Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Emphysema is characterized by abnormal irreversible enlargement of the air spaces in the lung resulting from progressive degradation of the extracellular matrix of the alveolar walls (1, 2). Evidence acquired over the past 30 years has led to the idea that matrix degradation results from an excess of inflammatory cell-derived proteolytic activity and a relative paucity of antiproteolytic defense in the lower respiratory tract, a theory often referred to as the protease-antiprotease hypothesis. In cigarette smokers, the traditional theory states that the effector cell is the neutrophil (polymorphonuclear leukocyte [PMN]), the crucial protease neutrophil elastase, and that cigarette smoke also inactivates 1-antitrypsin (1AT), the major antiproteolytic substance in the lung parenchyma.

In patients with hereditary 1AT deficiency, this fundamental scenario appears to be correct, and some reports suggest that progression of emphysema may be ameliorated by administration of 1AT (3, 4). However, in cigarette smokers the issue of what cell, what protease(s), and what antiproteolytic substances are involved has become increasingly controversial. Attempts to show that 1AT levels in smokers are different from those in nonsmokers, or that the 1AT present has reduced function, have led to contradictory results: some studies show lower levels of 1AT, or a decreased association rate between 1AT and neutrophil elastase, or the presence of oxidized methionine residues in the protein leading to a loss of inhibitory function; whereas an equal number of studies have shown that 1AT levels in smokers are the same as those in nonsmokers and that there is no evidence of oxidative inactivation or other loss of function (reviewed in 2, 5).

The idea that the PMN is the effector cell has become equally controversial. Cigarette smoke causes an increase in both lavage and tissue PMNs and macrophages (MACs) (8). Although increased numbers of PMNs can be detected in the alveolar walls in smokers, and experimental instillation of neutrophil elastase produces emphysema in laboratory animals (10), the numbers of PMNs present in human tissue do not correlate with the degree of lung destruction, whereas correlations are obtained with numbers of MACs (9, 12). The traditional view that MAC proteases show little elastolytic activity compared with neutrophil elastase has been modified by the recognition that a variety of MAC-derived metalloproteases, including gelatinase A and B, matrilysin, and MAC metalloelastase, all can degrade elastin (13). It has also become apparent from human and experimental studies that emphysema is characterized not only by breakdown of elastin but also by breakdown and resynthesis of collagen with scar formation (2, 5, 16, 17).

These observations have led to an alternate formulation of the protease-antiprotease hypothesis in which MACs are the crucial cells and MAC-derived proteases, particularly metalloproteases, are the effector agents (14, 18). Sansores and colleagues (19) found that air space and interstitial MACs from smoke-exposed guinea pigs showed increased elastolytic activity, and Finlay and associates (20) reported similar observations using cultured MACs derived from human lavage fluid. Selman and colleagues (21) showed upregulation of interstitial collagenase in smoke-exposed guinea pigs, and D'Armiento and coworkers (22) showed that transgenic mice overexpressing interstitial collagenase developed emphysema. Hautamaki and associates (18) created mice in which the gene for MAC metalloelastase had been knocked out and observed that these mice did not develop emphysema when exposed to cigarette smoke. Ohnishi and coworkers (23) found increased levels of various metalloproteases including matrix metalloproteinase (MMP)-1, MMP-2, MMP-8, and MMP-9 in human lungs with emphysema compared with lungs without. More recently, Ofulue and associates (24) reported that the development of emphysema in smoke-exposed guinea pigs correlated with MAC but not PMN numbers and MAC-derived proteolytic activity, and that depletion of MACs prevented smoke-induced emphysema whereas depletion of PMNs was not protective (25).

