INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has been well established that delayed skeletal muscle injury and inflammation can be observed several days after exercise, particularly after eccentric muscle contraction (1). Reid and co-workers (5) reported that ventilatory failure induced by resistive loading was associated with diaphragm fiber injury and inflammation after prolonged loading by tracheal banding. We also observed that delayed diaphragm injury and inflammation occurred 3 d following a very brief period of intense inspiratory resistive loading (IRL) (6). However, whether this delayed diaphragm injury could reduce diaphragm contractility remains unknown. Because it is apparent that respiratory muscle contractility, particularly diaphragmatic contractility, is important in sustaining ventilation, the impact of diaphragm muscle injury on force production is of interest.

Delayed or secondary muscle injury and inflammation in limb muscles have been shown to cause marked loss of force (7, 8); we therefore hypothesized that delayed diaphragm injury produced by acute intense IRL could also lead to a significant reduction in diaphragm force production. We have already shown that histological evidence of delayed diaphragm injury and inflammation occurs after high-intensity IRL but not moderate IRL (6). However, although moderate-intensity IRL did not produce histological evidence of diaphragm injury, it did lead to activation of one of the major degradative enzyme systems, Ca²⁺-activated neutral protease (calpain). In the present study, therefore, we examined the effect of high and moderate intensities of acute IRL on diaphragm force production in the New Zealand White (NZW) rabbit.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Three groups of NZW rabbits (n = 7/group) with comparable body weights (wt), ranging from 3.5 to 4.5 kg, were studied with the animals in the supine posture. Each rabbit was randomly assigned to one of the following groups: control, moderate IRL (target airway opening pressure [Pao] of 30 cm H₂O), and high IRL (target Pao 45 cm H₂O).

On Day 1

All rabbits were anesthetized with ketamine (initial dose 35 mg/kg body wt, administered intramuscularly) and xylazine (initial dose 7 mg/kg body wt, administered intramuscularly). The rabbits were intubated. A constant level of anesthesia was maintained by giving supplementary doses of anesthetics (one-half of the initial dose) every 30 min. Arterial blood gases were obtained from a catheter placed in an ear artery and measured with a Corning (Corning, NY) 168 pH/blood gas analyzer. A two-way valve (no. 2600; Hans Rudolph, Kansas City, MO) was connected to the endotracheal tube and an adjustable needle valve was connected to the inspiratory port to apply the inspiratory load.

Pao was measured from a side port inserted at a right angle into the endotracheal tube. Two 5-cm latex balloons were placed via the oropharynx, one in the stomach (filled with 1.0 mL of air) to measure abdominal pressure (Pab) and the other in the midesophagus (filled with 0.6 mL of air) to measure esophageal pressure (Pes). These balloons were connected by thin polyethylene tubing (50 cm, PE-200) to a differential pressure transducer (MP 45; Validyne Engineering, Northridge, CA). From the outputs of these two signals the transdiaphragmatic pressure (Pdi) was obtained (Pdi = Pab  Pes). Body temperature was kept between 37 and 39° C with a heating pad.

During the experiment, the rabbits in the control group went through the same protocol as the two IRL groups but breathed against no inspiratory load. The rabbits in the two IRL groups breathed against a moderate IRL and a high IRL, respectively, for 1.5 h. The intensities of the IRL produced a diaphragm pressure time index (PTIdi) of 0.17 ± 0.01 in the moderate-IRL group and 0.23 ± 0.05 in the high-IRL group. The Pdi_max was assumed to be 65 cm H₂O in the calculation of PTIdi and this figure was derived from our previous work (9). An appropriate oxygenation level in the IRL groups was ensured by applying 100% oxygen at 3 L/min to the inspiratory side of the adjustable needle valve. The Pa_CO2 increased substantially during both moderate and high IRL. It averaged 95 ± 15 mm Hg (mean ± SD) in the moderate IRL group (corresponding to pH 7.14 ± 0.13) and 131 ± 32 mm Hg in the high-IRL group (corresponding to pH 7.01 ± 0.14) at the end of the IRL period (90 min). Pa_O2 was maintained above 80 mm Hg in all the rabbits during the loading period. The animals were then allowed to recover.

On Day 3

All rabbits were reanesthetized. This time, instead of intubation, a tracheostomy was performed and a tracheostomy tube inserted between the third and fourth tracheal cartilage rings. The methodology for measuring Pao and Pdi was similar to that on Day 1. In addition, the abdomen was opened and a pair of bipolar stainless steel hook electrodes was placed in the costal region of the diaphragm bilaterally in order to monitor diaphragm action potentials elicited during bilateral electrical phrenic nerve stimulation.

