INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the normal human lung there is a network of elastic fibers that form discrete longitudinal bundles (LB) in the submucosa of the conducting airways (1, 2). Their macroscopic, microscopic, and ultrastructural appearances have been described by Monkhouse and Whimster (2). They are anatomically distinct from the subepithelial collagen layer that lies immediately beneath the epithelial basement membrane. This layer is thickened in asthma (3, 4) and contributes, with increases in bronchial smooth muscle and other structures (5), to the airway wall remodeling characteristic of the disease. Airway wall remodeling is considered to have physiologic effects, including decreased distensibility of the airway wall (6), exaggerated narrowing of the airway lumen when smooth muscle shortens (7), and/or irreversible airflow obstruction (8).

Lambert and colleagues (9) have suggested that increased stiffness of the inner part of the airway might attenuate airway narrowing. Using mathematical models and morphometric data from noncartilagenous sheep airways they found that the ability of the airway to resist deformation was a function of the thickness of the submucosa and the number of folds that formed when smooth muscle shortened. We propose that the longitudinal bundles of fibroelastic tissue described in this study may act as anatomic sites of airway mucosal folding, and variations in their number or size may alter the mechanical load opposing deformation of the mucous membrane.

No quantitative studies of longitudinal airway structures have previously been undertaken in subjects with asthma. Other investigators have reported increased elastic tissue in the submucosa of asthmatic airways (10, 11). Unfortunately, these studies did not provide sufficient anatomic detail to determine whether the changes involved the LB or some other structure. Therefore, the aim of our present study was to compare the area and elastic and collagen content of the airway LB and assess the relationship between these bundles and the number of folds in the airway mucosa in patients dying of asthma, patients with asthma dying of nonrespiratory causes, and nonasthmatic subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cases

The left lung was obtained from persons dying of acute asthma (n = 12), persons with asthma dying of other causes (n = 12), and persons dying of nonrespiratory causes with no history of asthma (n = 12). All cases were obtained through the Prairie Provinces Fatal Asthma Study. This study was initiated in 1992, as a prospective study of asthma deaths in the three provinces of Canada (Alberta, Saskatchewan, and Manitoba) with the objective of determining the cause(s) for the high rate of death from asthma in this region (12). The morphologic study described here was initiated when specimens from six men and six women from each group had been obtained. The lungs were fixed with 2.5% gluteraldehyde using a combination of airway instillation and vascular perfusion at a pressure of 20 cm H₂O. Information regarding a history of asthma, asthma symptoms and severity, use of asthma medications, smoking, hospitalization because of asthma, and basic demographic information were obtained from a questionnaire administered to the next of kin, medical examiner's notes, and pharmacy records. From the information obtained, cases were categorized as (1) control: no history of asthma, wheeze, use of asthma medications or other lung disease; (2) nonfatal asthma: died suddenly of nonrespiratory causes but subsequently found to have a definite history of asthma from relatives or medical examiner's files; (3) fatal asthma: when asthma was given as the cause of death at autopsy, when the report of events prior to death was consistent with asthma, when cardiovascular and cerebral causes of death were excluded as a cause of death, and when a history of asthma (as above) was obtained from the records. Cases were excluded from the study if they did not meet the criteria set out above.

Sampling

A stratified sampling procedure was used to sample an identical airway path to the posterior basal segment of the left lower lobe. Transverse sections of airways were obtained at nine equally spaced intervals along the axial path of the segment from the lobar bronchus at the hilus to the pleural surface. Thus, similar generations of airways were sampled for each case.

Histologic Preparation

Sections 5 µm thick of transverse airways from the posterior basal segment were stained with a modified Verhoeff-Masson-trichrome stain to identify elastic fibers and collagen within the airway wall. Selected contiguous sections 5 µm thick were stained using an immunoperoxidase technique for smooth muscle actin (Sigma Biosciences, St. Louis, MO) to determine the actin content of the LB, to identify myofibroblasts, and to differentiate between airway smooth muscle and the elastic bundles when they were in close proximity to one another.

Airway Dimensions

On all airways cut in transverse section (defined as an even thickness of epithelium, and even thickness from the basement membrane to the smooth muscle layer) and free from branching, measurements of longitudinal elastic bundle area and the number of mucosal folds were counted using a microscope, sidearm, and digitizer. Using a calibrated eyepiece grid, airway perimeter and area of the lumen were measured using linear intersect and point-counting techniques.

