Redox Balance and Cellular Inflammation in the Diaphragm, Limb Muscles, and Lungs of Mechanically Ventilated Rats

• ❖ High tidal volume ventilation produces oxidative stress and inflammation in the lung, but whether it produces similar effects in other organs is less clear

• ❖ In normal rats, high tidal volume ventilation produced minimal or no oxidative stress and inflammation in the skeletal muscle and the diaphragm

• ❖ High tidal volume ventilation may not induce diaphragmatic dysfunction via oxidative stress and inflammation

ACUTE lung injury (ALI) and acute respiratory distress syndrome (ARDS) are common conditions in intensive care units. Regardless of the severity of respiratory failure, the most common cause of death in patients with ARDS is shock and multiorgan failure.1Despite the lifesaving potential ofmechanical ventilation in patients with ALI and ARDS, complications arising out of this sort of therapy in critically ill patients should also be considered. In this regard, previous studies have shown that mechanical ventilation strategies using high tidal volume (V^T) and low positive end-expiratory pressure (PEEP) cause alveolar disruption and pulmonary edema, thus activating inflammatory pathways.2,3On this basis, it became clear that in patients with ARDS and ALI, certain strategies of mechanical ventilation might enhance lung injury, a condition known as ventilator-induced lung injury (VILI).4The pathophysiology of VILI, however, has not yet been fully elucidated.

Oxidative stress, defined as the imbalance between oxidants and antioxidants in favor of the former, has been suggested to be involved, among other factors, in the development and perpetuation of VILI.5For instance, a reduction in antioxidant activity along with an increase in malondialdehyde was observed in the fluid lining of mechanically ventilated lungs.6,7In addition, protein tyrosine nitration, a marker of nitrosative stress, was also shown to be increased in the lungs of patients with ALI.8Interestingly, evidence from animal studies demonstrated that inflammatory cells and proinflammatory cytokines also contributed to the pathogenesis of VILI.9,10 

Recently, the line has also been put forward that high V^Tmechanical ventilation induced organ injury other than lung injury and systemic inflammation in several animal models of VILI.11–17It remains unknown, however, whether high V^Tmechanical ventilation exerts deleterious effects on skeletal muscles, including the diaphragm, thus potentially influencing the weaning of patients from the ventilator. Moreover, studies investigating early molecular events induced by VILI in different organs and tissues are also lacking. On this basis, our study explored whether high V^Tmechanical ventilation per se  may induce early molecular changes such as increased oxidative stress levels and inflammation in the diaphragm, limb muscles, and lungs in an in vivo  experimental model of VILI applied to healthy rats for 1 h. In fact, the VILI model used in the current investigation has already been used for exclusively exploring pulmonary and systemic effects of high V^Tin healthy animals.10–19Accordingly, our objective was to specifically explore early molecular events involving oxidative stress, antioxidant mechanisms, and inflammation in the diaphragms, limb muscles (fast-and slow-twitch types), and lungs of healthy rats exposed to high V^Tmechanical ventilation for 1 h.

All animal experiments were conducted at the Hospital Universitario de Getafe (Madrid, Spain). This is a controlled study designed in accordance with both the ethical standards on animal experimentation (EU 609/86 CEE, Real Decreto 1201/05 BOE 252, Spain) at the Hospital Universitario de Getafe and the Helsinki convention for the use and care of animals. All experiments were approved by the Animal Research Committee at the Hospital Universitario de Getafe.

