INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung volume reduction surgery (LVRS) has become a widely used therapeutic option in selected patients with severe emphysema. It is believed that a large part of the subjective improvement experienced by the patients after LVRS is derived from improvement in inspiratory muscle function (1). In vivo, inspiratory muscle function can be assessed indirectly by measuring the maximal pressures developed by the muscles (mouth and transdiaphragmatic pressure). It has been shown that on an average basis, the maximal pressure generated by the inspiratory muscles increases after LVRS (2, 3).

Theoretically, this could be due to (1) changes in respiratory muscle and chest wall geometry or to (2) alterations in intrinsic properties of the respiratory muscles. The hyperinflation associated with emphysema has dramatic consequences on the mechanical advantage of the diaphragm (reviewed in [1]). Accordingly, there is evidence that the reduction in lung volume caused by LVRS is an important determinant of the improved diaphragmatic function observed in patients. However, how LVRS affects the intrinsic properties of the diaphragm is unknown and impossible to assess in patients.

We applied LVRS to an established model of elastase- induced emphysema in hamsters with the aim of analyzing the effects of LVRS on the diaphragm in vitro contractile properties and morphology.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Induction of Emphysema

Nine-week-old male Syrian Golden hamsters were purchased. One week after purchase, the animals were anesthetized with sodium pentobarbital (Nembutal; Sanofi Santé Animale Benelux, Brussels, Belgium; 60 mg/kg body weight [b.w.] intraperitoneally) and underwent transoral tracheal intubation with a 16-gauge catheter (Insyte-W catheter; Becton Dickinson, Madrid, Spain). Hamsters were instilled on a random basis with either porcine pancreatic elastase (Sigma 40 U/100 g b.w. and sodium chloride 0.9% added to a total volume of 0.4 ml, n = 20) or 0.4 ml sodium chloride 0.9% (n = 8, CTL). One of the animals instilled with elastase died soon after extubation due to tracheal perforation. The animals were kept in cages by 8 and provided with standard laboratory chow and water ad libitum. All experiments received approval of the animal experiments committee of the Medical Faculty of the Katholieke Universiteit Leuven.

Surgical Procedure

Eighteen weeks after intratracheal elastase instillation, the animals instilled with elastase underwent randomly either LVRS (n = 11, LVRS) or a sham operation (n = 8, SHAM). More animals were allocated to LVRS because of the expected mortality. They were anesthetized with sodium pentobarbital (Nembutal; Sanofi Santé Animale Benelux; 45 mg/kg b.w. intraperitoneally). Cefuroxime (Zinacef; Glaxo, Brussels, Belgium) was used as antibacterial prophylaxis (5 mg/100 g b.w. intraperitoneally). Hamsters were transorally intubated with a 16-gauge catheter (Insyte-W catheter, Becton Dickinson, Madrid, Spain). They were mechanically ventilated by means of a rodent ventilator (Harvard Apparatus Respirator, Model 645; South Natick, MA). A halothane tank (Fluotec 3; Cyprane, Heighley, UK) was placed in series between the inspiratory outlet of the ventilator and the animal's airways. Halothane (Fluothane; Zeneca, Destelbergen, Belgium) 1.5% in O₂ was intermittently administered during the surgical procedure. The surgical procedure was performed under clean and semisterile conditions. The anterior thorax was shaved and disinfected with iodinated alcohol. After incision of the tegument, a sternotomy was performed with the aid of an electric scalpel in order to avoid bleeding. The right pleura was incised. After apnea was produced by temporarily switching off the ventilator, the right lung was drawn out of the thoracic cage by means of lung grasping forceps. The two arms of a linear cutter (ETS-Flex 35 mm; Ethicon Endo-Surgery, Inc., Cincinnati, OH) loaded with thin staples used in human vascular surgery (Endopath TR35W; Ethicon Endo-Surgery, Inc., Cincinnati, OH) were then slid around the lung before the stapler-cutter was activated. Mechanical ventilation was then resumed. The same was performed on the left lung. Emphysema was macroscopically diffuse and was marked in the upper parts of the lungs. The resection aimed at removing 25 to 30% of the lung volume, as estimated visually. In practice, this was generally obtained by the resection of most of the right cranial and left upper lobes as well as the anterior part of the right middle and left lower lobes, according to Patra's description of the hamster's bronchial anatomy (4). The thorax was then closed by suturing the different planes and the animals were weaned from the ventilator. The animals of the SHAM group underwent the same procedure except for the lung resection. Once the animals resumed spontaneous ventilation they were extubated and placed in an atmosphere enriched in oxygen for a few hours.

