Western analysis for detection of iNOS protein at 24 h after CLP procedure. Macrophage lysate was used as a positive control for iNOS expression. Protein homogenates from CTL diaphragm and soleus muscles showed no detectable iNOS expression. In contrast, both diaphragm and soleus samples from the CLP group demonstrated strong signals for iNOS protein at 24 h post-CLP procedure. The data are representative of results obtained in three to four animals per group, and each lane represents separate muscle samples obtained from different animals.

To determine whether macrophages, a potential source of increased iNOS expression, were present in the diaphragms of septic animals, acid phosphatase staining was performed as previously described (20). Whereas macrophages were largely absent from the diaphragms of CTL animals, cells showing heavy acid phosphatase reactivity were seen in the two septic groups. These cells tended to be concentrated primarily in perivascular areas (see Figure 5a), particularly in the LPS group. However, macrophage infiltration was also found in the endomysial space and surrounding muscle fibers in some instances (see Figure 5b).

View larger version (112K): [in this window] [in a new window]   Figure 5.   Muscle histology showing inflammatory cell infiltration in the diaphragm of septic animals. (A) LPS group. Extensive perivascular inflammation and infiltration with macrophages, which are characterized by heavy and diffuse staining for acid phosphatase. (B) CLP group. Acid-phosphatase-positive macrophages are observed scattered in the endomysial space (small arrowheads) as well as surrounding a single myofiber (large arrowhead ) in close proximity to a blood vessel.

Effects of NOS Inhibition on Sepsis-induced Diaphragm Myofiber Membrane Injury

Given the apparent association between the presence of sarcolemmal injury and iNOS expression in the muscles of septic animals, as noted above, we hypothesized that NOS inhibition with LNMMA could potentially alleviate myofiber membrane damage in this setting. The effects of LNMMA administration on the percentage of diaphragm myofibers exhibiting abnormal permeability to the low molecular weight tracer dye after LPS injection are shown in Figure 6. There was a significant decrease in the prevalence of damaged diaphragm myofibers in the LNMMA group, whereas no significant difference was found after DNMMA administration. Furthermore, very similar results were obtained in the diaphragms of CLP rats after LNMMA and DNMMA delivery, as depicted in Figure 7. In contrast to the diaphragm, the increased prevalence of myofiber membrane damaged observed in the soleus muscle of CLP animals was not significantly reduced by LNMMA administration (7.0 ± 1.3 versus 4.3 ± 1.2%, respectively).

View larger version (16K): [in this window] [in a new window]   Figure 6.   Effects of NOS inhibition on diaphragm myofiber membrane injury in LPS-injected animals. Administration of LNMMA, an inhibitor of NOS activity, significantly reduced the percentage of myofibers with detectable sarcolemmal injury in comparison with nontreated LPS animals. On the other hand, there was no significant alteration in the extent of LPS-induced myofiber membrane damage after DNMMA administration. Values are mean ± SE. *p < 0.05 versus CTL group; p < 0.05 versus LPS group; §p < 0.05 versus LPS + LNMMA group.

View larger version (16K): [in this window] [in a new window]   Figure 7.   Effects of NOS inhibition on diaphragm myofiber membrane injury after CLP. The administration of LNMMA significantly reduced the percentage of myofibers with sarcolemmal injury in the diaphragm, whereas there was no significant alteration in the extent of sepsis-induced myofiber damage after DNMMA treatment. Values are mean ± SE. *p < 0.05 versus CTL group; p < 0.05 versus CLP group; §p < 0.05 versus CLP + LNMMA group.

In order to ascertain whether morphologic signs of diaphragm sarcolemmal injury were associated with corresponding alterations in membrane electrophysiologic properties, E_m was also measured, as presented in Table 1. As has been previously reported (2), LPS administration led to significant membrane depolarization (i.e., less negative E_m) in diaphragm myofibers. There was a significant correlation between the degree of membrane depolarization and the percentage of diaphragm myofibers exhibiting positive tracer dye staining (r = 0.68, p < 0.001). Furthermore, in keeping with the morphologic findings, treatment with LNMMA allowed significant albeit partial restoration of E_m toward normal values, whereas DNMMA administration produced no significant effects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The principal findings of the present study can be summarized as follows: (1) iNOS induction was documented in the diaphragm after acute endotoxemia (LPS) as well as subacute sepsis (CLP); (2) substantial myofiber membrane damage was found in the two sepsis models, with the magnitude of injury as well as iNOS expression being higher in the diaphragm than in the soleus for both models; (3) evidence of sarcolemmal injury by microscopy was correlated with reductions in the resting membrane potential of diaphragm myofibers; (4) the morphologic as well as the electrophysiologic signs of sepsis-induced diaphragm myofiber membrane injury were significantly alleviated by the administration of LNMMA, a potent inhibitor of NOS activity.