Three independent groups of pathogen-free healthy male Sprague-Dawley rats (Harlan Iberica, Spain, weight 325-375 g, n = 8/group) were studied using well-validated methodologies, which have been previously published by our group and others.10–19Two different ventilatory strategies were used in healthy rats to evaluate the effects of lung overdistension in the presence or absence of PEEP10–19: moderate V^Tmechanical ventilation (V^T9 ml/kg, PEEP = 5 cm H²O), Group 1, and high V^Tventilation (V^T35 ml/kg, PEEP = 0 cm H²O), Group 2. In both groups of rats, respiratory parameters were set as follows: respiratory rate, 70 bpm; inspiratory time, 0.3 s; expiratory time, 0.56 s; and inspired fraction of oxygen, 0.45. All animals were initially ventilated for an equilibration period of 10 min using moderate V^Tventilation parameters before they were assigned to either Group 1 or Group 2 (moderate and high V^Tmechanical ventilation, respectively). Both groups of animals were mechanically ventilated for 1 h, after being assigned to the corresponding group of the mechanical ventilation strategy. A third group of nonventilated rats (intact animals) was also studied (Group 3, control). Importantly, in ventilated animals, respiratory drive was abolished by the high respiratory rate set in the ventilator, because no spontaneous breaths were observed. Lung mechanics, hemodynamics, and circulatory conditions were measured both at time 0 and after the 1-h period of mechanical ventilation in each group of ventilated rats.

At the end of the experimental period (1 h), the following muscles and organs were removed in all animals during anesthesia: diaphragm, gastrocnemius, soleus, tibialis anterior, and lungs. The muscle and organ specimens were immediately frozen in liquid nitrogen and subsequently stored at −80°C. Furthermore, tissue specimens corresponding to the diaphragm, gastrocnemius, and lungs were also immersed in an alcohol-formol bath for 2 h and were embedded in paraffin. Frozen tissues were used for the redox marker analyses (immunoblotting), whereas paraffin-embedded tissues were used for the assessment of inflammation (immunohistochemical analyses). (See also Animals and Study Protocol in Supplemental Digital Content 1, in which full details on the animal experimentation and protocol is provided, http://links.lww.com/ALN/A569.)

All biologic analyses were conducted in the same laboratory at the IMIM-Hospital del Mar (Barcelona, Spain).

The effects of reactive oxygen species (ROS) and reactive nitrogen species on muscle proteins were evaluated using immunoblotting according to the methods published previously.20–26(See also Oxidative Stress Markers in Supplemental Digital Content 1, in which full details on the redox balance analyses are provided, http://links.lww.com/ALN/A569.)

To evaluate the presence of inflammatory cells in these muscles, immunohistochemical analyses were conducted in all the diaphragms and gastrocnemius of the three groups of animals by using similar methods published else where.27,28(See also Inflammation in Diaphragm and Gastrocnemius Muscles in Supplemental Digital Content 1, in which full details on the muscle inflammation assessment are described, http://links.lww.com/ALN/A569.)

Inflammatory infiltrates were determined in the lungs of the three groups of rats using hematoxylin-eosin staining according to the methodologies previously published.29,30See also Inflammation in the Lungs in Supplemental Digital Content 1, in which full details on the evaluation of lung inflammation levels are provided, http://links.lww.com/ALN/A569.

This study is designed on the basis of a two-tailed hypothesis. Clinical and physiologic data are presented as mean (SD). Biologic data are presented as median and interquartile range in the tables, whereas they are presented as box and whisker plots in the figures. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, 12.0 version for windows, SPSS Inc., Chicago, IL). Mechanical ventilation-induced changes in clinical and physiologic variables were explored using paired Student-t  test compared with baseline in both moderate- and high-V^Trat groups. Kruskal-Wallis test was used to examine significant differences among the three study groups, for each biologic variable and tissue. Furthermore, nonparametric Mann-Whitney U test was also used to specifically explore significant differences in biologic variables of the study between each group of mechanically ventilated rats and control animals. The sample size was based on previous studies10–19and on assumptions of 80% power to detect an improvement of more than 20% in measured outcomes at a level of significance of P  less than or equal to 0.05. In all the biologic variables involving the study of redox balance (total reactive carbonyls, malondialdehyde-protein adducts, protein tyrosine nitration, Mn-superoxide dismutase (Mn-SOD), and catalase), mean difference between groups was estimated at a minimum of 25%, and SD was approximately 25-30% of the mean value for each of the variables.