All of these studies continue to raise questions about the cells and proteases involved in smoke-induced connective tissue breakdown in the lung, the process that ultimately leads to emphysema, as well as the time course of these events. In this study we have used a very acute smoke-injury model to examine connective tissue breakdown and to determine which cells and proteases (or groups of proteases) are involved.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Smoke Exposure and Lavage Procedures

Each experimental group consisted of five C57-BL/6 mice. Mice were exposed to the whole smoke from two full Kentucky 2R1 cigarettes (obtained from the University of Kentucky) using a standard smoking apparatus (described by us elsewhere [26]). Control mice were sham-smoked. At various times following smoke exposure, mice were killed by halothane overdose, and the lungs were removed from the chest cavity. A 20-gauge catheter was inserted into the trachea and the lungs were lavaged six times with 1 ml of ice-cold saline for cell counts or with distilled water for connective tissue degradation analysis (water was used because concentration of salts during sample preparation for the high-performance liquid chromatography [HPLC] procedure interferes with the analysis of desmosine [DES] [27]).

For inflammatory cell measurements, the saline lavage was centrifuged at 200 × g at 4°C for 10 min. The supernatants were decanted and the cell pellets were resuspended in 200 µl of saline. Total cell counts were performed in a hemacytometer and differential cell counts were performed on a 10-µl drop of the cell suspension heat-fixed on a slide and stained with hematoxylin and eosin.

To examine the effects of a single smoke exposure over time, C57-BL/6 mice were exposed to smoke from two whole cigarettes; test and control groups were killed 6, 24, and 48 h after smoke exposure, and their lungs were lavaged. To examine the effects of cigarette smoke dose, groups of mice were exposed to one, two, or three whole cigarettes and killed 24 h later.

1AT Purification, Administration, and Serum Levels

Purified human 1AT (Prolastin) was purchased from Bayer, Inc. (Etobicoke, ON, Canada) and the preparation was further purified to remove contaminating proteins. Prolastin was reconstituted in 20 mM sodium phosphate buffer (pH 7.6) and filtered through an Ultra-Free 15 100K centrifugal filter device (Millipore, Bedford, MA) to remove proteins of 100-kD size or greater. The concentrate was washed several times with buffer and the resulting filtrates were pooled. The pooled filtrate was run through a Cibacron Blue column (Pharmacia, Oakville, ON, Canada) at 4°C to selectively remove albumin, the major contaminating protein. Human 1AT was washed out of the column with sodium phosphate buffer (pH 7.6). Finally, the solution was concentrated with an Amicon ultrafiltration cell using a 10K Diaflo membrane (Amicon, Beverly, OK). Purity of the resulting product was checked on a 12.5% homogenous Phastgel (Pharmacia) with silver staining.

Initial experiments were performed to determine the increase in serum 1AT after injection of various amounts of purified product and to determine the time course of the serum elevation (see RESULTS) after intraperitoneal injection. For this purpose blood was drawn from the tail vein and analyzed by a competitive enzyme-linked immunosorbent assay (ELISA) using purified human 1AT as a standard (Calbiochem, La Jolla, CA). The plates were coated with 100 ng/well 1AT and blocked with phosphate-buffered saline (PBS) containing 3% bovine serum albumin at pH 7.4. A total of 50 µl of sample was added into the wells, followed by 50 µl of antihuman 1AT antibody (Boehringer Mannheim, Laval, PQ, Canada) depleted against mouse serum. After the solution was mixed, the plate was incubated for 1 h at room temperature. After incubation the plate was washed five times with PBS. A total of 100 µl of secondary antibody (goat antirabbit immunoglobulin G-horseradish peroxidase [HRP] 1:10,000) was measured into each well. After 1 h incubation at room temperature, the plate was washed five times with PBS. The color reaction was developed using TMB ELISA substrate (ICN Biochemicals). A total of 100 µl of substrate was added into each well and incubated for 5 min. Color development was stopped by the addition of 100 µl 2 M HCl solution. Absorbance was measured at 450 nm.

After examination of the data, a dose of 20 mg of purified product in 1.0 ml of saline was selected and mice were injected 24 h before smoke exposure.

