INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis is a common cause of morbidity and mortality, particularly in elderly, immunocompromised, and critically ill patients. Indeed, severe sepsis is at present the most common cause of death in intensive care units in the United States (1).

Respiratory failure is a major clinical manifestation of sepsis, greatly contributing to the mortality of this pathologic condition (2). In this context, respiratory failure has been traditionally related to the development of adult respiratory distress syndrome (2). However, different experimental studies demonstrated that an impaired contractile function of respiratory muscles, including the diaphragm, is another important mechanism of respiratory failure during sepsis (3). Diaphragmatic dysfunction during sepsis appears to be mediated mainly by oxidative stress (5, 8, 9). In fact, we and others have shown an increased production of oxidants, such as superoxide anion and peroxynitrite, in the diaphragm of endotoxemic rats (7, 10). The expression of the inducible isoform of nitric oxide (NO) synthase (iNOS) is partly responsible for this effect (10). Despite experimental evidence showing generation of oxidants during sepsis, little is known about the specific antioxidant systems utilized by the diaphragm to counteract this pathological condition. This is an important point to consider because a better understanding of diaphragmatic antioxidant defenses engaged against sepsis could improve our knowledge of the mechanism of oxidative stress-mediated contractile failure.

The microsomal enzyme heme oxygenase (HO) catalyzes the oxidation of heme to biliverdin and carbon monoxide (CO) and is widely distributed in mammalian tissues (11). Two main isoforms, products of different genes, have been identified: heme oxygenase 1 (HO-1), the inducible form (also known as heat shock protein 32), and HO-2, the constitutive form (12). HO-1 expression is extremely sensitive to a variety of agents that cause oxidative stress (see Reference 13 for review), and can be induced in various tissues including skeletal muscle (14, 15). In this particular tissue, expression of the HO-1 gene is significantly increased after strenous exercise (14), a condition that can lead to muscular oxidative stress (16). Increasing experimental evidence suggests that, in sepsis and other pathological states, the HO pathway is a powerful protective cellular system. Otterbein and coworkers (17, 18) have previously shown that induction of HO-1 by hemoglobin dramatically decreases mortality and improves biological and hemodynamic parameters of rats challenged with Escherichia coli endotoxin. This protective effect is likely related to HO-1 antioxidant properties (11). Indeed, elimination of the prooxidant free heme and production of biliverdin and bilirubin, two potent free radical scavengers with antioxidant characteristics (19), can account for the defensive role of HO against oxidative stress. Moreover, HO-1 could also be implicated in cellular protection against NO and NO-derived species (13, 20). Whether the HO pathway is part of a protective mechanism against diaphragmatic oxidative stress and contractile failure during sepsis has yet to be examined. Such information could provide new insight into the physiopathology of septic diaphragmatic dysfunction.

Therefore, the present study was designed to investigate (1) HO-1 and HO-2 protein expression and cellular localization, and HO activity in the diaphragm of rats treated with E. coli endotoxin (lipopolysaccharide, LPS), and (2) whether modulation of HO activity and/or protein expression could influence oxidative stress-mediated damage and contractile dysfunction in the diaphragm. In addition, to investigate whether the diaphragm was the only muscle affected by LPS treatment, we analyzed HO-1 protein expression in the abdominal muscles, the peripheral skeletal muscle soleus, extensor digitorum longus (EDL), and the left ventricular myocardium.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Studies

Male Sprague-Dawley rats (200-220 g) were purchased from Janvier France (Le Genest St. Isle, France). No more than five animals were housed in the same cage. They were acclimatized to a 12-h light-dark cycle and allowed free access to food (Ralston Purina, St. Louis, MO) and water for a 5-d period prior to the experimental procedure. During the experiments, environmental temperature was maintained between 22 and 24° C. The experiments conducted in the present study were approved by the local Institutional Animal Care and Use Committee and the experimental protocol was in agreement with French legal recommendations related to animal studies (Ministère des Affaires Sociales et de la Solidarité Nationale, Paris, France).

Animals received randomly either sterile 0.9% sodium chloride solution (control animals, n = 60) or E. coli endotoxin (LPS animals, n = 210) suspension (4 mg/kg body weight), both given intraperitoneally in a total injected volume of 0.5 ml. Endotoxin was reconstituted with 0.9% sterile sodium chloride.

Both control and LPS animals were randomly divided into two groups. One group received an inhibitor of heme oxygenase activity, zinc protoporphyrin IX (ZnPP-IX) (21, 22), given intraperitoneally at a dose of 50 µmol/kg (in 50 mM sodium bicarbonate) and the other received sodium bicarbonate 1 h before LPS or saline administration. Control animals were killed 24 h after saline inoculation. LPS animals were killed 3 h, 6 h, 12 h, 24 h, 48 h, 72 h, 96 h, and 6 d after LPS injection. Endotoxemic animals receiving ZnPP-IX were killed 24 or 48 h after LPS treatment. Rats that had to be killed at 48 h received a second injection of ZnPP-IX 25 h after the first, in order to maintain HO inhibition for a total period of 2 d.