Compared with time 0, peak airway pressure was significantly increased in rats exposed to high V^Tfor 1 h, while exhibiting a significant decrease in the dynamic compliance of the respiratory system (table 1). Furthermore, high-V^Trats showed hypotension, hypoxemia, acidosis, and hyperlactatemia after the 1-h ventilation period compared with levels at time 0 (table 1).

Representative immunoblots corresponding to protein carbonylation in the diaphragm and gastrocnemius muscles of the rats are illustrated in figures 1A and B, respectively. In the diaphragms, total protein carbonylation was significantly lower in high-V^Tanimals than in either moderate-V^Trats or control animals (fig. 1C). In the gastrocnemius, protein carbonylation levels were significantly reduced in high-V^Trats compared with control animals (fig. 1D). High-V^Tand moderate-V^Tanimals showed significantly lower levels of muscle protein carbonylation in the soleus and tibialis anterior compared with control rats (figs. 1E and F, respectively). In the rat lungs, reactive carbonyl levels did not significantly differ among the three study groups (table 2).

In the diaphragm and gastrocnemius of the rats (figs. 2A and B, respectively), total malondialdehyde-protein adducts levels did not significantly differ among the three groups. In the soleus and tibialis anterior (figs. 2C and D, respectively), high-V^Tand moderate-V^Tanimals showed significantly lower levels of muscle malondialdehyde-protein adducts compared with control rats. In the rat lungs, however, high-V^Tanimals showed significantly greater levels of total malondialdehyde-protein adducts compared with either moderate-V^Trats or control animals (figs. 2E and F).

In the diaphragm and gastrocnemius of the rats (figs. 3A and B, respectively), high-V^Trats showed significantly lower levels of muscle protein tyrosine nitration compared with moderate-V^Tanimals, but not controls. In the soleus and tibialis anterior of the rats (figs. 3C and D, respectively), high-V^Tand moderate-V^Trats showed lower levels of muscle protein nitration compared with control animals. In the rat lungs, protein tyrosine nitration levels did not significantly differ among the three study groups (table 2).

In the diaphragm, gastrocnemius, soleus, and tibialis anterior of the rats, Mn-SOD protein levels did not significantly differ among the three groups of animals (table 2). In the rat lungs, however, Mn-SOD protein levels were significantly reduced in both high-V^Tand moderate-V^Tanimals compared with the controls (fig. 4A). Moreover, lung Mn-SOD levels were significantly lower in the high-VT rats compared with moderate-V^Tanimals (fig. 4A).

In the diaphragm, gastrocnemius, soleus, and tibialis anterior of the rats, catalase protein levels did not significantly differ among the three groups of rats (table 2). Nevertheless, in the rat lungs, catalase protein levels were significantly reduced in both high-V^Tand moderate-V^Tanimals compared with controls (fig. 4B).

Human tonsils were used as positive controls for both leukocyte and macrophage immunohistochemical identification (figs. 5A and B, respectively). Figures 5C and Dillustrate inflammatory cell infiltration within the diaphragm fibers of a high-V^Trat and a control animal, respectively. Although intramuscular inflammatory cell infiltration was extremely low in the respiratory and limb muscles of the three study groups of rats, there was a statistically significant increase in inflammatory cell counts in the diaphragm and gastrocnemius (figs. 5E and F, respectively) of high-V^Trats compared with either controls or moderate-V^Tanimals.

Hematoxylin-eosin staining of a lung section from both control and moderate-V^Trats are shown in figures 6A and B, respectively. Abundant inflammatory cells were observed within the lung inflammatory cell infiltrates in moderate-V^Trats (fig. 6C). Interestingly, lung inflammatory cell infiltrates were more abundant, but were of a smaller size, in high-V^Trats (figs. 6D and E) compared with moderate-V^Tanimals (fig. 6C). Abundant inflammatory cells were observed within the corresponding infiltrates (fig. 6F). The number of inflammatory infiltrates was significantly greater in the lungs of high-V^Trats compared with either control or moderate-V^Trodents (fig. 6G). However, total lung inflammatory infiltrated area was significantly increased in both high- and moderate-V^Trats compared with control animals (fig. 6H).

