Cellularity of the Alveolar Walls in Smokers and Its Relation to Alveolar Destruction Functional Implications 1-3
DAVID EIDELMAN,4 MARINA R SAETTA, HEBERTO GHEZZO, NAI-SAN WANG, JOHN R. HOIDAL, MALCOLM KING, and MANUEL G. COSIO
Introduction
Cigarette smoking is considered to be the major risk factor in the development of pulmonary emphysema (1-4); however, the mechanisms by which cigarette smoking causes emphysema are largely unknown. A widely held hypothesis proposes that the lung destruction occurring in cigarette smokers is mediated by release of elastolytic enzymes/from pulmonary inflammatory cells. However, a direct relationship between lung parenchymal inflammation and destruction has not been shown. Furthermore, as indicated by Janoff (5), we still need to know which cell is the predominant source of elastolytic proteins in smokers' lungs, the neutrophil or the macrophage. In 1961, Anderson and Foraker (6) observed that the earliest pathologic abnormality in centrilobular emphysema (CLE) was hypercellularity within the alveolar walls. This observation has not been extended. Rather, most recent studies in smokers have focused on inflammatory changes in the walls of the bronchioles (7-9), within air spaces, or in bronchoalveolar lavage fluid (10). ' We reasoned that the inflammatory cells within the alveolar walls, because of their proximity to the site of injury, are the ones that have the greatest potential to playa role in the pathogenesis of lung destruction. Alveolar septal damage can now be quantitated microscopically around the small airways and in the parenchyma with any of the methods recently described by us (11-13). These indices of destruction can separate smokers from nonsmokers at a time when there is no other evidence of emphysema. This study was undertaken to investigate the possible relationship of the cellularity in the alveolar walls of cigarette smokers, particularly neutrophils, to parenchymal destruction and to assess the effect of the
SUMMARY Inflammatory cells are believed to play an important role in the pathogenesis of emphysema; however, a relationship between presence of cells in the lung parenchyma and its destruction has never been shown. The aim of this study was to quantitate alveolar septal cellularity in smokers' lungs and to investigate its relationship with parenchymal destruction and lung function. The lungs of 23 smokers (SS) undergoing thoracotomy for localized pulmonary lesions were compared with those of eight nonsmokers (NS) and five smokers (AS) who died suddenly of non respiratory causes. Pulmonary function tests were performed within 1 wk of surgery in SS. For each subject, we quantitated alveolar wall cells (CEllS), an index of alveolar wall destruction (01), and the mean linear intercept (lm). As no significant differences were found between S and AS with regard to these indices, we combined them (Group S) for comparison with NS. Although Lm was not significantly different between Sand NS, (0.331 ± 0.072 versus 0.288 ± 0.038), CELLS and 01 were higher in S than in NS (48 ± 8 versus 25 ± 2 cells/mm, p < 0.001;47 ± 20 versus 17 ± 5, P < 0.001,respectively). Further, CELLS and 01 were significantly correlated (r = 0.799, P < 0.001). The number of polymorphonuclear cells (PMN) in S can exceed that in NS by as much as 5-fold; however, PMN were inversely correlated with parenchymal destruction (01) (r = -0.598, P < 0.01). Thus, smokers' lungs have alveolar septal hypercellularity, possibly inflammatory, and closely related to destruction involving cells other than the PMN. AM REV RESPIR DIS 1990; 141:1547-1552
alveolar damage on the loss of function of the lung. Methods The population studied consisted of three groups: nonsmokers (NS), autopsy smokers (AS), and surgical smokers (SS). For the first two groups, lungs were obtained at autopsy from subjects who died suddenly outside of hospital of nonrespiratory causes. The smoking history was obtained from next of kin. Pulmonary function was not assessed during life in these groups. For the surgical smoker group, lungs (8 cases) or lobes (15 cases: 7 upper and 8 lower lobes) wereobtained from 23 smokers undergoing thoracotomy for localized malignant nonmetastatic pulmonarylesions not involving the remaining lung parenchyma. All patients had a smoking history of at least 35 yr. None of the patients had interstitial lung disease, dust exposure, asthma, or heart disease. In order to investigate possible differences in the measurement of parenchymal cellularity between surgical and postmortem specimens and the possible effects of surgery, five lungs with mean linear intercept (Lm) values comparable to those of the surgical smokers wereselected from smokers of similar age who died suddenly outside
the hospital and who had no evidenceofpneumonia. According to close relatives, each of the autopsy smokers had smoked cigarettes regularly for many years.