In another set of experiments, HO-1 was induced by hemin injection (50 mg/kg, intraperitoneal) 24 h prior to administration of saline, LPS, or LPS plus ZnPP-IX. Animals were then killed 6 and 24 h after treatment with saline, LPS, or LPS plus ZnPP-IX. Care was taken to shield porphyrins from visible light to prevent its inactivation (23).

Each experimental group was composed of 6-12 animals.

Rats were anesthetized with sodium thiopental (50 mg/kg, intraperitoneal) and killed by transection of the abdominal aorta. The diaphragm, with its costal insertions and central tendon, was then quickly removed and transferred to a dissecting dish containing Krebs solution (137 mM NaCl, 4 mM KCl, 1 mM MgCl₂, 1 mM KH₂PO₄, 12 mM NaHCO₃, 2 mM CaCl₂, and 6.5 mM glucose) bubbled with a mixture of 95% O₂ and 5% CO₂. The diaphragm was pinned to maintain resting length during dissection. Temperature and pH were continuously maintained at 37° C and 7.4, respectively. Muscular samples from the lateral costal region of each hemidiaphragm were dissected. This region was chosen because it was shown to contain parallel fiber layers and to be composed of equally distributed fiber types (24). One sample was immediately prepared for contractility studies, another sample was mounted for immunohistochemistry examination (these examinations were performed in control [C] and LPS animals), and the remaining tissues were frozen in liquid nitrogen and stored at 80° C until biochemical assays were performed. For the HO activity assay, a whole diaphragm was used to perform each measurement.

Samples from the peripheral skeletal muscles soleus and EDL, from the abdominal muscle rectus abdominis, and from the left ventricular myocardium were collected from control and LPS-24 animals and stored as described above.

Western Blot Analysis

Samples from diaphragm and other skeletal muscles were homogenized in 0.35 ml of lysis buffer (50 mM Tris-HCl [pH 7.4], 0.1 mM EGTA, 1 µM EDTA, 1 µM leupeptin, 1 µM aprotinin, 1 µM phenylmethylsulfonyl fluoride [PMSF]) with an Ultraturrax T25 (Janke and Kunkel, IKA Works, Cincinnati, OH). Samples were then centrifuged at 3,000 × g for 15 min and supernatants were denatured by boiling in sample buffer for 5 min. Equal amounts of protein (100 µg/well) from each sample were electrophoresed on a 12% sodium dodecyl sulfate (SDS)-polyacrylamide gel using a Mini Protean II system (Bio-Rad, Hercules, CA) and transferred onto polyvinylidene difluoride (PVDF) membranes overnight at 4° C. After blocking in 5% nonfat milk and Tris-buffered saline (TBS)-0.05% Tween (TTBS), membranes were incubated for 1 h at room temperature with HO-1 monoclonal or HO-2 polyclonal antibodies (StressGen Biotechnologies, Victoria, BC, Canada) used at 1:1,000 dilution in TTBS-1% bovine serum albumin (BSA). Detection was performed with a chemiluminescence substrate. Positive controls for HO-1 and HO-2 proteins were obtained from rat spleen and testis homogenates, respectively. Furthermore, we evaluated the specificity of the antibodies by loading recombinant HO-1 protein side by side with HO-2 protein (both from StressGen Biotechnologies; 200 and 100 ng per lane, respectively) and using either anti-HO-1 or anti-HO-2 antibodies. After detection, membranes were stained with Red Ponceau solution and the total amount of proteins was assessed by densitometric analysis.

Immunohistochemistry

Muscular samples obtained from the costal region of the diaphragm were fixed in formol (10%), and embedded in paraffin. Four-micron sections were mounted on poly-L-lysine-coated microscope slides. After paraffin removal and hydration, the sections were rinsed in phosphate-buffered saline (PBS) and incubated in blocking buffer (PBS containing 1% bovine serum albumin and 2% normal goat serum) for 1 h at room temperature. They were then pretreated by microwave heating for antigen retrieval. After this, the slides were incubated with the same monoclonal anti-HO-1 antibody utilized for the Western blot (StressGen Biotechnologies). Incubation was effectuated during 1 h at room temperature and the concentration of the antibody was 1.2 µg/ml. The slides were washed in PBS and successively exposed to a biotinylated goat antiserum to mouse immunoglobulin G (Dako, Carpinteria, CA) diluted 1:500, and a complex of streptavidin-biotin- alkaline phosphatase (Dako). Localization of phosphatase alkaline was revealed by using Fast Red substrate solution (Dako). Three experiments were performed to assess the specificity of the immunostaining: first, preincubation of the primary antibody with recombinant HO-1 protein in a 1:10 molar ratio; second, by replacement of the primary antibody with a control isotype antibody, and then by omitting the primary antibody.