The phrenic nerves were exposed bilaterally just above the apices of the lung. An effort was made to expose the entire trunk of the nerve without inducing a pneumothorax. A pair of bipolar platinum wire electrodes (under mineral oil) was placed around the entire trunk of the phrenic nerve. Diaphragm force production was assessed by measuring twitch Pdi and Pdi at frequencies from 10 to 80 Hz during bilateral supramaximal phrenic nerve stimulation. The twitch Pdi and Pdi-frequency measurements were repeated after a high IRL with PTIdi around 0.22 in all the rabbits, including those in the control group, in order to examine whether the diaphragms from the animals subjected to moderate and high IRL on Day 1 were more fatiguable than controls.

To minimize the effect of changes in chest wall and diaphragm configuration during epiphrenic stimulation, the upper part of the abdomen including the lower part of the rib cage was bound (abdominal pressure increase of 10-13 cm H₂O) with an inflated cuff of a blood pressure-measuring apparatus during phrenic nerve stimulation. Five single supramaximal stimuli (0.2-ms duration, three times the twitch threshold voltage) were first delivered to the phrenic nerve to obtain twitch Pdi. The phrenic nerves were then supramaximally stimulated (model S45; Grass, Quincy, MA) for 1 s with square-wave pulses (0.2 ms in duration, three times threshold voltage) at frequencies of 10, 20, 30, 40, 50, and 80 Hz. All twitch stimulations at different frequencies were performed between breaths during brief airway occlusions at FRC. The twitch stimulation was performed at 1-min intervals and the stimulations at different frequencies were performed at 2-min intervals. Stimulations at different frequencies were performed after twitch stimulation to avoid twitch potentiation. With each stimulation, signals of Pab and Pdi were recorded with a computerized data acquisition system (Direc; RayTech, Vancouver, Canada). Data files were 8 s in length at 1,000 samples/s.

At the end of the experiment, the costal diaphragm was removed and samples were cut from it. The specimens were then quickly frozen in isopentane precooled with liquid nitrogen and stored in a refrigerator at 70° C until analysis.

Assessment of Diaphragm Injury

Cross-sections of the costal diaphragm specimens were sectioned at 10-µm thickness with a cryostat microtome (Reichart-Jung, Buffalo, NY) kept at 20° C and were stained with hematoxylin and eosin (H&E). We examined the costal diaphragm alone because our previous results had shown that most of the injury occurred in the costal diaphragm (6). Injury of these muscles was assessed by the point-counting technique as described in principle by Weibel (10) and used by Reid and co-workers in their study of diaphragm injury (5). Briefly, the area fraction (A_A) of normal muscle, abnormal and inflamed muscle, and connective tissue was determined from cross-sections of each biopsy sample (H&E stain) using a light microscope equipped with a camera lucida (Labophot; Nikon, Tokyo, Japan) and a computer system for point counting. The image of a 63-point grid from the computer monitor was projected via the camera lucida onto the image of the muscle cross-section viewed at ×400. Points projected on the cross-section were assigned to three categories: (1) normal muscle, (2) abnormal and inflamed muscle, including viable muscle with abnormal morphology, necrotic muscle, necrotic muscle with inflammatory cells, and inflammatory cells (around which no evident outline of a muscle cell could be seen), and (3) connective tissue (fat and collagen). The number of points in each of these three categories was expressed as a percentage of the total number of points in the three categories to determine each A_A value (%).

Data Analysis

Data obtained during stimulation were analyzed with a commercial software program (Anadat; RHT-InfoDat, Montreal, Canada) and the peak values of twitch Pdi and Pdi at different frequencies were measured. Differences in area fractions, twitch Pdi, and Pdi-frequency values, both before and after the high IRL on Day 3, among the three groups were examined by analysis of variance (ANOVA) with a Tukey test as the post hoc test. Student's paired t test was used to compare twitch Pdi obtained at baseline and after the high IRL on Day 3. Values represent means ± SE unless otherwise stated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Assessment of Diaphragm Injury

Three days after the initial inspiratory load, marked muscle injury and inflammation were observed in the costal region of the diaphragm on H&E-stained cross-sections in the high-IRL group, characterized by necrotic diaphragm fibers, flocculent degeneration, and profound influx of inflammatory cells into the necrotic fibers and the interstitial space. The interstitial space in the muscle was also widened. Table 1 shows the area fractions of the normal muscle, the abnormal muscle, and the connective tissue of the costal diaphragm as measured by the point-counting technique for the three groups. The A_A of abnormal muscle in the high-IRL group was significantly greater than that in the control group (p < 0.01). The A_A of connective tissue in the high-IRL group was also significantly higher than control (p < 0.01). However, the A_A of abnormal muscle and the A_A of connective tissue in the moderate-IRL group did not differ from control.