Western Blot Analysis of Bronchoalveolar Lavage Fluid for 1AT

Unconcentrated bronchoalveolar lavage (BAL) fluid (BALF) samples were separated on a 12% polyacrylamide resolving gel with purified human 1AT (Sigma, Mississauga, ON, Canada) as a positive control. After electrophoresis, the protein bands were immobilized on a nitrocellulose membrane and the membrane was probed with antihuman 1AT primary antibody (Boehringer Mannheim) and goat antirabbit HRP-conjugated antibody (ICN Biochemicals) and developed by enhanced chemiluminescence (Amersham, Oakville, ON, Canada).

Depletion of Neutrophils

Neutrophils were depleted by administration of an antimouse neutrophil antibody (APA, Cat. no. AIA31140; Accurate Scientific, Westbury, NY). Two intraperitoneal injections of 0.2 ml of antibody were given 24 h apart. Test lavages and blood counts 24 h after intratracheal instillation of 50 µg of lipopolysaccharide (LPS) (used here as an agent known to evoke a marked lavage neutrophilia), administered 24 h after the second antibody dose, showed that this procedure succeeded in depleting 100% of the PMNs from the peripheral blood and more than 90% of the PMNs in the lavage. For smoke experiments, mice were exposed to smoke or air 24 h after the second antibody dose.

DES and Hydroxyproline Analysis

The water lavageate was lyophilized and hydrolyzed in HCl at 110°C for 48 h, and then analyzed for DES and hydroxyproline (HP) on a Waters HPLC system (Waters Associates, Milford, MA) using our previously published protocol (26).

BALF Elastase-Like Activity

Elastase-like activity in BALF was measured by colorimetric assay. BALF samples were lyophilized and reconstituted in water to make a solution that was five times more concentrated than the originial sample. N-succinyl-Ala-Ala-Ala-p-nitroanilide (SLAPN; Sigma) was used as the substrate. The assay was carried out in 0.2 M Tris-HCl, pH 8.0, with and without the addition of 10 mM ethylenediaminetetraacetic acid as a metalloelastase inhibitor or 10 mM N-methoxysuccinyl-Ala-Ala-Pro-Val chloromethyl ketone as a serine elastase inhibitor.

BAL samples were assayed in duplicate wells of 96-well flat bottom plates (VWR Canlab, Mississauga, ON, Canada). Each well contained 100 µl of the appropriate assay buffer as described earlier, 50 µl of 0.5 mg/ml SLAPN substrate, and 50 µl of sample. Negative control wells contained 150 µl of assay buffer and 50 µl of substrate. Background absorbance of each BAL sample was assessed by incubating 150 µl of assay buffer with 50 µl of BALF (this value was subtracted from the absorbance of the test wells). The absorbance of the wells was measured at 405 nm wavelength.

Statistics

Most comparisons were made by analysis of variance. Because neutrophil counts were often zero (especially in control animals and the neutrophil depletion or 1AT treatment groups), comparisons were performed with the nonparametric Kruskal-Wallis test, although the graphs show means and standard deviations (SDs). Values of P < 0.05 or less were considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Figure 1 shows the effects of cigarette dose with death at 24 h after smoke exposure. Lavage PMNs but not MACs increased with number of cigarettes smoked, but with a plateau after two cigarettes. Lavage DES and HP also increased with smoke dose in a similar fashion.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Lavage PMNs (top left), DES (bottom left), and HP (bottom right) (measured at 24 h) increase with increasing numbers of cigarettes, with a plateau after two cigarettes. Lavage MAC numbers (top right) are not changed. *Significantly greater than control. Values are means ± SD.

Figure 2 shows the effects over time of a single exposure to two cigarettes. PMNs were increased at 6 h, although the difference from control was not statistically significant, and were significantly elevated at 24 h. By 48 h neutrophil counts had decreased. MAC numbers were not affected by smoke. Both lavage DES and HP were increased at 6 and 24 h, but returned to baseline at 48 h.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Lavage PMNs (top left), DES (bottom left), and HP (bottom right) increase to 24 h after a single exposure to two cigarettes (hatched columns), and then decrease to control values (filled columns). Lavage MAC numbers (top right) are not changed. *Significantly greater than control. Values are means ± SD.