HO Activity Assay

Diaphragm microsomes for heme oxygenase activity assay were prepared as previosuly described (25). Briefly, fresh muscles were homogenized on ice in 5 volumes of 0.25 M sucrose-0.1 M Tris-HCl buffer containing a protease inhibitor cocktail (Complete; Roche Diagnostic, Meylan, France). Samples were centrifuged at 27,000 × g for 10 min at 4° C, and the supernatant was removed and centrifuged for an additional 90 min at 105,000 × g. The microsomal pellet was resuspended in 250 µl of PBS-2 mM MgCl₂ buffer and stored at 80° C until the activity assay was performed. Microsomes were incubated at 37° C in the dark for 30 min in a reaction mixture containing hemin, NADPH, glucose 6-phosphate, glucose-6-phosphate dehydrogenase, and liver cytosol (25). Bilirubin production was measured spectrophotometrically and expressed as picomoles of bilirubin per milligram of protein per hour ( = 40 mM¹ cm¹).

NOS Activity Assay

To verify that the inhibitory effect of ZnPP-IX was specific for HO in our model, we measured diaphragmatic cNOS (constitutively expressed NOS) and iNOS activity by the conversion of L-[³H]arginine to L-[³H]citrulline, according to the method of Bredt and Snyder (26) as previously described (10, 27).

Contractile Function

Diaphragmatic strips were prepared as previously described (6). Briefly, a bundle from the middle part of the lateral costal region of the diaphragm (2 mm wide) was dissected, and the tendinous end was left intact while the costal end was cut off the ribs and ligated with fine copper wire (20-µm diameter). Strips were mounted horizontally into an open-topped Plexiglas tissue chamber continually perfused with flowing Krebs solution, oxygenated with a 95% O₂-5% CO₂ mixture and maintained at 37°C and pH 7.4. Diaphragmatic strips were immobilized by snaring the tendon with surgical silk into one end of the bath, and the other extremity was attached to an FT 03D force transducer (Grass Medical Instrument, Quincy, MA). Strips were electrically stimulated with supramaximal currents (1.2 times the current required to produce a maximal tension) via an S 48 stimulator (Grass Medical Instrument) and by means of rectangular platinum field electrods. Isometric force was measured with a force transducer mounted on a micrometer, which allowed adjustment of the preparation to the optimal length (length for which twitch tension development was maximal). Peak twitch force was obtained from a series of contractions induced by a single-pulse stimulus. The force-frequency relationship was assessed by sequential stimulation of strips at 10, 20, 30, 50, 100, and 120 Hz. Each stimulus was applied for 500 ms, with at least 30 s between each stimulus train. Force was expressed in grams by gram of muscle weight.

In some experiments, diaphragms from control and LPS-24 animals were stabilized for 30 min and then superfused with 10⁴ M ZnPP-IX for 30 min. After this, force was measured and then bundles were rinsed with fresh Krebs buffer. Force was measured 10 and 30 min thereafter.

Determination of Diaphragmatic Malondialdehyde Content

Samples were homogenized on ice with cold PBS without Ca²⁺ and Mg²⁺, and centrifuged for 15 min at 3,500 × g. Malondialdehyde (MDA) concentration was determined in the supernatant, using a colorimetric assay previously reported (8). The assay is a modification of the thiobarbituric acid (TBA) method described by Yagi (28). The assay was performed in duplicate for each sample. Briefly, 1 ml of tissue sample was added to 2 ml of 0.6% TBA and 1 ml of 10% trichloracetic acid, and heated at 80° C for 20 min. Samples were then immediately water cooled and centrifuged. Optical density of the supernatant was measured at 635 nm and compared with a standard curve of tetramethoxypropane. MDA concentration was expressed as nanomoles per milligram of protein.

Oxidized Protein Detection

An OxyBlot oxidized protein detection kit (Appligen Oncor, Illkirch, France) was used to detect carbonyl groups formed as a result of protein side-chain oxidation, as described previously (29). Briefly, proteins were derivatized to 2,4-dinitrophenylhydrazone (DNP) by reaction with 2,4-dinitrophenylhydrazine. Derivatized samples were loaded on 12% SDS-polyacrylamide gels (20 × 20 cm) and electrophoresed for 16 h. Proteins were then transferred on nitrocellulose membranes and bands were revealed with a rabbit anti-DNP antibody (1:300 dilution) and a secondary antibody coupled with alkaline phosphatase according to the manufacturer instructions. Nonderivatized samples were used as negative control. Carbonyl content was assessed by densitometric analysis of the bands, using a Perfect Image 2.01 image analysis system (Iconix, Courtaboeuf, France).