Pulmonary Function rests Pulmonary function tests were performed in smokers within 1 wk of surgery according to the standard NIH protocol (14). Values were expressed as percent predicted based on the prediction formulas of Morris and coworkers (15). A volume-displacement, pressurecompensated, body plethysmograph was used (Received in originalform August 30, /989 and in revised form November 27, /989) I From the Respiratory Division of the Royal Victoria Hospital, the Montreal Chest Hospital, the Meakins-Christie Laboratories, and the Department of Pathology, McGill University, Montreal, Quebec, Canada. 1 Supported by Grant No. MA-6743 from the Medical Research Council of Canada. 3 Correspondence and requests for reprints should be addressed to Manuel G. Cosio, M.D., Respiratory Division, L4.09, Royal Victoria Hospital, 687 Pine Avenue West, Montreal, Quebec, H3A IAI, Canada. 4 Fellow of the Medical Research Council of Canada.
1547
1548
EIDELMAN, SAETTA, GHEZZO, WANG, HOIDAL. KING, AND COSIO
to measure FRC and TLC. At least three reproducible static deflation pressure-volume (P-V) curves were obtained for each patient, and transpulmonary pressure was measured with a differential pressure transducer using the esophageal balloon technique as described by Milic-Emili and coworkers (16). All P-V points above FRC were fitted to a single exponential of the form V = A - Be-K P , as described by Gibson and colleagues (17).Values for the transpulmonary pressure at 90070 of measured TLC (PL90 ) were derived from the fitted parameters in order to avoid observer bias.
Morphologic Exam ination All specimens were fixed with lOOJo formalin by intrabronchial infusion at a constant pressure of 25 cm H 20 for at least 24 h. Five randomly selected blocks of tissue (template size, 2 x 2.5 ern)wereobtained from the subpleural slice from each lobe. Sections 6-Jlm thick were stained with periodic acid-Schiff and hematoxylin-phloxine-saffron (PAS-HPS) (18) for light microscopy. Only sections microscopically free of tumor or cutting artifact were included in the study.
Cellularity of Alveolar Walls At a magnification x 400, the length of the alveolar wall visualized in the center of the microscopic field was measured with a digitizing tablet (Hewlett-Packard 9111A; HewlettPackard, Waltham, MA) and a microcomputer (Hewlett-Packard 85). Each field was projected with a camera lucida onto the digitizing tablet, and the portion to be measured was traced and corrected for shrinkage. The total number of nuclei, taken as representative of cells lying within this alveolar wall, was counted (figure 1).Twenty fields randomly distributed across the slide were studied. The result was expressed as the number of cells/mm of al veolar wall, and the average value for each subject was computed. Polymorphonuclear leukocytes (PMN) were identified mainly by their size and nuclear shape, and to a lesser extent by the presence of cytoplasmic granules, and expressed as a percent of the total number of cells. No attempt to differentiate the other cells any further was made since by light microscopy it is difficult to accurately differentiate all the cell types within these groups. All slideswerecoded, randomized, and then evaluated by a single observer who did not have access to the code. In order to test the reproducibility of the measurements, 10slides were randomly selected for reading by a second observer and by the main observer 3 months later. Both the intraobserver and interobserver reproducibility were acceptable (r = 0.90, p < 0.001 and r = 0.83, p < 0.005, respectively).
Fig. 1. Photomicrographs of alveolar walls . A. Nonsmoker's lung . B. Smoker's lung. Stain, HPS-PAS; original magnification: x400.
a grid with 42 points was superimposed on the microscopic field, and the air spaces directly under these points were classified as normal alveolar or duct spaces (N) when they were surrounded by intact walls, and destroyed alveolar or duct spaces (0) when they were surrounded by walls with at least two breaks . Twenty fields per slide were examined, and the 01 was computed from the formula DI = 0 /(0 + N) x 100. For each subject the final value for the DI was the average value of all the slides.