In this study, in contrast to our initial hypothesis, early molecular events after administration of high V^Tmechanical ventilation to healthy rats were characterized by a significant reduction in both protein carbonylation and nitration in the diaphragm and limb muscles of the animals, while inducing a significant increase in protein oxidation in their lungs. Interestingly, in the muscle specimens, modifications in oxidative stress markers were not accompanied by any significant variation in the content of the antioxidant mechanisms Mn-SOD or catalase, whereas significant reductions in the levels of these antioxidant enzymes were observed in the lungs. Furthermore, although there was a statistically significant increase in inflammatory cell counts in the diaphragm and gastrocnemius muscles of high-V^Trats compared with control animals, such an increase was quite likely to be of small biologic relevance in terms of contribution to inflammatory and oxidative stress events in either respiratory or limb muscles of the mechanically ventilated animals. Finally, inflammatory cell infiltrates were greater in the lungs of mechanically ventilated rats compared with the nonventilated controls.

In the current investigation, against our original hypothesis, high V^Tmechanical ventilation induced a significant decrease in the levels of oxidative stress, as measured by protein carbonylation and nitration, in the diaphragms, gastrocnemius, soleus, and tibialis anterior muscles compared with nonventilated animals. Carbonyl formation (ketones and aldehydes) is an important detectable marker of protein oxidation. Carbonyl groups can be formed by the direct reaction of proteins with ROS, leading to the formation of protein derivatives containing highly reactive carbonyl groups,31,32and by Michael-addition reactions of lysine, cysteine, or histidine residues with unsaturated aldehydes (hydroxynonenal and malondialdehyde) formed during the peroxidation of polyunsaturated fatty acids.31,32Conversely, protein tyrosine nitration is usually considered a marker of excessive reactive nitrogen species production in the tissues. Peroxynitrite, which is formed from the near-diffusion limited reaction between nitric oxide and superoxide anions, is the most widely accepted mechanism of in vivo  tyrosine nitration in the skeletal muscles.33,34 

In the current investigation, modifications observed in protein carbonylation and nitration in both the respiratory and limb muscles of healthy animals ventilated at high V^Tmay also be part of the systemic effects of mechanical ventilation in the organs and systems other than the lungs. Indeed, high V^Tmechanical ventilation of healthy rats was already shown to induce cardiovascular, liver, and gut injury in addition to systemic inflammation.11–14,17,19In this study, high-V^Tventilated rats exhibited hypotension, hypoxemia, acidosis, and hyperlactatemia—findings that are in keeping with previous reports.17–19Although the organ and skeletal muscle blood flow was not measured in this study, it would be possible to conclude that perfusion of organs and muscles in high-V^Trats was quite likely to be reduced as a result of the hypotension exhibited by these animals after the 1-h period of mechanical ventilation. Decreased blood flow and hypoxemia may well account for the reduced oxidative stress levels detected in both the respiratory and limb muscles of high-V^Trodents after mechanical ventilation compared with control animals. Interestingly, protein carbonylation (both reactive carbonyls and malondialdehyde-protein adducts) and nitration (3-nitrotyrosine immunoreactivity) levels were also significantly decreased in the soleus and tibialis anterior muscles of mechanically ventilated rats, using moderate V^T(9 ml/kg). Although not tested in the current investigation, it could be hypothesized that mechanical ventilation per se  could alter vascular function and blood flow15,19in healthy animals, especially that of more distal and smaller muscles, such as the tibialis anterior and the soleus. Clearly, future studies will be designed to further explore this hypothesis.