Materials

Hemin chloride and ZnPP-IX were purchased from Porphyrin Products (Logan, UT). Escherichia coli lipopolysaccharide was from Difco (Detroit, MI). Sodium thiopental (Nesdonal) was purchased from Specia-Rhone-Poulenc (Romainville, France). L-[³H]arginine was from NEN/DuPont (Boston, MA). Tetrahydrobiopterin was supplied by Research Biomedical (Natick, MA). Reagents for SDS-PAGE and chemiluminescence substrates were from Bio-Rad. All other chemicals were purchased from Sigma-Aldrich (St. Quentin-Fallavier, France).

Statistical Analysis

Values are given as means ± SEM. The data obtained for the different groups were analyzed by one-way analysis of variance (ANOVA). When the F value was significant (p < 0.05), we performed a post hoc analysis with the Duncan new multiple range test in order to test the differences between means (30). Significance for all statistics was accepted at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of Endotoxin on HO-2 and HO-1 Protein Expression in Rat Diaphragm

Western blot analysis revealed a basal HO-1 protein expression in the diaphragm of control rats; the molecular weight of diaphragmatic HO-1 was identical to that of HO-1 expressed in rat spleen (Figure 1a). Expression of HO-2 protein was also detected in the diaphragm of control rats and the band was identical to HO-2 expressed in rat testis (Figure 1b). When rats were challenged with LPS, the expression of HO-2 did not change at the various time points examined; in contrast, diaphragmatic HO-1 protein levels were significantly augmented. Specifically, the intensity of the HO-1 signal markedly increased at 24 h and remained elevated up to 96 h after LPS treatment. HO-1 protein returned to basal levels 6 d after endotoxin administration. No cross-reactivity was observed between HO-1 and HO-2 antibodies (Figures 1a and 1b). Analysis of Red Ponceau staining revealed no difference in the amount of loaded proteins between the different experiments (data not shown).

View larger version (26K): [in this window] [in a new window]   Figure 1.   Time course of HO-1 (a) and HO-2 (b) protein expression, analyzed by Western blot, in total homogenates of rat diaphragm, at different times after LPS inoculation (from 12 to 144 h) or 24 h after ZnPP-IX administration in control rats. Positive control is a spleen homogenate for HO-1 and testis homogenate for HO-2. Recombinant HO-1 and HO-2 proteins were used to assess the specificity of the antibodies. The 32 and 36-kDa molecular size markers corresponding, respectively, to HO-1 and HO-2 are shown on the left. No cross-reactivity was observed between HO-1 and HO-2 antibodies. As stated by the manufacturer, a second band (molecular size 32 kDa) is present in HO-2 protein (b). Analysis of red Ponceau staining revealed no difference in the amount of loaded proteins between the different experiments (data not shown).

We further investigated the distribution of HO-1 protein in diaphragmatic tissue of endotoxemic rats. Immunohistochemical analysis showed that HO-1 was localized exclusively in diaphragmatic myocytes. At 24 h after LPS treatment, HO-1 staining was consistently and reproducibly observed in diaphragmatic myocytes in a diffuse cytoplasmic pattern and occupied a variable portion of the myocyte surface (Figure 2b). A weak staining was also detected in control animals (Figure 2a). Different technical conditions showed the specificity of the HO-1 antibody : (1) no immunostaining was observed after preincubation of the primary antibody with recombinant HO-1 protein (Figure 2d); (2) no tissue section showed positive immunostaining when the anti-HO-1 antibody was replaced by a control isotype antibody (Figure 2e); and (3) no HO-1 staining was observed when the anti-HO-1 antibody was omitted (data not shown).

View larger version (103K): [in this window] [in a new window]   Figure 2.   Immunohistochemical detection of HO-1 in the diaphragm. A slight immunostaining was observed in some myocytes of the control animal (a). Twenty-four hours after LPS inoculation, an increased immunostaining was observed in the myocytes (b). Twenty-four hours after hemin administration in control rats, an intense HO-1 immunostaining was observed in diaphragmatic myocytes (c). No immunostaining was observed in samples from an hemin-treated animal when the anti-HO-1 antibody was preincubated with recombinant HO-1 protein or replaced by a control isotype antibody (d and e, respectively).

Effect of Endotoxin on Diaphragmatic HO Activity

Having verified the presence of HO-1 and HO-2 proteins in rat diaphragm and having observed that endotoxin increases HO-1 levels, we wanted to examine whether changes in HO activity occur after LPS administration. Interestingly, HO activity in endotoxemic rats exhibited a biphasic response over time (Figure 3). In fact, HO activity transiently decreased during the first 12 h after LPS treatment and this effect was particularly evident at 6 h (33% of control, p < 0.05); however, the enzymatic activity was augmented by 22% (p < 0.05) and 17% over control values at 24 and 48 h, respectively. The increase observed in the late time points directly reflected HO-1 upregulation, as HO-2 protein levels remained unchanged.