Destructive Index
Mean Linear Intercept
Assessment of the destructive index (01), which quantifies the percent of destroyed air spaces in the lung parenchyma, was made as described (12). Briefly,at a magnification x 63
Measurement of Lm was made using a microscope with a x 10 objective and x lO eyepiece by the modified method of Thurlbeck (19). The size of the section was determined
by digitizing its area; correction for tissue shrinkage was made with Thurlbeck's formula to obtain the final value of Lm .
Data Analysis Group data are expressed as mean ± SO. The two-tailed t test for paired and unpaired data, as appropriate, was used for comparison of means. Regression analysis was carried out using the least squares method . Stati stical significance was accepted at the 5070 level of confidence. Results
The sex, age, values of FEV .. PL9 0 , DI, and Lm, cellularity in the alveolar walls, the PMN as a percentage of the total cell
1549
ALVEOLAR WALL CELLULARITY: MORPHOLOGY AND FUNCTION TABLE 1 PULMONARY FUNCTION AND MORPHOLOGIC INDICES IN SMOKERS Subject No.
Age (yr)
FEV, (% pred)
PL go (cm H2O)
DI (%)
Lm (mm)
Cells (mm- 1)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Mean SD Total number
65 53 57 63 60 64 70 61 69 70 68 72 77 66 54 74 58 66 45 56 64 67 64 63.6 7.4 23
72.2 86.4 80.0 77.2 72.8 40.0 79.1 86.0 113.2 95.2 94.0 78.1 17.4 79.7 65.8 100.8 82.5 105.0 88.0 94.0 91.0 44.0 66.0 78.6 21.7 23
12.3 11.8 5.8
40.9 41.8 63.6 35.3 60.7 57.9 42.0 43.2 52.1 42.2 33.8 52.7 76.4 56.1 69.8 62.7 33.4 30.1 29.6 29.8 9.8 93.1 72.4 49.1 19.0 23
0.318 0.243 0.351 0.292 0.322 0.420 0.328 0.281 0.424 0.257 0.328 0.342 0.540 0.261 0.468 0.406 0.325 0.259 0.233 0.309 0.296 0.386 0.238 0.332 0.078 23
42.0 42.7 60.8 51.7 53.2 58.2 35.9 37.6 43.9 40.9 42.1 50.8 63.4 37.8 49.6 46.7 37.1 42.8 51.4 47.2 43.2 56.3 60.9 47.7 8.2 23
13.8
9.6 11.3 9.0 7.2 5.4 11.0
12.9 10.1 14.8 16.7 16.1 2.2 5.5 10.3 4.1 17
PMN (mm- 1)
PMN (%)
0.67
1.60
2.55 2.69 0.45 1.28
4.20 5.20 0.85 2.20
1.43
3.80
0.98 2.02 0.97 1.27 2.23 0.89
2.40 4.80 1.90 2.00 5.90 1.80
3.49 1.88 2.78 0.99 2.64 0.78 0.26 1.59 0.92 19
9.40 4.40 5.40 2.10 6.10 1.38 0.40 3.47 2.29 19
The percentage of PMN in the alveolar wall of smokers ranged between 0.4 and 9.4%, and the total number per millimeter ranged between 0.3 and 3.5 PMN/ mm alveolar wall (tables 1 and 2). Among smokers, both the percentage of PMN and the absolute number of PMN/mm were inversely correlated with DI (r = - 0.598, p < 0.01 and r = -0.568, p < 0.01, respectively) (figure 4). PMN/mm decreased with increasing cellularity, but this did not reach statistical significance (r = -0.291). Among surgical smokers, alveolar eellularity and DI were both negatively correlated with the index of elastic recoil PLgo (r = - 0.558, p < 0.05 and r = - 0.823, P < 0.001, respectively) (figure 5). The correlation between Lm and PLgo did not reach statistical significance (r = - 0.458). All three morphometric indices were negatively correlated with FEY 1 (cells/mm, r = -0.641, p < 0.001; DI, r = -0.623, p
DAVID EIDELMAN,4 MARINA R SAETTA, HEBERTO GHEZZO, NAI-SAN WANG, JOHN R. HOIDAL, MALCOLM KING, and MANUEL G. COSIO
Introduction
Cigarette smoking is considered to be the major risk factor in the development of pulmonary emphysema (1-4); however, the mechanisms by which cigarette smoking causes emphysema are largely unknown. A widely held hypothesis proposes that the lung destruction occurring in cigarette smokers is mediated by release of elastolytic enzymes/from pulmonary inflammatory cells. However, a direct relationship between lung parenchymal inflammation and destruction has not been shown. Furthermore, as indicated by Janoff (5), we still need to know which cell is the predominant source of elastolytic proteins in smokers' lungs, the neutrophil or the macrophage. In 1961, Anderson and Foraker (6) observed that the earliest pathologic abnormality in centrilobular