Twitch Pdi

Table 2 shows the absolute twitch Pdi (cm H₂O) recorded both before and after the high IRL on Day 3 in all three groups. There were no significant differences in baseline twitch Pdi (recorded before the high IRL on Day 3) among the three groups. Twitch Pdi obtained after the high IRL on Day 3, however, was significantly lower than baseline twitch Pdi in all three groups, particularly in the high-IRL group. Figure 1 shows the twitch Pdi obtained after the high IRL on Day 3 (expressed as percentages of baseline value) among the three groups. The decrease in twitch Pdi after the high IRL on Day 3 was significantly greater in the high-IRL group compared with control (Figure 1, p < 0.05).

View larger version (41K): [in this window] [in a new window]   Figure 1.   Twitch Pdi obtained after the high IRL on Day 3, expressed as percentages of baseline values, in the control, moderate-IRL, and high-IRL groups. An asterisk indicates significantly smaller (p < 0.05) twitch Pdi in the high-IRL group compared with controls. Values represent means ± SE.

Pdi-Frequency Curves

Figure 2 shows the absolute Pdi at all stimulation frequencies before (left) and after (right) the high IRL on Day 3 in the control, moderate-IRL, and high-IRL groups. The baseline Pdi values (before IRL) in the high-IRL group were significantly lower than in controls (p < 0.05 at 30, 40, and 50 Hz; and p < 0.01 at 80 Hz). However, there was no difference in the baseline Pdi at all frequencies between the control group and the moderate-IRL group (Figure 2, left). The Pdi values at different frequencies after the high IRL on Day 3 were significantly lower than those of controls at  30 Hz in the high-IRL group but the Pdi-frequency values in the moderate-IRL group were not different from control values (Figure 2, right). In Figure 3 the absolute Pdi values at all stimulation frequencies obtained at baseline (before the high IRL on Day 3) are compared with those obtained after the high IRL on Day 3 for the control group (left), the moderate-IRL group (middle), and the high-IRL group (right). After the high IRL on Day 3 as compared with their baseline values, there were decreases in Pdi-frequency values in the control group and in the moderate-IRL group, but the decreases were more pronounced in the high-IRL group.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Pdi-frequency curves on Day 3 at baseline (left) and after the high IRL (right) in the three groups. Significant differences (*p < 0.05 and **p < 0.01) between the high-IRL group and the control group are indicated. Values represent means ± SE.

View larger version (16K): [in this window] [in a new window]   Figure 3.   The absolute Pdi values at all stimulation frequencies obtained at baseline (before the high IRL on Day 3) are compared with those obtained after the high IRL on Day 3 for the control panel (left) the moderate-IRL group (middle), and the high-IRL group (right). Significant differences (*p < 0.05 and **p < 0.01) between the high-IRL group and the moderate-IRL group are indicated. Values represent means ± SE.