Three hamsters of the LVRS group died soon after the end of the surgical procedure. Autopsy revealed an important amount of blood in one of the pleural cavities in the three cases. A fourth animal of the LVRS group died on the fifth operative day in a comatose state. Air leakage was never a problem.

Lung Volume Measurements

Nine weeks after the surgical procedure (and at equivalent age for the CTL animals), the animals were anesthetized with a mixture of sodium pentobarbital (3 mg/100 g b.w.), -chloralose (3.8 mg/100 g b.w.), and urethane (38 mg/100 g b.w.), as modified from Reid and coworkers (5). They were tracheostomized and the trachea was cannulated with a 16-gauge catheter (Insyte-W catheter; Becton Dickinson, Madrid, Spain) of which the narrowed tip was removed. Measurements of functional residual capacity (FRC) and total lung capacity (TLC₂₅) were performed using a setting modified from Koo and coworkers (6). Rather than using a constant volume plethysmograph, we used a flow plethysmograph. Accordingly, a pneumotachograph (Hans Rudolph 8420B; Hans Rudolph, Inc., Kansas City, MO) connected to a differential pressure transducer (MP45; Validyne, Northbridge, CA; range ± 2 cm H₂O) was placed in the wall of the body box and volume changes were calculated by integration of the flow signal. All volume measurements were performed at least in triplicate in order to achieve three reproducible measurements. These were averaged for statistical analysis.

Diaphragm in Vitro Contractile Properties

After completion of the lung volume measurements, the animals were again mechanically ventilated and the diaphragm was carefully removed en bloc after laparotomy. Two small strips of costal diaphragm were dissected in order to assess in vitro contractile properties as described in detail previously (7). Briefly, the strips were placed at their optimal length (L_o), defined as the length at which peak twitch tension (P_t) was obtained. The following measurements were then successively performed: P_t, with determination of the time to peak tension (TPT) and half-relaxation time (RT_1/2) from the corresponding time-tension curve, tetanic force (P_o), fatigue run (330 ms 25 Hz stimulus each 3 s during 5 min). All experiments were performed at 37° C. After completion of the experiment, bundle length, thickness, and width were measured at L_o using fine calipers. The bundle was dip dried and weighed. The cross-sectional area (CSA) was obtained by dividing bundle weight by muscle specific density (1.056) and L_o. All forces were expressed per unit CSA (8). The whole diaphragm as well as the right scalenus medius and gastrocnemius were weighed after tripping and dipping dry.

Histological and Histochemical Procedures

A muscle strip obtained from the costal region of the diaphragm was handled as described in detail previously (7). Briefly, the muscle strip was fixed with the fibers oriented perpendicularly to a cork holder, and frozen in cooled isopentane. Serial cross sections parallel to the cork were obtained. Sections were taken for routine hematoxylin- eosin staining. The other serial sections were stained for myofibrillar adenosine triphosphatase after incubation at two acid pH (4.3 and 4.5) (9). The preincubation at pH 4.5 offered the best separation of the different fiber types and was generally exclusively used. It has been shown that ATPase staining is an accurate means to characterize muscle fibers, allowing for a distinction among types I, IIa, and IIx/b (10) (Figure 1). Morphometric examination was performed with a Leica microscope (Leitz Laborlux S., Wetzlar, Germany) at ×20 magnification. The latter was coupled to a digital video camera (Sanyo B/W CCD Camera, Model VC-2512; Osaka, Japan) and connected to a computerized image analysis system (Quantimet 500; Leica, Cambridge Ltd., UK). Ten fields composed of fibers oriented in a direction transverse to their long axis were analyzed for each diaphragm. At least 150 fibers were analyzed in each diaphragm. To correct fiber CSAs for the muscle shortening occurring after excision, measured CSAs were divided by the ratio L_o/excised unstressed length.

