INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The diaphragm is a unique muscle with a primary role in inspiration. Chronic obstructive pulmonary disease (COPD) challenges the diaphragm by increasing inspiratory muscle demands due to higher resistive, threshold, and elastic loads (1), and by contributing to inspiratory muscle inefficiency or weakness (1). Similar to limb muscles, the diaphragm has been shown to respond to an overload by cellular and functional adaptations (2). In contrast to models of limb muscle overload, where rest and recovery occurs, diaphragm overload associated with COPD can be unrelenting and prolonged. The diaphragm's ability to adapt may be further impaired by factors that accentuate muscle weakness or limit regeneration. These include poor nutritional status, corticosteroids, and poor arterial blood gases (6). Thus, diaphragm dysfunction and injury may be due to the unrelenting overload compounded by adverse clinical factors that exceed the diaphragm's capacity to adapt.

Diaphragm injury has been shown in animal models of resistive ventilatory loading, including acute high-intensity loading for a couple of hours (7, 8) and chronic low-intensity loading over several days (9). Sarcolemmal defects in diaphragm muscle fibers have been shown to occur in dogs subjected to moderate intensity ventilatory loads repeated over several days (10). A couple of studies (11, 12) have described abnormalities in the lateral intercostal muscles of human subjects who experience ventilatory loading due to airflow obstruction, but there were several limitations in these studies that the present study has addressed.

Previous studies have reported a relatively low occurrence of morphological abnormalities in the lateral intercostal muscles (11, 12) that may be more postural than inspiratory in function (13). No study, however, has examined the primary muscles of inspiration in humans with airflow limitation except for description of fiber proportions (2). Although injury may occur in other respiratory muscles in response to ventilatory overload, animal studies indicate that the diaphragm is affected to a greater degree (7). Second, previous reports were limited to subjects with mild (11) or mild to moderate (12) chronic airflow obstruction and third, abnormal morphology was described qualitatively as being present or absent (11, 12) rather than quantitatively as a proportion. Finally, necrosis in the diaphragm from neonates who died from sudden infant death syndrome and inflammatory cells in both neonates and a small group of adults (14, 15) have been described. No studies to date, however, have described the location or the number of macrophages in the diaphragm. Macrophages are the predominant cell type observed after muscle injury (16, 17) and play important roles in both the inflammatory and regenerative stages of skeletal muscle injury (18). Neutrophils are usually observed very early after muscle injury (16) and other mononuclear cells such as lymphocytes are less commonly observed in muscle injury (16, 17).

The purpose of this study was to describe the nature of diaphragm injury, to quantify the injury and the number of macrophages at the light microscopic level, and to determine its association to airflow obstruction in a group of individuals with a large range of FEV₁ (16 to 122% predicted). We hypothesized that with increasing airflow obstruction (1) the area fractions (A_A) of abnormal diaphragm muscle would increase and the A_A of normal diaphragm muscle would decrease, and (2) the number of macrophages in the diaphragm would increase.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Thirty-one subjects undergoing thoracotomy surgery consented to participate in this study. Reasons for surgery included lung resection for cancer (n = 27), bullectomy for bullous disease (n = 1), lung transplantation for emphysema (n = 2), and lung volume reduction for emphysema (n = 1). Informed consent was obtained by using the procedure approved by the Ethics Committee at the University of British Columbia.

Inclusion criteria included adequate fitness for surgery and completion of preoperative pulmonary function tests. Exclusion criteria included acute exacerbation requiring hospitalization within 4 wk prior to surgery; chronic respiratory disease of a restrictive nature including interstitial lung disease, sarcoidosis, tuberculosis, neuromuscular disease, or chest wall abnormality; previous lower lobectomy or pneumonectomy; and inadequate muscle biopsy size or quality (minimum of five viable fields at ×500).

Muscle Sampling and Preparation

During surgery, a partial-thickness biopsy was obtained from the thoracic surface of the costal diaphragm midway between the central tendon and the costal insertion. Biopsies were embedded in O.C.T. (O.C.T Compound; Tissue-Tek, Torrance, CA), quick frozen in isopentane cooled to just above freezing in liquid nitrogen, and stored at 70° C.

Histology and Immunohistochemistry

Transverse cryosections 10 µm thick were cut and stained with hematoxylin and eosin (H&E). Transverse cryosections 6 µm thick were cut and mounted on slides coated with 3-aminopropylethoxysilane. The 6-µm sections were processed using the alkaline phosphatase- anti-alkaline phosphatase (APAAP) method and anti-macrophage monoclonal antibodies (Ber-MAC3; DAKO Corp., Carpinteria, CA) to label macrophages.