emphysema (CLE) was hypercellularity within the alveolar walls. This observation has not been extended. Rather, most recent studies in smokers have focused on inflammatory changes in the walls of the bronchioles (7-9), within air spaces, or in bronchoalveolar lavage fluid (10). ' We reasoned that the inflammatory cells within the alveolar walls, because of their proximity to the site of injury, are the ones that have the greatest potential to playa role in the pathogenesis of lung destruction. Alveolar septal damage can now be quantitated microscopically around the small airways and in the parenchyma with any of the methods recently described by us (11-13). These indices of destruction can separate smokers from nonsmokers at a time when there is no other evidence of emphysema. This study was undertaken to investigate the possible relationship of the cellularity in the alveolar walls of cigarette smokers, particularly neutrophils, to parenchymal destruction and to assess the effect of the
SUMMARY Inflammatory cells are believed to play an important role in the pathogenesis of emphysema; however, a relationship between presence of cells in the lung parenchyma and its destruction has never been shown. The aim of this study was to quantitate alveolar septal cellularity in smokers' lungs and to investigate its relationship with parenchymal destruction and lung function. The lungs of 23 smokers (SS) undergoing thoracotomy for localized pulmonary lesions were compared with those of eight nonsmokers (NS) and five smokers (AS) who died suddenly of non respiratory causes. Pulmonary function tests were performed within 1 wk of surgery in SS. For each subject, we quantitated alveolar wall cells (CEllS), an index of alveolar wall destruction (01), and the mean linear intercept (lm). As no significant differences were found between S and AS with regard to these indices, we combined them (Group S) for comparison with NS. Although Lm was not significantly different between Sand NS, (0.331 ± 0.072 versus 0.288 ± 0.038), CELLS and 01 were higher in S than in NS (48 ± 8 versus 25 ± 2 cells/mm, p < 0.001;47 ± 20 versus 17 ± 5, P < 0.001,respectively). Further, CELLS and 01 were significantly correlated (r = 0.799, P < 0.001). The number of polymorphonuclear cells (PMN) in S can exceed that in NS by as much as 5-fold; however, PMN were inversely correlated with parenchymal destruction (01) (r = -0.598, P < 0.01). Thus, smokers' lungs have alveolar septal hypercellularity, possibly inflammatory, and closely related to destruction involving cells other than the PMN. AM REV RESPIR DIS 1990; 141:1547-1552
alveolar damage on the loss of function of the lung. Methods The population studied consisted of three groups: nonsmokers (NS), autopsy smokers (AS), and surgical smokers (SS). For the first two groups, lungs were obtained at autopsy from subjects who died suddenly outside of hospital of nonrespiratory causes. The smoking history was obtained from next of kin. Pulmonary function was not assessed during life in these groups. For the surgical smoker group, lungs (8 cases) or lobes (15 cases: 7 upper and 8 lower lobes) wereobtained from 23 smokers undergoing thoracotomy for localized malignant nonmetastatic pulmonarylesions not involving the remaining lung parenchyma. All patients had a smoking history of at least 35 yr. None of the patients had interstitial lung disease, dust exposure, asthma, or heart disease. In order to investigate possible differences in the measurement of parenchymal cellularity between surgical and postmortem specimens and the possible effects of surgery, five lungs with mean linear intercept (Lm) values comparable to those of the surgical smokers wereselected from smokers of similar age who died suddenly outside
the hospital and who had no evidenceofpneumonia. According to close relatives, each of the autopsy smokers had smoked cigarettes regularly for many years.