Figure 3 shows the effects of anti-PMN antibody administered before smoke exposure as described in MATERIALS AND METHODS. The antibody reduced lavage PMN levels to zero and prevented smoke-mediated increases in DES and HP, again without affecting MAC numbers.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Anti-PMN antibody reduces PMN counts to zero (top left) (bar in graph is a place marker) in both control and smoke-exposed animals, and also returns smoke-induced elevations of DES (bottom left) and HP (bottom right) to control values. Lavage MAC numbers are not changed (top right). *Significantly greater than control. Values are means ± SD.

Figure 4 shows the effects of exogenous human 1AT administered as described in MATERIALS AND METHODS on serum 1AT levels and the time course of the serum elevation. Serum levels were increased two to three times over normal C57-BL/6 levels (28) at 24 h and declined to baseline over 7 d. Figure 5 illustrates Western blots of lavage fluid in animals given human 1AT; the human protein is present in the lavage fluid of both smoke- and air-exposed animals. Only a single band at 52 kD was observed and no complexes were identified. Figure 6 shows the effects of exogenous 1AT administration 24 h before smoke exposure: 1AT totally abolished the smoke-mediated increases in PMNs, DES, and HP. MAC numbers were not affected by exogenous 1AT.

View larger version (7K): [in this window] [in a new window]   Figure 4.   Time course of serum human 1AT level after administration of 20 mg of human 1AT. Values are means ± SD.

View larger version (70K): [in this window] [in a new window]   Figure 5.   Western blot of lavage fluid at 48 h after 1AT administration (24 h after smoke exposure) using antibody against human 1AT. Human 1AT is present in both control (nonsmoked) animals (top panel: lanes 1-5, no human 1AT; lanes 6-9, 20 mg human 1AT; std indicates pure human 1AT) and smoked animals (bottom panel: lanes 1-4, no human 1AT; lanes 5-8, 20 mg human 1AT; and std indicates pure human 1AT).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Administration of human 1AT 24 h before smoke exposure abolishes smoke-induced elevations in PMN counts (top left), DES (bottom left), and HP (bottom right). MAC numbers are not affected (top right). *Significantly greater than control. Values are means ± SD.

Figure 7 shows lavage serine and metalloelastase-like activity at 24 h after two cigarettes. Cigarette smoke exposure elevated both serine and metalloelastase activity, and 1AT returned the serine but not the metalloelastase levels to control values.

View larger version (21K): [in this window] [in a new window]   Figure 7.   Smoke elevates both serine and metalloelastase activity in lavage fluid after 1AT administration. Administration of human 1AT 24 h before smoke exposure reduces serine elastase activity to control values, but does not affect metalloelastase activity. *Significantly greater than control. Values are means ± SD.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this paper we have examined the acute effects of cigarette smoke on connective tissue breakdown. The controversies regarding the effector cell(s) and protease(s) leading to emphysema were considered in the opening paragraphs. The quantification of connective tissue breakdown in smoke- exposed humans has been similarly problematic and reports using various methods have given inconsistent results. Most investigators have examined urinary DES, and early studies using radioimmunoassays concluded that there was no detectable difference in urine DES levels between control and 1AT-deficient subjects (29) and no correlation between urine DES and smoking status (30). More recent studies suggest that plasma and BALF elastin-derived peptides are increased in smokers with evidence of chronic obstructive pulmonary disease (COPD) (31, 32), and that the highest levels of urinary DES are present in those individuals with the most rapid decline in pulmonary function (33). Augmentation with 1AT reduces urinary DES levels in individuals with 1AT deficiency (34) and preliminary testing of a synthetic serine elastase inhibitor also indicates it is capable of reducing urinary DES excretion in some patients with COPD (35).