Impaired neuromuscular transmission related to decreased E_m has been reported to be an important cause of diaphragm dysfunction in septic animals (2). The observed changes in E_m would be expected to promote sodium channel inactivation, thereby reducing sarcolemmal excitability and action potential propagation (21). Changes in E_m at the level of T-tubules could similarly have an adverse effect on excitation-contraction coupling (21). Therefore, our data indicate at least one important mechanism that could underlie NO-induced reductions in diaphragm force-generating capacity during sepsis (8). Although there was a significant correlation between morphologic evidence of a breakdown in sarcolemmal integrity and altered E_m values, the prevalence of the latter change among diaphragm myofibers appeared to be greater. This could indicate that the inability to maintain normal E_m during sepsis, which is likely due to dysfunction of ion channels and/or energy-dependent membrane pumps (e.g., Na-K ATPase) (21), is a more widespread alteration that precedes the actual loss of physical integrity of the membrane in affected myofibers.

In the present study, two different animal models were employed to test the hypothesis that upregulation of iNOS expression would be associated with diaphragm myofiber injury during sepsis. Acute endotoxemia by LPS administration is a commonly used model in the study of sepsis, and there is credible evidence supporting a significant role for endotoxin in the pathogenesis of the sepsis syndrome in humans (1). Nonetheless, it has been argued that the CLP model, which involves a focal infection with more sustained exposure to the biologic determinants of sepsis, may be closer to the clinical situation and therefore of greater relevance (1). Furthermore, the cytokine responses and efficacy of potential therapeutic agents can be shown to differ between acute intravascular and peritonitis models of sepsis (22, 23). In our study, the magnitude of iNOS induction in the diaphragm was higher in CLP animals than in the LPS group, whereas essentially equivalent levels of diaphragm myofiber sarcolemmal injury were found in the two sepsis models. These findings could indicate a nonlinear relationship between NO production and diaphragm damage, or a difference in the rate of NO synthesis for a given level of iNOS expression in the two groups of experimental animals. With respect to the latter possibility, disparities between NOS enzymatic activity and NOS protein levels have previously been reported (9). This could be explained by differences in the relative availability of endogenous cofactors (9, 15) and inhibitors (24) of NOS. In addition, LPS administration has recently been reported to provoke increased expression not only of iNOS but also of the endothelial (ecNOS) and neuronal (nNOS) isoforms (9), suggesting that augmented NO production by the constitutive NOS isoforms (25, 26) could also contribute to the development of diaphragm myofiber injury in sepsis.

An interesting finding in the present study was the observation that both iNOS expression and morphologic evidence of sarcolemmal injury were more pronounced in the diaphragm than in the limb muscle of septic animals. In three previous studies in which LPS was delivered intraperitoneally (8, 9, 18), greater iNOS expression and/or calcium-independent NOS activity were also found in the diaphragm as compared with limb skeletal muscles. In the case of intraperitoneal delivery of LPS as well as the CLP model employed in our study, the greater expression of iNOS in the diaphragm could have been due to direct local exposure of the abdominal surface of the muscle to endotoxin. On the other hand, the fact that a similar differential response between diaphragm and limb muscle was observed in the present study after intraarterial LPS administration suggests that additional elements may also be involved. Hemorrhagic shock experiments (27) suggest that hypotension itself could have also contributed to our findings, and we have observed reductions of about 50% in mean arterial pressure after LPS administration with recovery to normal baseline values by approximately 1 h post-LPS (data not shown). However, we do not believe our data can be adequately explained on this basis alone, since effects of systemic hypotension would be expected to involve limb muscles as well as the diaphragm. In addition, Illner and Shires (28) studied septic rabbits (induced by intraarterial injection of E. coli) and reported that muscle membrane depolarization preceded the onset of hypotension in most instances, suggesting that hypotension was generally not responsible for the membrane dysfunction produced during sepsis. The idea that sepsis mediators can affect muscle membrane properties independent of hypotension is strongly supported by the work of Tracey and colleagues (29), who demonstrated direct tumor-necrosis-factor-mediated membrane depolarization of skeletal muscles in an organ bath preparation. Therefore, we speculate that the differential response of diaphragm and limb muscle observed in our study could be related to enhanced delivery of sepsis mediators to the diaphragm because of higher blood flow as well as greater mechanical activity of the diaphragm when compared with limb muscles, particularly in the anesthetized state.