Longitudinal Bundles

The area and perimeter of all well-defined aggregates of positively stained elastic fibers (bundles) situated in the submucosa and lying between the epithelial basement membrane and the smooth muscle layer were measured. Perichondrial elastic bundles associated with cartilage plates in the airway adventitia and elastic fibers associated with smooth muscle bundles were not measured. The amount of positive staining for elastic tissue, collagen, and myofibroblasts for each bundle was scored semiquantitatively (0 to 5) on the Verhoeff-Masson-trichrome stained slides by one investigator (FHYG), with a score of zero representing no positive staining and a score of 5 representing 100% positive staining.

Mucosal Folds

Mucosal folds were defined as points along the basement membrane where there was a departure from the linear contour of the mucosal surface greater than that caused by variations in epithelial height. We have previously validated this method by measuring the perpendicular distance from the outside border of the airway smooth muscle layer to the basement membrane at more than 150 equidistant points around the circumference of the basement membrane. These distances (y-axis) were then plotted on a linear scale (x-axis) and fitted with a smooth curve spline fit to reproduce the symmetry of the basement membrane (13). A fold was defined as a discrete increase in the linear curvature of the basement membrane or epithelial surface at a given point on the circumference of the basement membrane. Comparison of this method with observations based on epithelial height (as above) showed very close agreement for the number of mucosal folds per airway, suggesting that counting of mucosal folds by eye is as reliable as quantitating them in a more objective manner. The degree of airway constriction was calculated by expressing the ratio (as a percentage) of the actual measured lumenal area to the theoretically expanded lumenal area (i.e., the lumenal area in a perfect circle).

Interobserver and intraobserver variation for each measurement was calculated by having two observers (blinded to the case group) measure the same five airways five times each.

Data Analysis

To compare airways of similar generation but from different persons, airways were grouped by the order in which they were obtained, that is, at equal intervals from the lobar bronchus to the pleura. This resulted in the lobar bronchus being labeled as airway size Group 9, and the most peripheral airway section obtained being airway size Group 1. Airways were then further grouped into airway sizes to represent mainly cartilaginous airways (levels 5 to 9) and mainly membranous airways (levels 1 to 4). For some analyses, the measured area of bundles was normalized for each airway by dividing the bundle area by the area of the expanded lumen. The mean value for each case in each airway size group was used. Differences for mean bundle area and number of mucosal folds between the case groups (control, nonfatal, and fatal asthma) and between the smoking and nonsmoking groups, for different airway size groups, were tested using a one-way analysis of variance (ANOVA). Where significant differences were detected between groups, Fisher's post-hoc comparison was performed. Differences for the mean percentage of elastic, collagen, or myofibroblast content were analyzed using Spearman's rank correlation. Differences between the number of mucosal folds relative to the size of the airway and the amount of muscle shortening, and between the number of mucosal folds and the number and area of longitudinal elastic bundles, were analyzed using linear regression analysis. Multiple linear and logistic regression was used to examine the effects of case-group, sex, age, and smoking on bundle area. A probability value of less than 5% was considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject characteristics are shown in Table 1. There were 6 males and 6 females in each case group. The control group was slightly older and contained more smokers than the asthma groups. There were no significant differences in airway perimeters for any sampling site between the three groups. The LB were readily identified on Verhoeff-Masson-trichrome stained sections in all three groups by their characteristic location within the submucosa in close association with the mucosal folds (Figure 1, top panel ). The elastic fibers within the bundles were oriented along the longitudinal axis of the airways and were located midway between the subepithelial collagenous reticular layer and the bronchial smooth muscle (Figure 1, middle panel ). Immunostains revealed that smooth muscle actin was an integral component of the bundle (Figure 1, bottom panel ), being located in the cytoplasm of cells that had the appearance of myofibroblasts.

View larger version (117K): [in this window] [in a new window]   View larger version (144K): [in this window] [in a new window]   Figure 1.   (Top panel ) Verhoeff-Masson-trichrome-stained section of airway from a case of fatal asthma. This medium-powered view shows three longitudinal bundles in the centers of three mucosal folds. There is thickening of the subepithelial collagen layer and smooth muscle hypertrophy/hyperplasia of the airway wall, both characteristic features of asthma. (Middle panel ) High-powered view of one bundle seen at left in top panel showing the longitudinal orientation of the individual elastic fibers. The longitudinal bundle (LB) is anatomically distinct from the thickened subepithelial collagen layer (arrowheads) and bronchial smooth muscle. (Bottom panel ) Section from a case of fatal asthma stained for smooth muscle actin. Staining of myofibroblastlike cells with a longitudinal orientation is seen. These cells appear to be an integral component of the longitudinal bundles. Contrast the myofibroblasts in the longitudinal bundle with the obliquely orientated bronchial smooth muscle fibers seen at the bottom.