Apart from problems of interfering substances, measurement of urinary or plasma DES is not entirely specific because it will pick up the normal elastin turnover from all organs in the body; Gottlieb and colleagues (33) noted that in normal individuals, only 19% of urinary DES is derived from the lung. Measurements of lavage DES are potentially more informative, but few data are available. Ofulue and associates (24) recently reported that, using a radioimmunoassay, elevated levels of both DES and elastin split products could be detected in the lavage fluid of rats 2 mo after initial smoke exposure, but not before. The present experiments extend this phenomenon more proximally: elevated levels of DES were evident within 6 h of smoke exposure, indicating that smoke can induce connective tissue breakdown as a very early event. The fact that HP levels were also elevated by 6 h serves to reinforce the idea that the proteolytic attack resulting from cigarette smoke exposure degrades a whole spectrum of matrix components, rather than just elastin.

Because smoke increases epithelial permeability it is remotely possible that smoke-induced leakage of normal serum DES and HP might spuriously increase lavage DES and HP. However, we were not able to detect any DES in normal mouse serum with our HPLC technique, with a detection limit determined from spiked samples of 3 pmol/ ml. By contrast, the lavage DES levels in our control animals were 20 to 30 pmol/ml and in the smoke-exposed animals on the order of 70 pmol/ml. Given that a normal mouse serum volume is about 1.5 ml, serum leakage could not account for more than a very small part of the increase in lavage connective tissue breakdown products.

The present experiments suggest that, in this very acute model of smoke-induced injury, neutrophils act as the initial effector cells involved in connective tissue breakdown. In our animals neutrophils constitute, on average, some 1 to 3% of lavage cells in the smoke-exposed mice, values comparable to those seen in human smokers (8). This increase is more significant than might be thought from such small numbers because in absolute terms, exposure to two cigarettes typically increases neutrophils manyfold (see figures). The importance of increased neutrophil numbers is evident from the experiments using antineutrophil antibody; by reducing neutrophil counts essentially to zero, connective tissue breakdown was abolished. Similarly, the efficacy of exogenous 1AT in preventing connective tissue breakdown suggests that a serine protease is involved (but see later discussion), and the effect of 1AT in abolishing the elevated serine elastase activity in the smokers is evident in Figure 7. Although our experiments do not prove that that protease is neutrophil elastase, this is the likely culprit because it is the most potent elastase in the neutrophil armamentarium and accounts for most neutrophil-derived elastolytic activity. However, other neutrophil proteases such as cathepsin G and proteinase 3 may also play a role.

Our experiments raise the possibility that 1AT not only may act to neutralize serine proteases, but also may have an anti-inflammatory effect. DES and HP levels decreased in the animals given human 1AT in these experiments, but neutrophil levels returned to baseline as well, suggesting that 1AT might be preventing connective tissue breakdown by decreasing neutrophil recruitment. Anti-inflammatory effects have been documented for several different antiproteases. Both 1AT and 1-antichymotrypsin have been shown to decrease neutrophil chemotaxis in vitro (36, 37), at least partly by inhibiting the neutrophil-derived protease, cathepsin G, or by interacting with a neutrophil surface-associated chymotrypsin-like enzyme, both of which appear to function in this context to potentiate the chemotactic response (37). Secretory leukocyte protease inhibitor (SLPI) also decreases PMN influx but by a different mechanism: Lentsch and coworkers (38) have proposed that SLPI increases cytoplasmic IB and prevents activation of nuclear factor-B, leading to decreased endothelial intercellular adhesion molecule-1 expression and decreased egress of PMNs from the circulation into the lung. We are currently investigating the actions of 1AT in this regard.