The significant reduction of sepsis-induced sarcolemmal injury and altered E_m in the diaphragm by treatment with LNMMA is strong evidence that NO played an important role in producing these abnormalities. It should also be noted that LNMMA exerted significant effects despite being administered after the induction of sepsis, which increases the likelihood that a similar approach might be applicable in septic patients. In theory, NOS inhibition by LNMMA could have reduced endotoxin delivery to the muscle by decreasing diaphragm blood flow. However, in the acute LPS model, LNMMA was only administered at 90 min after intraaortic endotoxin infusion, making this explanation unlikely. In addition, in the CLP model, NOS inhibition with L-nitro-arginine methyl ester (L-NAME) actually produces a redistribution of blood flow in favor of the diaphragm (14). Therefore, we believe that LNMMA most likely acted at the level of the diaphragm upon one or more of the cell populations residing in the muscle itself. Boczkowski and colleagues (8) recently reported increased immunohistochemical detection of iNOS in inflammatory cells present within the diaphragm at < 12 h after LPS inoculation in rats, whereas positive iNOS staining was also found in diaphragm myofibers at 12 to 24 h post-LPS injection. We also found evidence for macrophage infiltration of the diaphragm in the two sepsis models, which could represent at least one potential source of NO in addition to endothelial cells (9) and myofibers themselves (8, 25, 26) as noted earlier. Additional studies will be needed to determine the precise contribution of macrophages versus other cell types to the increased iNOS expression and attendant changes in sarcolemmal properties observed with these two models.

This is the first study to provide direct evidence that NO-induced sarcolemmal damage occurs in the diaphragm after the induction of sepsis. However, the precise cellular mechanisms underlying this finding remain to be determined. Enhanced NO production could have damaged the sarcolemma by promoting lipid peroxidation as well as nitrosation of thiol-containing proteins in the membrane (13). In this regard, it is noteworthy that membrane channels and pumps involved in maintaining normal ion concentration gradients contain thiol groups and could thus serve as target sites for NO effects (13, 30). In addition, it has been reported in a number of cell types that peroxynitrite has the ability to cause DNA damage, thereby triggering a deregulated increase in activity of the energy-consuming nuclear repair enzyme poly(ADP-ribose) synthetase (PARS) that can lead to an eventual depletion of cellular energy stores (11, 12). Furthermore, NO may cause relative energetic depletion of cells by interfering with enzymes involved in the glycolytic pathway, Krebs cycle, and electron transport chain (11, 12, 31). Such energetic deficits have been reported to trigger cell death if sufficiently severe (11, 32). Activation of PARS has also been implicated in the hyperpermeability of pulmonary epithelial cells observed after peroxynitrite exposure (12), possibly by interfering with the function of energy-dependent membrane pumps involved in maintaining ionic homeostasis. In theory, membrane pump dysfunction could also permit deregulated entry of extracellular calcium into myofibers. This could then lead to activation of lipolytic and/or proteolytic systems, with consequent degradation of membrane phospholipids and proteins, respectively (33). Therefore, a number of direct and indirect NO-driven mechanisms could potentially act either in parallel or synergistically to promote diaphragm myofiber membrane damage during sepsis.

It has been proposed that plasma membrane "wounding" caused by mechanical insults is a common event in the normal lifespan of a number of cell types (34), including skeletal muscle (35). McNeil and Khakee (35) reported that immediately after a form of exercise known to produce muscle injury, approximately one fifth of rat hindlimb muscle fibers demonstrated evidence of having undergone some degree of sarcolemmal disruption, presumably because of contraction-induced mechanical stresses imposed on the membrane during exercise. However, after a 24-h rest period the majority of fibers appeared to reseal and did not develop overt necrosis (35). We have also reported the presence of sarcolemmal disruptions in a small proportion of diaphragm myofibers after inspiratory resistive loading in nonseptic animals; the lack of necrotic changes or inflammatory cell infiltration suggested that this injury was also sublethal to diaphragm myofibers in most instances (17). The above findings have at least two implications with respect to the present study. First, the extent to which diaphragm myofiber sarcolemmal damage associated with sepsis is reversible and the cellular repair mechanisms involved in preventing full-blown necrosis remain to be determined. Second, it is possible that activity-induced (caused by mechanical stress) and sepsis-induced (caused by NO and probably other mediators) injury to the sarcolemma have additive or even synergistic effects, which would help to explain the higher degree of sarcolemmal damage observed in the diaphragm when compared with the limb muscle. If such an interaction between mechanical forces and sepsis mediators does in fact exist, the early institution of ventilatory support in sepsis could conceivably limit damage to diaphragm myofibers by placing the muscle at rest and thereby reducing the level of contraction-induced mechanical stress. Future studies will be needed to more fully assess the potential utility of NOS inhibition as well as other measures such as ventilatory support in the prevention and treatment of sepsis-induced diaphragm myofiber injury.