Thin cryosections mounted on coated slides were fixed in acetone for 10 min, air dried, then rehydrated with Tris-buffered saline (TBS) at pH 7.6. Sections were incubated in 5% normal rabbit serum diluted in TBS buffer containing 1% bovine serum albumin (1% BSA) for 15 min. Sections were incubated in the primary antibody for 1 h at room temperature. The primary antibody was a monoclonal mouse anti-human macrophage clone (Ber-MAC3), diluted to a concentration of 1:50 in 1% BSA. Next, sections were washed twice with TBS. The secondary antibody (rabbit anti-mouse IgG; DAKO Corp.) was applied for 30 min at room temperature, diluted to a concentration of 1:20 in 1% BSA, followed by two 5-min washes with TBS. Sections were incubated with APAAP complex (DAKO Corp.) for 30 min at room temperature, diluted to a concentration of 1:50 in 1% BSA, followed by two 5-min washes with TBS. The alkaline phosphatase substrate, Naphthol AS-B1 phosphate in 1% New Fuchsin, was prepared and immediately applied to the sections for 10 min. Sections were rinsed with TBS followed by tap water, lightly counterstained using Meyer's hematoxylin, and mounted with coverslips using a permanent mounting medium.

Positive controls were human lymph node sections processed as described above. Negative controls were muscle sections processed as per the above technique, using mouse immunoglobulin G (IgG) (DAKO Corp.) as the primary antibody, diluted to the same mouse IgG concentration as Ber-MAC3.

Quantitative Evaluation of H&E Sections

Computer-assisted point counting was used to quantify the A_A of normal muscle, abnormal muscle, and connective tissue in diaphragm cross sections. The set-up consisted of an IBM-compatible computer with a stereology software package (The Gridder; WillRich Technologies, American Megatrends Inc.) and a Nikon light microscope with a camera lucida. Using the software program, a grid consisting of 63 point-intercepts (7 × 9 rectangular pattern) was projected from the computer monitor via the camera lucida and superimposed onto the image of the muscle cross section viewed down the light microscope. The observer was blinded to the identity of the slide. At a magnification of ×500, each point-intercept was assigned to a specific category and entered into the Gridder program (see Table 1). The A_A for each category was defined as the percentage of points that fell on each feature defined in Table 1 relative to the total number of points superimposed on all viable fields of each biopsy cross section. A_A of normal muscle, abnormal muscle, and connective tissue were determined by calculating the proportion of points in (1) category 1, (2) categories 2, 5, 6, 7, and 8, or (3) categories 3 and 4, respectively (see Table 1 for description of categories). Photographs were used for reference when assigning point-intercepts to categories. For each muscle cross section, all possible nonoverlapping fields were counted in a standardized manner.

Quantitative Evaluation of Macrophages

Macrophages were quantified using the multipurpose image analysis system, Bioview (Infrascan, Richmond, BC, Canada). Muscle cross-sectional images (at least 15 fields per histology slide) were captured at a magnification of ×125 using a high-resolution video camera (25.4 mm Vidicon, 60 Hx resolution, 81 series; DAGE-MTI Inc., Michigan City, IN) attached to a light microscope (Microphot-FX; Nikon, Tokyo, Japan). The images were displayed on a 20-in. color monitor (1,024 × 1,024 pixels, 24 bit) (Sony Multiscan HG, Tokyo, Japan).

Macrophages and muscle fibers were counted if they were located within a rectangular large unbiased counting frame superimposed on the image. An unbiased counting frame allows the number of cells to be counted without artifact caused by cells on the edge of the field of view. Briefly, the counting frame consists of four lines that connect to form a rectangle. Two of these lines are considered to be "inclusion" lines and the remaining two are considered "exclusion" lines. Cells were counted if they were either wholly within the counting frame or touching the inclusion lines. Cells were not counted if any portion of the cell touched either of the exclusion lines (19). The area of the counting frame was 0.24 mm² and was calibrated at the beginning of each session. The number of macrophages was expressed per fiber and per mm².

Spirometry

Spirometry values were obtained for each subject in a pulmonary function laboratory by a registered respiratory therapist utilizing standard protocols as per ATS Standards of Care (20). Percentage of predicted forced expiratory volume in one second (% predicted FEV₁) values were obtained using validated prediction equations (21).