It is worth noting that in none of the muscles examined, protein levels of the potent antioxidants Mn-SOD or catalase were modified by any of the mechanical ventilation strategies used in this study. These findings reinforce the concept that reductions in protein carbonylation and nitration levels detected in the diaphragm and limb muscles of mechanically ventilated healthy rats were most likely due to decreased ROS production within the skeletal muscle fibers. On the grounds that certain amounts of ROS and reactive nitrogen species are required for normal muscle force production, it could also be anticipated that decreased oxidative stress levels in the diaphragm and the limb muscles of mechanically ventilated animals may depress the muscle performance. Although not examined in the present investigation, Reid et al.  35already demonstrated that depletion of ROS by incubation of diaphragm fiber bundles with the antioxidant catalase or by enzymatic removal of superoxide anions substantially depressed muscle contractile function. The results of their investigations led to the conclusion that in unfatigued muscles, ROS are mandatory for optimal contractile performance by facilitating, for instance, excitation-contraction coupling.35In conclusion, mechanical ventilation-induced oxidative stress cannot be assumed in all organs or tissues, at least in early stages. Administration of antioxidants36–38might not all be justified in these specific models of VILI.

Despite the statistically significant increase in inflammatory cells detected in the diaphragms of ventilated rats (both high and moderate V^T) and in the gastrocnemius of high-V^Tanimals, such results are likely to be of small biologic relevance, because total levels of inflammatory cells were, in fact, extremely low in both respiratory and limb muscles of the three groups of rodents. These findings are in keeping with recent data obtained in our laboratory (Esther Barreiro, M.D., Ph.D., unpublished observations, 2009), in which intramuscular inflammatory cell counts were also shown to be extremely low in the diaphragm, external intercostals, and vastus lateralis muscles of chronic obstructive pulmonary disease patients and control subjects.

A redox imbalance in the lungs has already been demonstrated in several models of VILI.5–8In our study, protein carbonylation levels, as measured by malondialdehyde-protein adducts, but not protein tyrosine nitration, were significantly greater in the lungs of high-V^Trats than in controls. Lung content of the antioxidants Mn-SOD and catalase, however, were significantly reduced in both groups of mechanically ventilated rodents compared with controls. Previous reports have also shown a decrease in antioxidant capacity in healthy lungs exposed to high V^Tventilation6,39and an increase in malondialdehyde concentrations in the alveolar lining fluid.7The underlying mechanisms accounting for such a redox imbalance are probably based on the cyclic stretch of the lungs induced by mechanical ventilation. In this regard, exposure of the lung epithelial and endothelial cells to cyclic stretch increased ROS production as early as 30 min.40,41Besides, intense cyclic stretch also reduces glutathione and increases glutathione oxidation,40,42suggesting that redox imbalance as a result of cyclic stretch is a major contributor to the onset and perpetuation of VILI.5 

On the basis of the previous documented immunohistochemical tyrosine nitration of lung proteins in patients with ARDS,8nitrosative stress through protein nitration by reactive nitrogen species was also a proposed contributor to the pathogenesis of ALI. In this study, however, lung protein tyrosine nitration levels did not significantly differ among the three groups of animals. Moreover, we do not believe that tyrosine nitration of lung proteins through the action of leukocyte myeloperoxidases has taken place in mechanically ventilated rats, because the rise in inflammatory cell infiltrates was not accompanied by increased protein tyrosine nitration in their lungs. It is quite likely that mechanical ventilation for longer periods of time would have been associated with enhanced protein tyrosine nitration levels through either myeloperoxidase- or peroxynitrite-mediated mechanisms.

In the lungs of high-V^Tventilated rodents, the number and size of the inflammatory infiltrates was significantly greater than in the control group. Furthermore, the size of the infiltrated area, but not the number of infiltrates, was also significantly greater in moderate-V^Trats compared with nonventilated animals. This led us to the observation that inflammatory infiltrates, although less abundant, were slightly larger in moderate-V^Trats compared with high-V^Tanimals. Moreover, it should also be mentioned that the increase in inflammatory cell infiltrates observed in the lungs of mechanically ventilated animals is consistent with previous reports, in which the levels of several inflammatory cytokines, such as interleukin-1 β, interleukin-6, tumor necrosis factor-α, interleukin-8, and CXC chemokines,10,43,44were also shown to be increased in the lungs of healthy animals exposed to high V^Tmechanical ventilation.