Pulmonary Function rests Pulmonary function tests were performed in smokers within 1 wk of surgery according to the standard NIH protocol (14). Values were expressed as percent predicted based on the prediction formulas of Morris and coworkers (15). A volume-displacement, pressurecompensated, body plethysmograph was used (Received in originalform August 30, /989 and in revised form November 27, /989) I From the Respiratory Division of the Royal Victoria Hospital, the Montreal Chest Hospital, the Meakins-Christie Laboratories, and the Department of Pathology, McGill University, Montreal, Quebec, Canada. 1 Supported by Grant No. MA-6743 from the Medical Research Council of Canada. 3 Correspondence and requests for reprints should be addressed to Manuel G. Cosio, M.D., Respiratory Division, L4.09, Royal Victoria Hospital, 687 Pine Avenue West, Montreal, Quebec, H3A IAI, Canada. 4 Fellow of the Medical Research Council of Canada.
1547
1548
EIDELMAN, SAETTA, GHEZZO, WANG, HOIDAL. KING, AND COSIO
to measure FRC and TLC. At least three reproducible static deflation pressure-volume (P-V) curves were obtained for each patient, and transpulmonary pressure was measured with a differential pressure transducer using the esophageal balloon technique as described by Milic-Emili and coworkers (16). All P-V points above FRC were fitted to a single exponential of the form V = A - Be-K P , as described by Gibson and colleagues (17).Values for the transpulmonary pressure at 90070 of measured TLC (PL90 ) were derived from the fitted parameters in order to avoid observer bias.
Morphologic Exam ination All specimens were fixed with lOOJo formalin by intrabronchial infusion at a constant pressure of 25 cm H 20 for at least 24 h. Five randomly selected blocks of tissue (template size, 2 x 2.5 ern)wereobtained from the subpleural slice from each lobe. Sections 6-Jlm thick were stained with periodic acid-Schiff and hematoxylin-phloxine-saffron (PAS-HPS) (18) for light microscopy. Only sections microscopically free of tumor or cutting artifact were included in the study.
Cellularity of Alveolar Walls At a magnification x 400, the length of the alveolar wall visualized in the center of the microscopic field was measured with a digitizing tablet (Hewlett-Packard 9111A; HewlettPackard, Waltham, MA) and a microcomputer (Hewlett-Packard 85). Each field was projected with a camera lucida onto the digitizing tablet, and the portion to be measured was traced and corrected for shrinkage. The total number of nuclei, taken as representative of cells lying within this alveolar wall, was counted (figure 1).Twenty fields randomly distributed across the slide were studied. The result was expressed as the number of cells/mm of al veolar wall, and the average value for each subject was computed. Polymorphonuclear leukocytes (PMN) were identified mainly by their size and nuclear shape, and to a lesser extent by the presence of cytoplasmic granules, and expressed as a percent of the total number of cells. No attempt to differentiate the other cells any further was made since by light microscopy it is difficult to accurately differentiate all the cell types within these groups. All slideswerecoded, randomized, and then evaluated by a single observer who did not have access to the code. In order to test the reproducibility of the measurements, 10slides were randomly selected for reading by a second observer and by the main observer 3 months later. Both the intraobserver and interobserver reproducibility were acceptable (r = 0.90, p < 0.001 and r = 0.83, p < 0.005, respectively).
Fig. 1. Photomicrographs of alveolar walls . A. Nonsmoker's lung . B. Smoker's lung. Stain, HPS-PAS; original magnification: x400.
a grid with 42 points was superimposed on the microscopic field, and the air spaces directly under these points were classified as normal alveolar or duct spaces (N) when they were surrounded by intact walls, and destroyed alveolar or duct spaces (0) when they were surrounded by walls with at least two breaks . Twenty fields per slide were examined, and the 01 was computed from the formula DI = 0 /(0 + N) x 100. For each subject the final value for the DI was the average value of all the slides.