Statistical Analysis

The Pearson product-moment correlation coefficient (r) was used to determine the relationship between the % predicted FEV₁ and four variables: A_A of abnormal muscle, A_A of normal muscle, number of macrophages per fiber, number of macrophages per mm². Significance of the correlation coefficients was tested using one-tailed t tests. A significance level of p < 0.05 was chosen.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Descriptive Characteristics

Descriptive characteristics of the 21 subjects included in the study are presented in Table 2. Ten of the 31 subjects were excluded because of inadequate biopsy size or quality (see next section). Diagnoses for those included were lung cancer (n = 17), emphysema (n = 2), bullous disease (n = 1), and bronchiolitis obliterans (n = 1). Mean age was 62 ± 10 yr (range: 43 to 81 yr). Mean body mass index (BMI), defined as the ratio of the weight (kg) over the square of the height (m²), was 24.6 ± 3.9 kg/m² (range: 16.7 to 31.1 kg/m²). Mean % predicted FEV₁ was 74 ± 34% (range: 16 to 122%).

Area Fractions of Normal and Abnormal Diaphragm

Abnormal features included fibers with one or more internally located muscle nuclei (Figure 1, left panel), fibers with subsarcolemmal or perinuclear lipofuscin pigmentation, small angular fibers (Figure 1, left panel), inflamed fibers (Figure 1, right panel), and inflammatory cells in the endomysium or perimysium (Figure 1, right panel). Necrotic fibers and small basophilic fibers were rarely observed.

View larger version (53K): [in this window] [in a new window]   Figure 1.   Photomicrographs of abnormal morphology in H&E-stained human diaphragm cross sections. (Left panel ) Note the internal nuclei (black arrows) and the small angular fibers, one of which is in the center of the photomicrograph indicated by the open arrow. (Right panel ) Note the inflamed fiber shown in the center of the photomicrograph and the increased nuclearity between fibers in the interstitial space. Scale bars = 10 µm.

The A_A of normal muscle, abnormal muscle, and connective tissue were 66.2 ± 9.0%, 17.6 ± 7.2%, and 16.3 ± 4.2%, respectively. Mean values and ranges of data derived from point counting and subsequently combined for statistical analysis are shown in Table 3. The largest abnormal muscle category was the internal nuclei category, with a mean A_A of 14.7 ± 6.9% (range: 2.7 to 31.7%). Percent predicted FEV₁ was inversely related to the A_A of abnormal muscle (r = 0.53, p < 0.01) (Figure 2, upper panel) and directly related to the A_A of normal muscle (r = 0.37, p < 0.05). Percent predicted FEV₁ was not related to the A_A of connective tissue (Figure 2, lower panel).

View larger version (13K): [in this window] [in a new window]   Figure 2.   Scatter plot showing relationship between % predicted FEV1 and AA of abnormal diaphragm muscle (upper) and connective tissue (lower).

Macrophage Quantification

In diaphragm cross sections processed using immunohistochemistry, macrophages were distinguished by their deep pink cytoplasm and plasma membrane, which contrasted sharply with the pale basophilic muscle fibers and the basophilic muscle and interstitial nuclei (Figure 3). Macrophages were scattered throughout the endomysial and perimysial connective tissue layers. A few regions contained two or more closely approximated macrophages in widened regions of the interstitial space. No macrophages were identified within muscle fibers.

View larger version (102K): [in this window] [in a new window]   Figure 3.   Photomicrograph of macrophages in human diaphragm cross section. Note the three darker staining macrophages identified by antimacrophage monoclonal antibodies (Ber-MAC3; DAKO Corp., Carpinteria, CA). Scale bar = 10 µm.

The mean number of macrophages per fiber was 0.41 ± 0.18 (range: 0.08 to 0.74). The mean number of macrophages per mm² was 52 ± 19 (range: 18 to 85). Percent predicted FEV₁ was not related to the number of macrophages per fiber and (r = 0.05, p > 0.05) or to the number of macrophages per mm² (r = 0.01, p > 0.05) (Figure 4).

View larger version (12K): [in this window] [in a new window]   Figure 4.   Scatter plot showing relationship between % predicted FEV1 and number of macrophages in diaphragm expressed per mm2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we found that an increased severity of airflow obstruction is associated with an increased A_A of abnormal diaphragm muscle (Figure 2) and a decreased A_A of normal diaphragm muscle in people with a large range of airflow obstruction (% predicted FEV₁ range: 16 to 122% predicted) undergoing thoracotomy surgery. The percentage A_A of abnormal diaphragm ranged between 4 and 34% and included fibers with internally located nuclei, lipofuscin pigmentation, small angulated fibers, and some inflammation. This study is the first to report a quantitative description showing the A_A of abnormal diaphragm morphology in stable individuals with airflow limitation.