View larger version (19K): [in this window] [in a new window]   Figure 3.   Time course of HO activity in the microsomal fraction of the diaphragm (expressed as picomoles of bilirubin produced per milligram of protein during 30 min of incubation) after LPS or LPS plus ZnPP-IX administration. Each column and error bar represent the mean ± SEM. *Different from control; # different from LPS-6; &different from LPS-12 (p < 0.05 for each paired comparison).

Effect of Endotoxin on Diaphragmatic Force

Diaphragmatic bundle dimensions did not differ among the various groups of animals considered and bundle lengths were between 14.7 and 16.1 mm.

Diaphragmatic force was significantly reduced in endotoxemic animals. Indeed, 12 and 24 h after LPS treatment, diaphragmatic peak twitch and maximal tetanic tension (tension observed at 100-Hz stimulation) were significantly lower than in the control group (p < 0.0001; Table 1). In addition, analysis of the diaphragmatic force-frequency relationship at these time points showed that force generated in response to all frequencies of stimulation was reduced in LPS compared with control animals (data not shown). A complete recovery of diaphragmatic force was observed 48 h after LPS challenge (Table 1).

Malondialdehyde and Oxidized Protein Content in Diaphragmatic Tissue of Endotoxemic Rats

LPS administration also produced changes in diaphragmatic MDA and oxidized protein contents. Specifically, MDA was significantly elevated in LPS-treated animals at 6, 12, and 24 h (p < 0.05 versus control animals) and returned to basal values at 48 h (Table 2). Western blot analysis performed on diaphragmatic samples from endotoxemic rats at 24 h showed a significant increase in oxidized proteins compared with controls (p < 0.05; Figure 4a). Several positive bands were detected, with molecular masses of 29, 39, 43, and 52 kD, respectively. These bands were present with a low intensity in control samples and no new bands were noted in septic animals, just an increased intensity of preexistent bands. The addition of the optical densities of all positive bands was significantly higher in LPS-24 as compared with control samples (Figure 4b). Immunoblotting samples not treated with DNP did not shown any positive band (data not shown). Collectively, these results indicate the occurrence of oxidative stress in rat diaphragm after LPS treatment.

View larger version (24K): [in this window] [in a new window]   Figure 4.   Effect of endotoxemia and HO inhibition on diaphragmatic carbonyl protein content in the different groups of animals. Carbonyl groups were derivatized to 2,4-dinitrophenylhydrazone and detected by Western blot with an anti-DNP antibody (a). Optical density of all the bands was analyzed and expressed as a percentage of the control group value (b). Each column and error bar represent the mean ± SEM. * Different from control; # different from LPS-24 (p < 0.05 for each paired comparison).

Blockade of the HO Pathway by ZnPP-IX Deteriorates Endotoxin-mediated Diaphragmatic Dysfunction

To ascertain whether the heme oxygenase system plays a role in protection against endotoxin-mediated diaphragmatic dysfunction, we used an inhibitor of this enzymatic pathway. In LPS-untreated animals, administration of ZnPP-IX significantly reduced diaphragmatic HO activity at 6 h (p < 0.05 versus control; Figure 3). Similarly, ZnPP-IX markedly decreased HO activity by 40 and 25% in LPS-treated rats at 6 and 12 h, respectively (p < 0.05 versus LPS). Interestingly, HO-1 protein expression in the diaphragm of untreated rats was slightly augmented 24 h after injection with ZnPP-IX (Figure 1a); this is not surprising as blockade of heme oxygenase activity has been shown to induce the HO-1 gene (22).

Administration of ZnPP-IX in untreated animals produced a small but nonsignificant decrease in diaphragmatic force (Table 1). Notably, the decrease in diaphragmatic force observed in endotoxemic rats was markedly worsened by treatment with the heme oxygenase inhibitor. Specifically, peak twitch and maximal tetanic tension at 24 and 48 h were significantly lower in endotoxemic rats administered ZnPP-IX than in animals treated with LPS alone (p < 0.05 at each time point; Table 1). Identical results were obtained for all the frequencies of stimulation examined (data not shown). Most importantly, ZnPP-IX completely prevented the recovery of diaphragmatic force observed 48 h after treatment with LPS alone.