Destructive Index
Mean Linear Intercept
Assessment of the destructive index (01), which quantifies the percent of destroyed air spaces in the lung parenchyma, was made as described (12). Briefly,at a magnification x 63
Measurement of Lm was made using a microscope with a x 10 objective and x lO eyepiece by the modified method of Thurlbeck (19). The size of the section was determined
by digitizing its area; correction for tissue shrinkage was made with Thurlbeck's formula to obtain the final value of Lm .
Data Analysis Group data are expressed as mean ± SO. The two-tailed t test for paired and unpaired data, as appropriate, was used for comparison of means. Regression analysis was carried out using the least squares method . Stati stical significance was accepted at the 5070 level of confidence. Results
The sex, age, values of FEV .. PL9 0 , DI, and Lm, cellularity in the alveolar walls, the PMN as a percentage of the total cell
1549
ALVEOLAR WALL CELLULARITY: MORPHOLOGY AND FUNCTION TABLE 1 PULMONARY FUNCTION AND MORPHOLOGIC INDICES IN SMOKERS Subject No.
Age (yr)
FEV, (% pred)
PL go (cm H2O)
DI (%)
Lm (mm)
Cells (mm- 1)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Mean SD Total number
65 53 57 63 60 64 70 61 69 70 68 72 77 66 54 74 58 66 45 56 64 67 64 63.6 7.4 23
72.2 86.4 80.0 77.2 72.8 40.0 79.1 86.0 113.2 95.2 94.0 78.1 17.4 79.7 65.8 100.8 82.5 105.0 88.0 94.0 91.0 44.0 66.0 78.6 21.7 23
12.3 11.8 5.8
40.9 41.8 63.6 35.3 60.7 57.9 42.0 43.2 52.1 42.2 33.8 52.7 76.4 56.1 69.8 62.7 33.4 30.1 29.6 29.8 9.8 93.1 72.4 49.1 19.0 23
0.318 0.243 0.351 0.292 0.322 0.420 0.328 0.281 0.424 0.257 0.328 0.342 0.540 0.261 0.468 0.406 0.325 0.259 0.233 0.309 0.296 0.386 0.238 0.332 0.078 23
42.0 42.7 60.8 51.7 53.2 58.2 35.9 37.6 43.9 40.9 42.1 50.8 63.4 37.8 49.6 46.7 37.1 42.8 51.4 47.2 43.2 56.3 60.9 47.7 8.2 23
13.8
9.6 11.3 9.0 7.2 5.4 11.0
12.9 10.1 14.8 16.7 16.1 2.2 5.5 10.3 4.1 17
PMN (mm- 1)
PMN (%)
0.67
1.60
2.55 2.69 0.45 1.28
4.20 5.20 0.85 2.20
1.43
3.80
0.98 2.02 0.97 1.27 2.23 0.89
2.40 4.80 1.90 2.00 5.90 1.80
3.49 1.88 2.78 0.99 2.64 0.78 0.26 1.59 0.92 19
9.40 4.40 5.40 2.10 6.10 1.38 0.40 3.47 2.29 19
The percentage of PMN in the alveolar wall of smokers ranged between 0.4 and 9.4%, and the total number per millimeter ranged between 0.3 and 3.5 PMN/ mm alveolar wall (tables 1 and 2). Among smokers, both the percentage of PMN and the absolute number of PMN/mm were inversely correlated with DI (r = - 0.598, p < 0.01 and r = -0.568, p < 0.01, respectively) (figure 4). PMN/mm decreased with increasing cellularity, but this did not reach statistical significance (r = -0.291). Among surgical smokers, alveolar eellularity and DI were both negatively correlated with the index of elastic recoil PLgo (r = - 0.558, p < 0.05 and r = - 0.823, P < 0.001, respectively) (figure 5). The correlation between Lm and PLgo did not reach statistical significance (r = - 0.458). All three morphometric indices were negatively correlated with FEY 1 (cells/mm, r = -0.641, p < 0.001; DI, r = -0.623, p