The A_A of abnormal muscle and connective tissue in the human diaphragm (17.6 ± 7.2 and 16.3 ± 4.2%, respectively, for a combined A_A of 33.9%) is higher than that described previously in animal studies. The A_A of abnormal diaphragm and connective tissue was 18% in hamsters tracheal banded for 6 d (9), 11% in hamsters tracheal banded for 30 d (22), and 15% in rabbits 3 d after a 1.5-h period of inspiratory resistive loading via an endotracheal tube (8). The clinical significance of these findings is threefold. First, the force loss from injury is disproportionately much higher than the amount of abnormality seen under the light microscope as shown in the diaphragm (8) and in limb muscles (23). Second, recovery of strength from injury is much more prolonged than reversal of fatigue. After eccentric loading of the elbow flexors, in otherwise healthy humans, the half-time of recovery appeared to be as long as 5 to 6 wk (24). Lastly, injury of the diaphragm may be an ongoing process in some individuals because it is difficult to rest this muscle, and coexisting clinical factors may potentiate weakness and fatigue (such as malnutrition, poor arterial blood gases, and medications, i.e., corticosteriods) and, thus, increase its susceptibility to muscle injury (6).

Similar abnormal morphology was found in the diaphragm, compared with abnormal features described previously in the diaphragm and other muscles of respiration, including internal nuclei (11, 12, 25), inflammatory cells (11, 14, 15), necrotic fibers (11, 15), basophilic fibers (11, 15), and small angular fibers (12). None of the previous studies, however, quantified the extent of the abnormal respiratory muscle morphology, thus increasing the difficulty in comparing and contrasting the results of the present study with previous work.

In the present study, approximately 15% of the cross-sectional area represented muscle fibers with internal nuclei. Previous research has not quantified the area fraction of fibers with internal nuclei, but expressed their prevalence as a percentage of total fiber counts. In these reports, the proportion of centrally nucleated fibers was 2 to 3% (26) and 4% (27) in normal limb muscles, and 5 to 10% in the limb muscles of subjects with neuromuscular disease (26). Although we did not count the number of centrally nucleated fibers, our results show that approximately 22% of the total point-intercepts that projected upon muscle fibers (versus elements of the extracellular matrix) fell upon muscle fibers with internal nuclei, indicating a relatively large cross-sectional area of fibers with internal nuclei.

The explanation for internal nuclei is unclear. Internal nucleation has been considered a quantifiable and cumulative index of necrosis and regeneration in the muscle of mdx mouse dystrophy (28). It appears that some fiber nuclei may remain internally located after cessation of necrosis and regeneration (28). Others have proposed that peripheral nuclei assume an internal location in order to direct ongoing regenerative efforts in response to muscle damage or forced contractile activity (29). It is possible that internal nuclei may reflect fiber type transformation, a suggestion proposed by Lindsay and coworkers (25), who observed internal nuclei and abnormal (neonatal) myosin expression in the diaphragm of individuals with heart failure. Alternatively, central and subsarcolemmal nuclei may indicate the maturation of myotubes at an advanced stage into adult fibers and peripheral nuclei may result from satellite cells fusing into existing fibers (29). Finally, internal nuclei may also be reflective of pending muscle fiber degeneration. The explanation for internal nuclei in the human diaphragm requires further study.

This study showed a relatively low A_A of inflamed or necrotic fibers in the diaphragm. There were few inflammatory cells in the interstitium and diaphragm morphology was not characterized by an inflammatory response. These findings contrasted with those of the animal models of resistive ventilatory loading, in which diaphragm injury was characterized by necrotic fibers, flocculent degeneration of the cytoplasm, and an influx of inflammatory cells in the interstitium and in necrotic fibers (7, 9). The difference in our study was not altogether unexpected, however, as subjects did not experience the acute increases in resistive loading imposed in the animal models. In the animal models, the main feature was acute increased resistive loading in a young, otherwise healthy adult animal whereas in our group of subjects, increased loading may result from increased elastic, threshold, and resistive loads as FEV₁ decreases. Our patients were clinically stable so that morphological abnormalities in their diaphragm likely reflect past episodes rather than recent episodes of increased loading. Further, other confounding variables such as medication (i.e., corticosteroids), nutrition, and other medical conditions may have contributed to or increased susceptibility to injury in this group of patients.