To determine whether force loss after the IRL occurred at high and/or low frequencies, the Pdi at each frequency was expressed as a percentage of the control value. There was no decline in baseline diaphragm force at any frequencies in the moderate-IRL group (Figure 4, left). However, the baseline diaphragm force in the high-IRL group was reduced at both high and low frequencies compared with both control and moderate-IRL groups. The decrease in force at 20 Hz was 24% of control and at 50 Hz was 27% of control. After the high IRL on Day 3, the decreases in diaphragm force at both low and high frequencies in the high-IRL group were more pronounced (p < 0.05 at 20 and 30 Hz and p < 0.01 at 40, 50, and 80 Hz) compared with both control and moderate-IRL groups as seen in Figure 4 (right). The decrease in Pdi for the high-IRL group at 20 Hz was 38% of control and at 50 Hz was 48% of control.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Pdi-frequency curves, expressed as percentages of mean control values, obtained before (left) and after (right) the high IRL on Day 3 for the moderate- and high-IRL groups. Significant differences (*p < 0.05 and **p < 0.01) between the high-IRL group and the moderate-IRL group are indicated. Values represent means ± SE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study provides both histological and functional evidence that marked diaphragm inflammation and injury occur 3 d after an acute period of high-intensity IRL and that this secondary diaphragm inflammation significantly decreases the force-generating capacity of the diaphragm. We observed that about 9% of the injured muscle appeared abnormal on histological examination by the point counting technique. In the high-IRL group, the decrease in baseline Pdi at 20 Hz was 24% of control and at 50 Hz was 27% of control. However, moderate IRL produced no significant injury histologically and did not produce a reduction in diaphragm force. Moreover, we observed that the injured diaphragm appeared more easily fatiguable or possibly more susceptible to repeated injury, as evidenced by the finding that after the high IRL on Day 3 the decreases in diaphragm force at both low and high frequencies in the high-IRL group were more pronounced compared with both control and moderate-IRL groups. These observations indicate that the decrease in diaphragm force occurs when injury is evident histologically and that the degree of IRL during acute loading, required to produce this type of injury, is well above the fatigue threshold.

The main purpose of the present study was to examine the impact of delayed diaphragm injury and inflammation on diaphragm force development, as there was no information (to our knowledge) addressing this question. In the present study, the diaphragm force production in the animals subjected to a high IRL on Day 1 was markedly reduced. However, only 9% of the costal diaphragm demonstrated abnormal muscle on histological sections on Day 3, whereas the decreases in diaphragm force production were much greater (30% on Day 3). The amount of injury on histology was similar to that (13% incidence of abnormal fibers) reported by Reid and coworkers (5) in chronic tracheal-banded hamsters. Secondary or delayed injury in limb muscles after exercise has been shown to be associated with disproportionate force loss (7). McCully and Faulkner (8) examined the effect of muscle injury on maximum isometric force (P₀) in mouse limb muscles. They found that 3 d after lengthening contractions, 37 ± 4% of muscle fibers had degenerated on histological sections and P₀ was 22 ± 3% of the control values. The amount of damage in this study was much larger than in other studies where contractions were not purely eccentric (1, 2, 11, 12). Nevertheless, the amount of injury histologically was found to be 37%, which was much less than the 78% drop in force developed by the muscle.

The disproportionately greater force loss compared with the amount of diaphragm tissue injury on histological examination is likely due to the following reasons. First, light microscopy of muscle sections may not reveal the ultrastructural abnormalities and molecular abnormalities that occurred with injury. Second, muscle injury may occur focally within a myofiber so that a single tissue cross-section would not show the total proportion of muscle fibers injured at other points along the length of the muscle. A previous study has shown that in longitudinal sections stained with ruthenium red and counterstained with azure II-methylene blue the areas of muscle damage were not uniform throughout the length of the muscle or even within a single fiber (13). Therefore, although the examination of a single cross-section may provide an estimate of the proportion of injured fibers, it is perhaps too simplistic to expect this to be proportional to the force loss.

In the present study, as we previously reported (6), marked diaphragm fiber injury and inflammation were present in the costal diaphragm 3 d following a short period of high-intensity inspiratory resistive loading. No injury and inflammation were evident immediately after IRL (see control group, Table ). Injury was characterized by necrotic diaphragm fibers, flocculent degeneration, and profound influx of inflammatory cells both in the necrotic fiber and in the interstitial space. Previous studies in limb muscles indicate that muscle injury and inflammation can occur several days after exercise, particularly after eccentric muscle contraction (1). Limb muscle injury and inflammation are known to start immediately after a vigorous muscle contraction and the muscle inflammation peaks at about 3 d after the strenuous muscle activity (1, 7, 8). This muscle injury and delayed inflammation may be called secondary muscle injury and inflammation. In support of this secondary muscle injury concept, Armstrong and coworkers (1) reported that there was an initial peak in plasma lactate dehydrogenase (LDH) and creatine phosphokinase (CPK) levels immediately after exercise and another large peak appeared later on Day 3. The delayed peak at Day 3 was reported to correspond to the secondary injury and extensive muscle inflammation revealed microscopically at this time (7).