View larger version (143K): [in this window] [in a new window]   Figure 1.   ATPase stain (pH: 4.5) of a transverse section of the diaphragm from a CTL animal. Type I, IIa, and IIx/b fibers are identified (original magnification: ×20).

Data Analysis

Two diaphragm bundles were obtained from each animal for assessment of contractile properties. Mean values from each animal were taken for statistical analysis. Data from the different groups were compared using one-way analysis of variance (ANOVA). Differences between each group were assessed using a Newman-Keul's multiple comparison test. All analyses were performed using the NCSS Statistical Software (NCSS, East Kaysville, UT). Significance was set at p < 0.05. Data are shown as mean ± SD in the text, tables, and figures.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung Volumes

FRC and TLC₂₅ differed significantly between the three groups (Figure 2). SHAM and LVRS showed a 190 and 120% increase in FRC, respectively, compared to CTL. LVRS had a 25% lower FRC than SHAM. SHAM and LVRS had a 77 and 43% higher TLC₂₅, respectively, than CTL. TLC₂₅ was significantly lower by 20% in LVRS as compared with SHAM.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Lung volumes. Functional residual capacity (FRC) and total lung capacity (TLC25) are shown in CTL (open bars), SHAM (closed bars), and LVRS (gray bars). **,*** Significantly different from CTL (p < 0.01 and 0.001, respectively); , significantly different from SHAM (p < 0.05 and 0.01, respectively).

Body and Muscle Weight

Body weight was similar between groups at the time of tracheal instillation and at the time of surgery. However, at the time of the study, the body weight was significantly lower by 10 and 12% in SHAM and LVRS, respectively, compared to CTL (Table 1). The loss of weight secondary to both surgical procedures was present from the fourth postoperative day on (data not shown). Diaphragm weight was 14 and 18%, respectively, lower in SHAM and LVRS compared with CTL. However, when normalized for body weight, diaphragm mass was similar in the three groups (Table 1). By contrast, the weights of other muscles such as the scalenus medius and the gastrocnemius did not differ significantly between groups.

Diaphragm Contractile Properties

A strong tendency for a reduced L_o in SHAM compared to CTL was present (ANOVA: p = 0.06). However, no marked increase in L_o in LVRS compared to SHAM was seen (Table 2). It is noteworthy that the statistical analysis of L_o was disturbed by an unexplained outlier in SHAM. This animal had a 19.6 mm L_o, the longest of all groups, in spite of lung volumes in the highest range (FRC, 4.3 ml; TLC₂₅, 13.3 ml). When removing this outlier, the ANOVA p value for L_o became < 0.005 with the SHAM and the LVRS having a significantly lower L_o than CTL in multiple comparison test, with in addition a strong tendency for a difference between SHAM and LVRS (p = 0.1).

P_t was significantly decreased in LVRS compared to CTL and SHAM. SHAM also had a lower P_t than CTL but this did not reach significance. Conversely P_o was not significantly different across the three groups. There was a significant decrease in TPT in LVRS compared with SHAM. RT_1/2 did not differ significantly between groups (Table 2).

Resistance to fatigue was significantly increased in LVRS and SHAM compared with CTL, as demonstrated by a significantly smaller decline in tension at 210 and 300 s during the fatigue run (Figure 3). This also approached significance at 240 and 270 s (ANOVA: p = 0.06 and 0.07, respectively).