The necrotic fibers observed in the present study differed from those previously described by Silver and Smith (15), and Kariks (14). These investigators described fibers with a loss of cross-striations, hypereosinophilia, and a hyaline appearance. In contrast, the necrotic fibers observed in this study were defined as a mass of inflammatory cells and cellular debris, with breakdown of the plasma membrane. Previous investigators have also described macrophage (15) or phagocyte (14) invasion in muscle fibers, a finding that has been interpreted to indicate regeneration (15) or healing (14). The comparatively rare incidence of necrosis in our study may be related to the difference in subject characteristics. In the reports by Silver and Smith (15) and Kariks (14), subjects experienced acute respiratory distress and eventually died, whereas in our study, subjects did not experience acute respiratory distress and biopsies were obtained during surgery.

Our study is the first to report lipofuscin in diaphragm muscle. Deposits of lipofuscin pigmentation have been described in human vastus lateralis 3 d posteccentric exercise, a finding that was interpreted to indicate acute lysosomal protein degradation (30). Lipofuscin pigments are residual bodies that result from lysosomal autolysis of lipids (31). Lipofuscin is sometimes referred to as an "age pigment," as it is most common in the long-lived cells of the nervous system (31). An increased presence of lipofuscin in skeletal muscle may be reflective of an increased lysosomal activity over the lifetime of the involved fiber, but whether this relates primarily to age or to overuse of a muscle fiber is unknown.

Small angular fibers described in our study may represent fibers that have atrophied due to malnutrition, disuse, or steroid myopathy (26). Similar atrophic fibers were described in previous reports (12, 14). Because our study did not control for nutrition, past history of mechanical ventilation, or corticosteroid ingestion, the relevance of this feature in our study is uncertain.

The results of our study contrast those of Hards and coworkers (11), who found no relationship between the incidence of accessory muscle abnormalities and pulmonary function measures. This difference may be attributed to the possibility that more injury is present in the diaphragm than in the accessory muscles of respiration, a finding that has been noted in animal models of resistive ventilatory loading (7) and in individuals with heart failure (25). Indeed, the external and intercostal muscles are believed to be mainly postural muscles that are inactive or minimally active during resting or stimulated breathing (13). The parasternal intercostal muscles are more active (13) but were not sampled by Hards and coworkers (11). Our findings may also have differed because a broader range of subjects was evaluated. Specifically, the inclusion of subjects with severe airflow obstruction may have improved the power of our correlation analysis. Lastly, our study evaluated "abnormal muscle" as a ratio variable (A_A) rather than as an ordinal variable (absence or presence). Describing injury along a continuum of severity rather than as present or absent may have been a more powerful way to show an association with varying levels of airflow obstruction.

Our study showed the presence of macrophages in the diaphragm, but their number was not related to the degree of airflow obstruction (Figure 4). This relationship may be more complex than this analysis allowed, given the multiple roles and complex interactions of macrophages within the body. Further analysis of distinct subpopulations of macrophages may show the evolution of their expression with increased airflow obstruction, similar to the distinct subpopulations found in rat limb muscle after acute muscle injury (32). Identification of distinct subsets of macrophages may show a different time course of resident macrophages versus macrophages that may have been attracted to the site of injury as circulating monocytes.

There are some limitations of this study to consider when reflecting on the clinical importance of the data. First, airflow limitation as indicated by the FEV₁ is only one measure that assesses increased loading in COPD. Other measures of pulmonary function such as hyperinflation or respiratory muscle strength and endurance function are worthy of study and might be associated with diaphragm injury. Second, this is a relatively small sample of subjects who all underwent thoracotomy for cancer and other lung disease. It is possible that other systemic factors besides airflow limitation may contribute to the morphological abnormalities of the diaphragm such as corticosteroids, paraneoplastic disease, or other medical conditions. Lastly, our assessment of diaphragm biopsies in this sample of individuals does not provide any temporal description of histological characteristics relative to episodes of increased loading and subsequent injury, adaptation, regeneration, and/or repair. Therefore, results should be interpreted with some caution until more extensive studies have been done to examine for potential contributing factors to and other clinical factors associated with diaphragm injury.

This study showed that as severity of airflow obstruction increases, the proportion of abnormal diaphragm increases and the proportion of normal diaphragm decreases. The prevalence of macrophages in the diaphragm was not related to the severity of airflow obstruction, and may have been affected by many uncontrolled factors. An awareness that diaphragm injury may in part contribute to diaphragm dysfunction will lead to development of more effective therapeutic interventions to promote its function.