We found a reduction in diaphragmatic force at both high and low frequencies of phrenic nerve stimulation. This indicates that both high-frequency fatigue and low-frequency fatigue occurred in the injured diaphragm subsequent to IRL. The force loss in high-frequency fatigue is believed to occur as a result of failed excitation-contraction coupling, whereas in low-frequency fatigue it is the result of failure at the contractile element itself. Diaphragm injury involves disruption of the sarcolemma as well as disruption of the contractile elements. Therefore, if these mechanisms do explain the difference between high- and low-frequency fatigue, then it is not surprising that both low- and high-frequency force are reduced with muscle injury.

Recovery from high-frequency fatigue classically occurs within 10 to 15 min. We repeated high-frequency stimulation in four of the high-IRL group 30 min after the end of high IRL on Day 3 and found no recovery of force at high frequencies. Thus loss of force at high frequencies in injured muscles may well be a reflection of something other than the usual mechanism attributed to fatigue and failure of excitation-contraction coupling. Results from eccentric contraction-induced injury in skinned muscle fibers suggest that about one-half of the force loss is due to failure of excitation-contraction coupling (14). In the setting of injury or damage this force loss may not be quickly reversible.

With low-frequency fatigue, force loss usually recovers with rest (15). However, in our model the force loss progresses even with rest. According to limb muscle literature, the force loss due to injury recovers slowly over a period of 7-30 d after the initial injury (16). Therefore, muscle injury may need to be considered as a separate entity that involves different processes contributing to loss and recovery of force.

The term tension-time index has been used in the literature (17) to describe human diaphragm fatigue. It is used in this study as it has been shown that when this index exceeds 0.15 the muscle is contracting at a fatiguing level. This degree of effort was associated with signs of ischemia in the muscle (18). Ischemia is a condition that is known to potentially activate free radicals and free radical activation could be associated with injury. Indeed, we found that just at the fatigue threshold with a pressure-time index of 0.17 (moderate-IRL group) there were some signs of inflammation in the diaphragm, but the extent of this inflammation was not different from that of controls and there was no associated force loss. Nevertheless, we have reported that one of the important muscle protein degradation systems, the calpain system, becomes activated even at this level of muscle contraction. However, only well above the fatigue threshold (high IRL) were both injury and substantial force loss recorded in the current study. Thus, it appears that whether diaphragm injury occurs depends on the level of muscle contraction and this level may have some as yet unknown relationship to the pressure-time index (17).

The pressure-time index, however, does not consider the type of muscle contraction, i.e., eccentric, concentric or isometric. Injury can occur in limb muscles with any form of muscle contraction. The diaphragm has been shown to be capable of contracting eccentrically. Measurements of diaphragm length changes by sonomicrometry have revealed that the diaphragm contracts quite heterogeneously under both spontaneous and loaded breathing conditions (19). During spontaneous breathing, contraction is uniformly concentric when the diaphragm is active. However, with loading parts of the diaphragm can be observed to contract eccentrically, particularly in the costal region (19). Thus eccentric muscle contraction during IRL may be contributing to diaphragmatic injury. These observations may also explain why less injury was observed in the parasternal intercostal muscles, where eccentric contraction may not occur (20).

Finally, other mechanisms may have also contributed to the force loss observed in this study. Indeed, it has been shown the tumor necrosis factor- (TNF-) or endotoxin, the inflammatory mediators released during sepsis, may be responsible for the diaphragm dysfunction observed in animal models of sepsis or in vitro experiments (21). In the current study we found that the inflammatory response was an important component of the diaphragm injury as evidenced by the profound influx of inflammatory cells into necrotic fibers and interstitial spaces. Therefore, inflammatory mediators, including free radicals released from inflammatory cells, might have contributed to the reduction of diaphragm force generation.

The injury that we have observed is associated with significant force loss. Short periods of inspiratory resistive loading used in this protocol resulted in delayed muscle injury and force loss. The effect of longer loading at lower levels of load or more frequent intermittent loads at these high intensities is unknown. Clearly, on the basis of our observations respiratory muscle injury is likely to be present in clinical conditions where ventilation is loaded either briefly or over more extended periods and, depending on the amount of injury, respiratory muscle contractility is anticipated to be compromised. Many respiratory diseases are characterized by continuous low loads or acute intense increases in load. These conditions can occur, for example, in association with episodes of acute bronchospasm in asthmatics or acute exacerbations of chronic obstructive pulmonary disease or, for example, during weaning trials in mechanically ventilated patients. More studies are required to examine the conditions that lead to respiratory muscle injury, the conditions that allow recovery from injury that lead to muscle adaptation, and the conditions that lead to persistent injury or muscle fibrosis.