View larger version (17K): [in this window] [in a new window]   Figure 3.   Fatigue run. Open circles, closed squares, and open squares refer to CTL, SHAM, and LVRS, respectively. *LVRS and SHAM significantly different from CTL (p < 0.05). (*) ANOVA: p value = 0.06 and 0.07, respectively, at 240 and 270 s.

Diaphragm Histochemistry

A significant 25% increase in type IIa and a significant (18 and 26%, respectively) decrease in type IIx/b fiber proportion in both SHAM and LVRS compared with CTL were observed. No significant differences in fiber CSA for any fiber type were present. Accordingly, the proportion of the total CSA occupied by type IIa fibers increased significantly, in association with a significantly decreased proportion of the CSA occupied by type IIx/b fibers in both SHAM and LVRS compared with CTL (Table 3, Figure 4).

View larger version (28K): [in this window] [in a new window]   Figure 4.   Relative proportion of the diaphragm muscle tissue cross-sectional area (CSA) occupied by the three fiber types: black bars, type I fiber; light gray bars, type IIa fiber; dark gray bars, type IIx/b fiber. **,*** Significantly different from CTL (p < 0.01 and p < 0.001, respectively).

Diaphragm Histopathology

A qualitative analysis of the hematoxylin-eosin staining showed a clear increase in the number of small nucleated cells around the muscle fibers, of muscle fibers with centralized nuclei, of angulated cells, and of split cells in both SHAM and LVRS (Figure 5). No evidence for a difference in these patterns between LVRS and SHAM was present.

View larger version (125K): [in this window] [in a new window]   View larger version (144K): [in this window] [in a new window]   Figure 5.   Transverse section of the diaphragm from a CTL (A) and a SHAM (B) animal demonstrating an increased number of small nucleated cells around the muscle cells, of muscle cells with centralized nuclei, of angulated cells, and of split cells in SHAM. Hematoxylin-eosin stain (original magnification: ×20).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we made four striking observations demonstrating that LVRS did not alter the adaptations occurring at the level of the diaphragm in hamsters with emphysema. First, LVRS did not reverse diaphragm atrophy associated with emphysema. Second, length adaptation occurred both in SHAM and LVRS. Third, the fiber type shift observed in SHAM was not different from the one seen in LVRS. Finally, functional and structural alterations suggestive of chronic overuse were present in both SHAM and LVRS. The observations are discussed consecutively. The model, however, has some limitations, which are addressed first.

Limitations of the Model

We performed LVRS in a well-known model of emphysema in hamsters (11). It must be stressed that this is a model of pure emphysema, without significant intrinsic impairment of airway function, and, accordingly, without increase in the resistive work performed by the inspiratory muscles (12). Although the equivalent in patients is considered to be the ideal indication for LVRS, this situation rarely occurs in humans. Even if the resistive load faced by the inspiratory muscles is not increased in our hamster model, the dramatic alterations in the geometry of the diaphragm (and other inspiratory muscles) clearly increase the load these muscles face at the muscle fiber level. Accordingly, the model of elastase-induced emphysema appears suitable to address different questions related to the physiological basis of alterations occurring in respiratory mechanics and respiratory muscle function after LVRS. LVRS has been shown to improve different parameters related to airway function, blood gases, and in vivo respiratory muscle function in patients with severe emphysema (reviewed in [1]). In the present study, we focused on the intrinsic properties of the diaphragm as these are difficult to assess in patients. Although we did not assess in vivo diaphragmatic properties, we assume that, as demonstrated in patients (2, 3), the reduction in FRC after LVRS was associated with an improved in vivo diaphragmatic function. This is all the more likely since the reduction in FRC observed in our hamsters was of the same magnitude (25%) as the one observed in patients (see below).

Animals were operated on 18 wk after elastase instillation and studied 9 wk after operation. It has been demonstrated that the degree of emphysema reaches a plateau between 12 and 26 wk after elastase instillation (13). As skeletal muscle demonstrates a high level of plasticity (14, 15), the 9-wk interval after surgery appears sufficient to evoke adaptive changes in the diaphragm.