In the current investigation, the first limitation is related to the clinical relevance of the findings encountered by using this specific model of VILI, in which lung injury was induced by the administration of high V^Tmechanical ventilation to healthy rats for a short period of time. In humans, 12 ml/kg has been shown to induce harmful effects on lungs,45whereas this V^Twould be harmless in healthy rats. Furthermore, standard mechanical ventilation in humans has been recently shown to induce marked atrophy of the diaphragm fibers and increased proteolysis as a result of inactivity.46It is worth noting that the ventilatory strategies used in the current investigation were within the scope of those used in previously published reports from our group and other investigators,10–19,47–49in which key physiologic alterations defining the pulmonary and nonpulmonary effects of high V^Tmechanical ventilation were demonstrated. In this regard, it is common practice to apply high-extreme conditions (clinical and/or physiologic) in completely healthy animals to generate contrast and to ensure that biologic differences among groups cannot be the subject of methodological concerns. On this basis, it is possible to conclude that compared with humans, administration of much larger V^Tis required in healthy rats to elicit lung injury in addition to pathophysiologic and biologic responses. Furthermore, these different ventilatory strategies were used in this study with the aim of exploring lung overdistension in the presence or absence of PEEP. Importantly, in previously published investigations,10–19these ventilatory strategies were shown to yield similar end-inspiratory lung volumes. Hence, we believe that administration of a high V^Tcan be perfectly justified in this model.

The second limitation is related to the short period of mechanical ventilation administered to the animals. However, our main goal in this experimental study was to explore the early molecular events involving redox balance of slow- and fast-twitch muscles and lungs in response to high V^Tmechanical ventilation in an animal model of VILI. Future studies will be definitely designed to assess medium- and long-term effects of VILI in organs and muscles by using a wider range of ventilatory strategies including lower V^T. Importantly, in the present investigation, inflammation and posttranslational oxidative modifications of both muscle and lung proteins have been evaluated in healthy rodents exposed to high V^T, thus showing significant variations in both the respiratory and limb muscles and the lungs compared with control animals.

Despite these recognized limitations, we believe that the study of events taking place very early after the pulmonary insult (high V^Tstretch) is of clinical relevance, because early occurrence of oxidative stress during high V^Tmechanical ventilation cannot be assumed in all organs or tissues. In fact, a reduction, rather than an increase, in oxidative stress was observed in both the respiratory and limb muscles of healthy rats exposed to large V^Tand VILI for a relatively short period of time.

Our study provides evidence that even with the use of extremely large V^Tand the creation of VILI in healthy rodents, there was no increase in oxidative stress in the diaphragm or limb muscles. In fact, at a particular time point, high V^Tmechanical ventilation did not induce the same oxidative and inflammatory events in all the rat tissues, whereas oxidative stress and inflammation increased in the lungs and the diaphragm, and both slow- and fast-twitch limb muscles exhibited a significant decline in oxidative stress markers and low levels of intramuscular inflammatory cell infiltration.

The authors thank Francesc Sánchez, Sandra Mas, and Maitane Pérez (Laboratory Technicians, Pulmonology Department-Muscle and Respiratory System Research Unit, Institut Municipal d'Investigació Mèdica-Hospital del Mar, Parc de Recerca Biomèdica de Barcelona, Barcelona, Catalonia, Spain) for their technical assistance in the laboratory and Dr. Rafael Marcos, M.D., Ph.D. (Associate Professor, Unitat d'Epidemiologia i Registre de Càncer de Girona, Pla Director d'Oncologia, Departament de Salut, Institut d'Investigació Biomèdica de Girona, Girona, Catalonia, Spain), for his generous help with the statistical analyses.