To avoid systemic effects of ZnPP-IX, we assessed diaphragmatic force after in vitro incubation of bundles from control and LPS-treated (24 h) animals with the heme oxygenase inhibitor. After exposure to ZnPP-IX for 30 min, bundles from both control and LPS-treated animals showed a similar and significant decrease in force (p < 0.05 versus baseline; Figure 5). However, while bundles from controls remained stable during the 30 min after ZnPP-IX exposure, bundles from LPS-treated animals exhibited a progressive decrease in force, resulting in a 50% decrease 30 min after ZnPP-IX (p < 0.05; Figure 5). This contractile response was similar for both peak twitch and maximal tetanic tension.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Maximal tetanic force changes after in vitro incubation of the bundles with ZnPP-IX. Bundles from diaphragms of control (open squares) and LPS-24 animals (solid triangles) were incubated for 30 min with 104 M ZnPP-IX. The bath was then rinsed and force was measured 10 and 30 min after the end of the incubation. Force was expressed as the ratio between the force at different times and the force before the beginning of the incubation (baseline). Each point and error bar represent the mean ± SEM. * Different from baseline values (p < 0.05); # different from control bundles (p < 0.05).

In untreated animals, ZnPP-IX significantly increased diaphragmatic MDA levels 24 h after administration (p < 0.05; Table 2), whereas no change was observed at 6 h. In endotoxemic animals, ZnPP-IX exacerbated the increase in MDA levels observed in the diaphragm 24 h after treatment with LPS alone (p < 0.05; Table 2). MDA in endotoxemic animals administered ZnPP-IX returned to basal levels at 48 h. Similarly, oxidized proteins in the diaphragm of endotoxemic rats injected with ZnPP-IX significantly increased compared with animals treated with LPS alone (p < 0.05; Figures 4a and 4b).

Administration of ZnPP-IX to control and endotoxemic animals did not modify cNOS and iNOS activity (Table 3).

Induction of HO-1 by Hemin Pretreatment Ameliorates Endotoxin-mediated Diaphragmatic Dysfunction

We wanted to explore whether upregulation of the HO-1 pathway could protect against LPS-mediated dysfunction of the diaphragm. Twenty-four hours after hemin administration, diaphragmatic HO-1 protein expression and HO activity were significantly increased as compared with controls (Figures 6a and 6b). When LPS was injected in hemin-treated rats, the high levels of HO-1 protein expression and HO activity were maintained during the following 24 h (Figures 6a and 6b). Interestingly, pretreatment with hemin completely prevented the decrease in force observed 24 h after LPS injection because peak twitch and maximal tetanic tension were similar to control groups (Table 1). Identical results in force were obtained for all the frequencies of stimulation (data not shown). MDA levels were significantly reduced in the diaphragm of endotoxemic rats pretreated with hemin compared with animals treated with LPS alone (Table 2). Notably, administration of ZnPP-IX to endotoxemic animals pretreated with hemin decreased HO activity and increased MDA content, thereby suppressing the protective effect on diaphragmatic force afforded by hemin (Figure 6b and Tables 1 and 2).

View larger version (39K): [in this window] [in a new window]   Figure 6.   Effect of HO-1 induction with hemin prior to endotoxemia. (a) Western blot analysis of HO-1 protein expression in whole diaphragmatic homogenates obtained 24 h after administration of either LPS (LPS-24), hemin (hemin-24), or LPS preceded 24 h before by hemin administration (hemin-LPS-24). Spleen homogenate is used as positive control. (b) HO activity in diaphragmatic microsomes (expressed as picomoles of bilirubin produced per milligram of protein during 30 min of incubation) from LPS-6 and LPS-24 animals, and animals treated with hemin (hemin 24); from LPS animals pretreated 24 h before with hemin and killed 6 and 24 h after LPS (hemin-LPS-6 and hemin-LPS-24, respectively); and from hemin-LPS-6 animals receiving ZnPP-IX 1 h after LPS (hemin-LPS-ZnPP-6). Each column and error bar represents the mean ± SEM. * Different from control (C); #different from LPS-6; § different from LPS-24; ** different from hemin-LPS-6 (p < 0.05 for each paired comparison).

Induction of HO-1 Protein in Diaphragm and Nondiaphragmatic Skeletal Muscles of Endotoxemic Rats

In control animals, HO-1 was expressed in all muscles examined but the protein levels were higher in left ventricular myocardium, soleus, and diaphragm than in EDL or rectus abdominis (Figure 7). After LPS administration a significant increase in HO-1 protein expression was detected in all muscles except EDL (Figure 7).