Others have performed LVRS in rabbits with emphysema induced by elastase aerosolization (16). They demonstrated that the increase in maximal expiratory flows was positively associated with the amount of lung resected (17). However, excessive lung excision was associated with a decrease in carbon monoxide diffusing capacity (DL_CO), with a resection of 30% of lung volume suggested to be optimal in this respect (18). According to FRC and DL_CO measurements, however, the level of emphysema achieved in this rabbit model at the time of surgery was mild (16, 18). It contrasts with the dramatic increase in FRC observed in the present study as well as in patients with severe emphysema selected for LVRS according to the currently accepted criteria.

As emphysema was diffuse in the present model, we did not target the lung resection to more affected zones. LVRS allowed for a reduction in the lung volumes in the range of 25%, close to the 30% target we aimed to achieve. This closely mimics the results obtained in patients (1). The significant mortality in the first hours after the procedure appeared to be mostly related to hemorrhage at the level of the lung suture. Air leaks never caused significant problems.

Diaphragm Structure

We found a significant atrophy of the diaphragm in SHAM that was not altered by LVRS. This is critical since muscle mass is a major determinant of maximal muscle force. In parallel, after the surgical procedure, both the SHAM and LVRS animals experienced a proportionally similar loss in weight as compared with CTL. This appears to suggest that the diaphragm atrophy would be nonspecific. However, the weight of the gastrocnemius and, interestingly, the scalenus medius did not alter significantly compared with controls. This suggests that diaphragm was specifically affected in both SHAM and LVRS.

Diaphragm atrophy has been formerly reported in patients with emphysema (19). Although the precise mechanisms responsible for it remain incompletely understood, the influence of multiple factors such as hypoxemia, systemic inflammation, and undernutrition has been evoked. The unaltered mass of other skeletal muscles in the present study suggests that factors specific to the diaphragm could have played a role in this atrophy. The shift toward a fiber type displaying a smaller CSA (see below) explains part of this muscle atrophy. Moreover, the chronically increased load on the diaphragm could be involved in this process by another mechanism. Indeed, it has been shown that the peroneus longus muscle of the cat (which demonstrates a fiber type distribution very similar to that of the hamster diaphragm) atrophies when subject to chronic stimulation. This atrophy is related to the rate of stimulation (20). As the load on the diaphragm increases with emphysema, the stimulation frequency of the diaphragm is expected to rise in this condition as well, as demonstrated in patients with severe chronic obstructive pulmonary disease (COPD) (21).

Similarly to Farkas and Roussos (22), we found a 10% reduction in L_o in the SHAM, which approached significance. This length adaptation has been shown to be allowed for by a loss of sarcomere in series (22). After LVRS, the reduction in lung volumes was only associated with a 2.5% increase in L_o compared with SHAM. When the outlier was removed from the analysis, however, the increase in L_o was in proportion to the decrease in FRC. This suggests that after LVRS, the diaphragm adapts once again to its new operating length. This observation is in keeping with studies showing that the length adaptation is a two-way phenomenon, with shortening of a muscle associated with a loss, and its lengthening associated with an addition of sarcomeres in series (23).

To the best of our knowledge, this is the first time that a shift in fiber type is observed in the model of elastase-induced emphysema in hamsters. However, most of the previous studies in hamsters with emphysema did not distinguish between the various type II fibers. The only available study with a distinction between type IIa and IIx/b fibers did not show any differences in fiber type proportion between animals with emphysema and control animals (24). We used a dose of 40 U/100 g body weight as opposed to the 25 U/100 g body weight used in the latter study. This is expected to cause more severe hyperinflation and, accordingly, more adaptive processes in the diaphragm. The shift toward a fiber type allowing for an increased resistance to muscle fatigue is a classical adaptive phenomenon in muscles that face a chronically increased load (14). It fits well with the high level of emphysema in SHAM. It also parallels the adaptations that occur in the diaphragm of patients with COPD, involving a shift from type II to type I fibers (25). This probably has a functional significance similar to the shift observed in the present study. It is in keeping with other adaptations observed in rats. Indeed, a limitation of the shift to the type II fibers also occurs in rat limb muscle in response to chronic stimulation (14).