The number of LB tended to increase in cases of asthma and in relation to severity, but this was not significantly different between case groups (Table 2). The absolute values for the total LB area (mm²) in the cartilaginous airways for the three groups subdivided by smoking status are shown in Figure 2. Asthma and smoking had significant (p < 0.05) positive relationships with bundle area. Analysis of the normalized data (bundle area [mm²]/expanded luminal area [mm²]) showed that LB area was greater (p < 0.05) in the large airways (Table 3) in cases of fatal asthma (0.058 ± 0.043) compared with cases of nonfatal asthma (0.036 ± 0.032) and control cases (0.031 ± 0.028). In the small airways (Table 3) the LB area was greater in cases of fatal asthma (0.037 ± 0.038) than in cases of nonfatal asthma (0.031 ± 0.033) and control cases (0.022 ± 0.034), but this difference was not statistically significant (p = 0.076). The effect of smoking on LB area in large and small airways is shown in Table 4. Because most of the control cases were smokers, the effect of smoking was difficult to assess in this group. Analysis of variance showed a significant (p = 0.04) overall effect of smoking on LB area in cases of asthma. This was largely due to the contribution from the cases of nonfatal asthma. The mean bundle area was significantly greater for men than for women (p = 0.026) and increased with increasing age regardless of sex (p = 0.028). The proportion of the bundle areas that stained for elastin was significantly greater (p < 0.05) in control cases (57%) than in nonfatal (38%) and fatal (33%) asthma cases. The proportion of the bundle areas occupied by myofibroblasts was significantly (p < 0.05) greater in cases of fatal (24%) and nonfatal (23%) asthma than in control cases (7%), as assessed by increased staining for smooth muscle actin in the asthma groups. The proportion of collagen staining was similar in each group.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Longitudinal bundle area (mm2) for the cartilaginous airways by case group and smoking status. Positive significant (p < 0.05) effects on bundle area are seen for asthma and smoking status (diamond indicates significant difference from nonsmokers; asterisks indicate significant difference from nonasthma control group).

The number of mucosal folds was similar in both airway size groups in cases of fatal asthma, cases of nonfatal asthma, and control cases (Table 5). There was a positive correlation between the number of mucosal folds and the number of LB in the cases of fatal asthma (r = 0.79, p < 0.05), cases of nonfatal asthma (r = 0.69, p < 0.05), and in the control group (r = 0.55, p < 0.05).

The amount of airway constriction, defined by the ratio of the measured lumenal area as a percentage of the relaxed lumenal area, is shown in Figure 3. All groups showed evidence of airway constriction of varying degree. The airway narrowing in the fatal asthma group was significantly greater (p < 0.05) than in the other groups for small airways and bronchi, but not for the larger cartilaginous airways.

View larger version (41K): [in this window] [in a new window]   Figure 3.   The amount of airway constriction expressed as the ratio of the measured lumenal area to the calculated relaxed lumenal area (%) in cases of fatal asthma, cases of nonfatal asthma, and control cases. The airways have been grouped by airway size. Some degree of airway constriction is seen for all groups. The airway constriction for the fatal asthma group was significantly greater (p < 0.05) than for the other groups for small- and medium-sized airways (diamonds indicate significant difference from nonasthma control and nonfatal asthma groups).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study the LB area in the submucosa of small and large airways was greater in cases of fatal asthma than in cases of nonfatal asthma and control cases, although the number of bundles was similar in each group. The elastin content of the LB was proportionally decreased in cases of asthma compared with cases without asthma, and there was a corresponding increase in myofibroblasts in the bundles in cases of asthma. The LB area was greater in smokers than in nonsmokers in cases of asthma, especially nonfatal asthma. The LB area was also greater in men than in women and increased with increasing age in both sexes, independent of case group and smoking status. The number of mucosal folds was similar in each of the case groups and there was a positive correlation between the number of LB and the number of mucosal folds.