View larger version (12K): [in this window] [in a new window]   Figure 7.   HO-1 expression in different striated muscles in LPS-24 and control (C) animals. Positive control is a spleen homogenate. The 32-kD molecular mass marker is shown on the left. EDL = extensor digitorum longus.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study we examined the role of the HO pathway in the modulation of diaphragmatic muscle function in endotoxemic rats. We report that (1) diaphragmatic HO activity was reduced during the first 12 h after LPS administration, with no concomitant changes in HO-1 and HO-2 protein expression. This effect was followed at 24 h by upregulation of HO-1 protein and increase in HO activity, which was sustained up to 48 h. Interestingly, diaphragmatic HO-1 expression was primarily localized in skeletal myocytes; (2) diaphragmatic force generation was markedly impaired during the first 24 h of endotoxemia and was associated with muscular oxidative stress. A full recovery of diaphragmatic function was observed at 48 h; (3) inhibition of diaphragmatic HO activity by ZnPP-IX significantly worsened the deleterious effects of endotoxemia on force generation and oxidative stress at 24 h. In addition, the heme oxygenase inhibitor prevented the full recovery of diaphragmatic function at 48 h; (4) induction of HO-1 protein and increased HO activity by hemin administration prior to LPS challenge completely abrogated endotoxin-mediated diaphragmatic dysfunction and oxidative stress; and (5) endotoxemia increased HO-1 protein expression not only in the diaphragm but also in other skeletal muscles and in the myocardium. These data demonstrate for the first time a major protective role of the HO system against diaphragmatic oxidative stress and muscular contractile failure caused by endotoxin treatment. Furthermore, these results are consistent with a pathogenic mechanism that, to the best of our knowledge, has not been described before in the respiratory muscles or any other organ. Specifically, endotoxemia produces a biphasic change in diaphragmatic HO activity that reflects the course of the pathological state: an initial decrease in activity associated with muscle failure followed by a late increase likely involved in the recovery of muscle function.

Modulation of the HO System in the Diaphragm and Other Skeletal Muscles during Endotoxemia

We found that HO-1 and HO-2 proteins are expressed in the diaphragm of control rats; thus, both isoforms are likely contributing to the detectable basal HO activity. The presence of HO-1 and HO-2 has already been reported in other rat skeletal muscles (15, 31, 32). In LPS-treated animals, an early (6 and 12 h) decrease in diaphragmatic HO activity occurred without changes in the expression of HO protein isoforms, suggesting that this effect is caused by posttranslational modifications. Notably, a study by Ding and coworkers (33) shows that HO-2 enzymatic activity can be inhibited by NO and peroxynitrite in vitro; the cysteine residues present in the protein structure appear to be the targets of these two oxidizing molecules. In analogy with our findings, the same authors demonstrated that HO-2 protein expression remained unchanged despite decreased HO-2 activity (33). In contrast, HO-1 activity is not affected by NO or peroxynitrite, possibly because this isoform does not contain sulfhydryl groups (33). Because we have previously reported an early augmented production of NO and peroxynitrite in the diaphragm of endotoxemic rats (6, 10), the initial decrease in HO activity observed in this study after endotoxin challenge may be the consequence of HO-2 protein inactivation by these nitrogen species.

The late (24 and 48 h) increase in HO activity detected in response to LPS is likely reflecting upregulation of HO-1 protein, as HO-2 expression in endotoxemic animals did not change over time. However, in view of data showing that HO-2 catalytic activity is enhanced by phosphorylation in cultured neuronal cells (34), the participation of HO-2 in increasing HO activity cannot be excluded a priori. Diaphragmatic HO-1 induction during endotoxemia could be mediated by increased production of NO and reactive oxygen species (6, 10); this hypothesis is plausible because we have demonstrated that NO and NO derivatives are potent stimulators of HO-1 in skeletal muscle cells and other cell types (15, 20). In addition, HO-1 transcript is highly induced in skeletal muscle in vivo by strenuous exercise (14), a condition known to generate muscular oxidative stress (16). The time course of diaphragmatic HO-1 expression observed in this study after endotoxin resembles that described in lung epithelium after tracheal infusion of LPS (35), suggesting that HO-1 induction can be considered a long-lasting response to LPS challenge.

Endotoxin treatment increased HO-1 expression not only in the diaphragm but also in striated muscles with high oxidative metabolism such as soleus (predominant in Type I fibers) and cardiac tissue. Interestingly, LPS did not induce HO-1 in EDL, a muscle rich in Type II fibers (low oxidative metabolism). This is in agreement with findings by Vesely and coworkers (31) showing a fiber type-specific pattern of HO-1 induction in skeletal muscles of rats treated with hemin. Collectively, these data suggest that increased generation of oxidant species mediates high HO-1 expression in muscles containing Type I fiber, a response that may lead to protection against oxidative stress. Consistent with this concept, it has been reported that other antioxidant enzymes are greatly expressed in muscles with high oxidative metabolism compared with those characterized by low oxidative metabolism (36, 37).