We observed a significant increase in resistance to muscle fatigue in LVRS with a similar strong tendency in SHAM compared with controls. This is well explained by the shift in fiber type from type IIx/b to IIa observed in both SHAM and LVRS.

Nevertheless, the preservation of the above mentioned alterations in LVRS could be viewed as unexpected since the impairment in the diaphragm geometry is expected to decrease after LVRS. The decreased FRC allowed by LVRS may, however, be insufficient to reverse the shift in fiber types induced by emphysema. Another hypothesis is that the stretch stimulus imposed to the diaphragm immediately after LVRS could have sustained the shift from type IIx/b to type IIa fibers. Indeed, muscle stretch has been associated with such an adaptive process (26).

Although both SHAM and LVRS had a reduced body weight compared with CTL at the time of study, it is unlikely that the shift in fiber type observed would be secondary to nutritional deprivation. Indeed, loss of weight was probably related to the surgical stress, since it was present from the fourth postoperative day. Moreover, in contrast to our findings, malnutrition in hamsters with emphysema has been shown to be associated with atrophy in both type I and II fibers (27). In addition, undernutrition is not known to affect fiber proportion (28).

Diaphragm Function

A number of observations in our study suggest that chronic overuse was present in emphysema and that it was not altered by LVRS. First, we observed a significant decrease in P_t in LVRS compared with CTL. As previously reported in hamsters with emphysema (29, 30), the same tendency was observed for SHAM but this did not reach significance. Moreover the P_t/P_o was decreased in LVRS. It appears that this effect was independent of the fiber type proportion as it was present only in LVRS. In limb muscles, a low P_t/P_o has been suggested to occur in situations of chronic muscle overuse (31) and of long-term stimulation at high frequencies (32). Similarly, a paradoxical shift toward fast muscle twitch properties has been demonstrated in chronically overused muscles despite a shift toward nearly exclusive expression of type I myosin heavy chain (MHC-I). This was associated with an altered repartition of the myosin light chain isoforms within the muscle fibers that can potentially explain these findings (33). This mechanism could also explain the lower TPT in LVRS. However, as hyperinflation regressed, the diaphragm was expected to operate under less detrimental conditions after LVRS. Accordingly, observations of alterations compatible with muscle overuse in the diaphragm of hamsters with emphysema after surgical therapy that are equally pronounced as before should not be expected. Diaphragm stretch probably occurs in the immediate postoperative period due to a rapid reduction in lung volume. Stretch is known to be a severe stress for muscles. Along these lines, eccentric contractions are known to be fatiguing for the skeletal muscles (34). It is thus conceivable that this stress induced muscle overload in LVRS, at least temporarily.

Second, the histopathological study of the diaphragm of both SHAM and LVRS demonstrated alterations such as centralized nuclei, angulated and split cells, as well as an increased number of small nucleated cells around the muscle cells. Although these alterations are nonspecific, they are also observed after chronic high-frequency stimulation (20) or lengthening contractions (35).

Altogether, LVRS does not appear to be associated with a positive effect on the contractile properties of the diaphragm in hamsters with emphysema. The length adaptation appears to be reversible. The increased resistance to muscle fatigue secondary to severe emphysema is maintained but so are other alterations that have a potentially negative impact on the function of the diaphragm in situ. Accordingly, the diaphragm atrophy and the reduced twitch force could potentially compromise diaphragmatic function. If we assume that, as in patients, the in vivo diaphragm function is improved after LVRS, these observations suggest that changes in diaphragm geometry play a central role in this improved function.