The longitudinal elastic fiber network of the bronchial tree has previously been described (1, 2), although, as far as we are aware, it has not been systematically quantified. In this study we measured only discrete bundles situated in the submucosa, and only where there was a definitive homogenous structure stained positively for elastic tissue, myofibroblasts, and collagen. The LB were easily distinguished from the subepithelial collagenous reticular layer, which contains no elastic fibers, and from the bronchial smooth muscle bundles. We are confident that the increases in bundle area observed in this study were due to hypertrophy and/or hyperplasia of these distinct anatomic structures and was not due to an overall increase in nonspecific connective tissue elements as has been described (10, 11). In the more distal airways the bundles were less discrete and tended to form a more diffuse meshwork of subepithelial elastic fibers; nonetheless, focal concentrations of elastic fibers were identified and quantified.

The enlargement of the LB constitutes an additional tissue component contributing to thickening of the airway wall in asthma (3, 7). These findings suggest that airway inflammation may stimulate tissue hypertrophy, possibly through increased production of growth factors. Thus, growth factors liberated during the repair phase of the inflammatory response may have a general effect on mesenchymal as well as epithelial components in the airway wall.

The increase in bundle area in this study was also related to age, sex, and cigarette smoking. The effect of age on bundle area may be due to exposure to cigarette smoke and/or other irritants, or it may be due to a longer duration of asthma, whereas the effect of sex on bundle area may be due to increased lung/airway size since we did not correct for body size. The effect of cigarette smoking on bundle area was observed in all three groups, but it was only statistically significant in the nonfatal cases. Possible explanations for the lack of the effect of smoking in the cases of fatal asthma include masking of its effects by more marked airway wall inflammation and remodeling (5) and a decreased cigarette intake in patients with more severe symptoms.

The increased LB area will contribute to the thickening of the inner airway wall (5) and increase the amount of lumenal narrowing that occurs when airway smooth muscle shortens (7, 14) and, as such, contribute to airway hyperresponsiveness. Other functional effects are likely to depend on the effect of changes in the composition of the bundles on the mechanical properties of the bundle and therefore on the mechanical properties of the airway wall (15).

The continuous longitudinal arrangement of the elastic and collagen fibers throughout the bronchial tree suggests that they may be important in elastic recoil of the airways after mechanical stretching during inspiration. The intimate relationship between elastic and collagen fibers in the airway submucosa is shown by the electron microscopic findings of Bousquet and colleagues (16). They showed that only three of 21 asthmatics examined had normal elastic fibers when compared with control subjects. A relative increase in the amount of collagen and myofibroblasts in the hypertrophied bundles from the asthmatics may result in decreased distensibility of the airway (6) or impaired longitudinal stretching and/or relaxation of the airways during the respiratory cycle. There are few other studies of the matrix content of the longitudinal bundles in asthma. In an autopsy study of a single case of fatal asthma in which the patient was a long-term smoker and had long-standing severe asthma with chronic cough and partially irreversible airway obstruction, Gabrielli and colleagues (11) reported elastic, collagen, and myofibroblast hyperplasia. They suggested that these changes alone may have resulted in increased stiffness of the airways and contributed to irreversible airflow obstruction.

Although the pathogenesis of the LB hypertrophy seen in cases of asthma may result from the release of mitogenic factors associated with airway inflammation, mechanical factors also deserve consideration. Hyperinflation of the lung and atelectasis occur during acute episodes of asthma (17). These would likely produce abnormal and asymmetric longitudinal forces on the airways. These forces may be the stimulus for myofibroblast proliferation similar to the proliferative effects of mechanical stretching on airway smooth muscle (18).

The LB observed in this study were invariably associated with a distinct protrusion or fold in the basement membrane. Lambert and colleagues (9) have suggested that folding of the inner part of the airway wall around areas of inhomogeneity will create a force resisting airway closure when smooth muscle shortens. He showed mathematically that an increase in the number of folds, for a given degree of smooth muscle shortening, will result in an increase in the resistance opposing muscle shortening and therefore lumenal narrowing. If this is the case then an alteration in the number of the bundles may alter the mechanics of airway narrowing after smooth muscle constriction. This has been a suggested role of large submucosal blood vessels in the sheep (19). However, these vessels are not seen in human airways (20). In the present study, the number of bundles and folds were the same in the three case groups. Although the cases of fatal asthma had the highest amount of airway constriction, no cases examined in this study had perfectly round airways, suggesting that all airways had some degree of muscle tone, which is necessary to count mucosal folds. Therefore the number of LB is unlikely to be an important determining influence on the degree of lumenal narrowing in patients with asthma compared with nonasthmatics. The results of the present study suggest that changes in the structural proteins of the airway wall in asthma, resulting from chronic airway inflammation and/or abnormal stresses, will affect the longitudinal as well as the transverse mechanical properties of the airways.