Role of the HO Pathway in Protection against LPS-mediated Diaphragmatic Dysfunction

Diaphragmatic force generation decreased 12 and 24 h after LPS administration and fully recovered at 48 h. A crucial protective role for the HO system against the reduction of force caused by endotoxin is sustained by the following experimental evidence. First, the initial reduction of HO activity in endotoxemic animals was accompanied by impairment of diaphragmatic contraction, which was further deteriorated by blockade of the HO pathway with ZnPP-IX. Second, the late upregulation of HO-1 with concomitant increase in HO activity by LPS correlated with the restoration of diaphragmatic contractile function and ZnPP-IX treatment impaired this recovery. Third, induction of HO-1 by hemin prior to LPS challenge considerably attenuated endotoxin-mediated diaphragmatic dysfunction. Although ZnPP-IX is a commonly used inhibitor of HO activity, previous reports argued about the specificity of this drug because high concentrations of metalloporphyrins also inhibit NOS and guanylate cyclase (38). However, as inhibitory effect of ZnPP-IX on NOS activity is unlikely because we did not find any difference in cNOS and iNOS activities in control and endotoxemic animals treated and untreated with ZnPP-IX. Moreover, although ZnPP-IX induced a small decrease in diaphragmatic force in control animals, this decrease did not reach significance, thus ruling out a direct deleterious effect of this inhibitor on muscular function. It must be noted that, in control animals, ZnPP-IX had no effect on protein oxidation but increased MDA levels. We have no explanation for this discrepancy, but it could reflect a different sensitivity of proteins and lipids to oxidative damage.

The fact that ZnPP-IX significantly reduced HO activity in control and in LPS- and hemin-treated animals confirms its ability to inhibit the HO enzymes in our model.

Systemic induction of the HO-1 gene by LPS could improve diaphragmatic metabolic supply and, thus, muscle function (41) by regulating vessel tone, as demonstrated in the liver during sepsis (42). However, in the present study, immunohistochemical analysis revealed that HO-1 is mainly expressed in myocytes and not in the vessels of the diaphragm from endotoxemic rats. This result suggests that the protective effect of HO-1 against LPS-mediated diaphragmatic failure is related to the action of this stress protein to modulate specific muscle functions. This hypothesis is further corroborated by our in vitro experiments showing that direct application of ZnPP-IX to diaphragmatic bundles from control and LPS animals reproduced the deleterious effect of systemically administered ZnPP-IX.

Several mechanisms could be responsible for the cytoprotective action of the HO system against diaphragmatic contractile failure induced by LPS. First, this protective role could be related to antioxidant properties of HO. Indeed, HOs are important intracellular antioxidant enzymes by virtue of their ability to degrade the prooxidant heme and generate biliverdin and bilirubin, two effective free radical scavengers. Bilirubin is probably the most abundant endogenous antioxidant in mammalian tissue (43) and has been shown to efficiently scavenge peroxyl radicals in vitro and in vivo (34, 44, 45). In the present study, endotoxin produced diaphragmatic oxidative stress, as revealed by increased MDA and carbonylated protein contents, two well-established indexes of lipid and protein modifications by oxidative damage (29, 46, 47). It must be noted that in endotoxemic rats treated with ZnPP-IX, inhibition of HO markedly attenuated diaphragmatic function 48 h after LPS inoculation while there was no longer evidence of increased oxidative stress, assessed by MDA content. The explanation for this finding is not clear, but a similar dissociation has been observed previously in endotoxemic rats (8) and in rats with chronic respiratory loading (48). It could be related to a temporal dissociation between a delayed recovery of muscular function after peroxidation of membrane lipid structures involved in excitation-contraction coupling. Indeed, the temporal correlations between MDA levels, membrane structure and/or function, and muscular contractility are not well defined. An additional mechanism of protection by the HO system could be related to downregulation of iNOS expression by degradation of heme, an essential cofactor for iNOS protein assembly and activity (13, 49). This hypothesis can be sustained by our findings showing that the increase in iNOS expression and activity by LPS in the diaphragm were no further detected at the time of induction of HO-1 expression and activity (48 h after LPS) (6). In addition, an increased level of ferritin related to HO induction can be responsible for the protective role of the HO pathway. Indeed, ferritin, an iron chelator, protects against cellular damage induced by liberation of free iron during oxidative stress. (50, 51). However, there are no data in the literature regarding the effect of ferritin on skeletal muscle contractility. Finally, a potential defensive role of CO, the other product of heme oxygenases, cannot be excluded a priori. Indeed, CO can inhibit NO synthase activity (52) and may influence other intracellular functions implicated in the progression of sepsis and other pathophysiological states (53).

In conclusion, the present study demonstrates that the HO pathway is a major cellular protective system against oxidative stress in the diaphragm during sepsis. Pharmacological agents that upregulate HO-1 expression appear therefore as a promising strategy for preventing sepsis-induced muscular dysfunction, as well as other muscular pathological situations related to oxidative stress, such as muscular fatigue.