However, 1AT can also act to upregulate the inflammatory response, inasmuch as complexes of 1AT and neutrophil elastase (39) or 1AT that has been proteolytically inactivated by macrophage elastase (40) act as neutrophil chemoattractants. Moreover, both elastin- and collagen- derived peptides are chemotactic for neutrophils (41, 42), so it is also possible that by inhibiting neutrophil-mediated matrix proteolysis, 1AT prevents formation of chemotactic peptide fragments and thus shuts off a positive feedback cycle. The current data on cigarette smoke exposure do not allow us to distinguish these possibilities. But it is interesting that when we administered 1AT along with LPS to mice, we were able to return lavage DES completely to control values even though lavage PMNs remained markedly elevated at 24 h (R. Dhami, unpublished data), suggesting that direct 1AT-mediated protease neutralization can be important in preventing connective tissue breakdown.

Our data also raise some questions about the exact effects of small differences in 1AT levels. As noted at the beginning of this paper, attempts to show directly that functional levels of 1AT are reduced in smokers have never yielded consistent results. C57-BL/6 mice have 1AT levels that, in human terms, are within the normal range, albeit slightly lower than other standard strains such as NMRI or BALB/c mice (28). Nonetheless, this combination of smoke and endogenous 1AT leads to acute connective tissue breakdown, and the same level of smoke exposure also produces mild emphysema in C57- BL/6 mice if administered daily over 6 mo ([18], and R. Dhami and associates, unpublished data). These findings indicate the protease burden derived from acute cigarette smoke exposure is certainly able to overwhelm the protective effects of endogenous 1AT in C57 mice. It is interesting in this regard that Sandford and colleagues (43) have recently reported the association of cigarette smoke-induced COPD with the 1AT phenotype MZ, where the 1AT levels are about 50% of those seen in MM subjects, but not in MS subjects, where the level is about 75% of MM; and Campbell and coworkers (44) have shown that serum from MZ and MS individuals provides slightly but significantly less proteolytic protection than does serum from MM individuals in an in vitro assay system. The observations not only emphasize the importance of 1AT (and by implication, neutrophil elastase), but also suggest that even fairly small variations in 1AT levels may lead to a situation in which smoke-induced proteolysis overwhelms host defenses.

Our results clearly differ from several recent reports in the literature that emphasize the role of MACs and metalloproteases in smoke-mediated proteolysis, as described in this paper's opening paragraphs. The obvious and crucial difference is that our experiments are very acute and provide no information about long-term changes. We cannot rule out the possibility that MACs, or a combination of neutrophils and MACs (and a combination of serine and metalloproteases), play a greater role over time, as those reports would imply. It is also possible that neutrophil- mediated connective tissue breakdown is entirely a transient process but this seems unlikely since chronic smokers have elevated lavage PMN levels (8); or that compensatory mechanisms serve to neutralize PMN-mediated proteolysis over time, which is one conclusion to be drawn from the data of Ofulue and associates (24) who reported that smoke-induced lavage and interstitial PMNs are increased at 1 mo and then decline in number, whereas MAC increases and increases in lavage DES or elastin split products are not seen until 2 mo. Emphysema, which in that study correlates with lavage DES levels, follows MAC numbers and elastolytic activity. We did not attempt to examine MAC function in the current study and it is possible that MAC protease production is increased; the metalloelastase activity we observed might be due to MAC- or neutrophil-derived enzymes such as gelatinase B  (MMP-9). However, because adminstration of 1AT decreases the serine but not the metalloelastase activity and also returns lavage DES and HP to control levels, metalloproteases do not appear to be playing any acute role in connective tissue breakdown.

It is important to emphasize that the acute production of neutrophil-mediated smoke-induced connective tissue breakdown seen in our model does not necessarily translate into the development of emphysema over time. Unfortunately, the current experiments are necessarily very short-term because repeated administration of anti-PMN antibody or of human 1AT to these animals rapidly leads to the development of serum sickness, making this approach unsuitable for longer studies. However, our data do show that, acutely, smoke-evoked, neutrophil-derived serine proteases can cause connective tissue breakdown in the lung, and that 1AT is protective, perhaps through an anti-inflammatory mechanism.