Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Characteristic features in cystic fibrosis (CF) are chronic, neutrophil-dominated airway infection (especially with Pseudomonas aeruginosa) and hypertrophy of airway secretory cells, leading to excessive mucus secretion. This chronic airway inflammation results in high concentrations of neutrophil-derived mediators in airways, in particular inflammatory cytokines (1, 2), long-lived oxidants (3), and neutrophil-derived serine proteases (4).

The neutral serine protease family (5) is composed of human neutrophil elastase (HNE), cathepsin G, and proteinase 3 (PR3) (6), also called myeloblastin (7). Cathepsin G is a chymotrypsin-like serine proteinase that has different substrate specificity and biologic activities from HNE and PR3 (8). Both HNE and PR3 are trypsin-like serine proteases with comparable proteolytic activity (9), and together they share a 54% nucleotide sequence homology (7, 10, 11). HNE has been implicated in lung diseases because of its ability to digest structural elastin (12, 13). PR3 also degrades extracellular matrix proteins (9); PR3 has been found in CF airways (14), and local instillation of PR3 is reported to cause experimental emphysema (15). HNE is a potent secretagogue in airway glands and in goblet cells (16). CF sputum induces mucus secretion, even after 30,000-fold dilution; this activity is inhibited by ICI 200,355 (17, 18), which inhibits both HNE and PR3. The secretagogue activity of PR3 is unknown, and the relative potential contribution of PR3 in causing neutrophil-mediated secretion has not been determined.

In the present studies, we evaluated purified PR3 as an airway secretagogue. ICI (Imperial Chemical Industries) 200,355 inhibits both HNE and PR3 activity, whereas the secretory leukoproteinase inhibitor (SLPI) inhibits only HNE (19, 20). We used this difference in activity to evaluate the relative contribution by HNE and PR3 in CF sputum and in neutrophil lysates. In addition, we examined the relationship between the concentration of HNE and PR3 in CF sputum, and both taurine concentration, a marker of inflammation (21), and forced expiratory volume in 1 s, expressed as percent of forced vital capacity (FEV₁%), a marker of airway potency.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Patients

Airway secretions were collected from 49 CF patients, mean age 15 ± 3.2 yr (range 6 to 26), who were hospitalized for severe infection. Thirty-one patients had chronic colonization of sputum with P. aeruginosa. Of these, 13 had fungi and 7 had Staphylococcus aureus infection. Among the remaining 18 patients, 11 had S. aureus, 5 had fungus infection, and 2 had both. All patients had clinical features typical of CF, including positive sweat tests.

Airway secretions were also collected from 10 patients with chronic bronchitis who were hospitalized for severe infections. The mean age was 45 ± 16 yr. All patients were seen at the Hôpital des Enfants Malades or at Hôpital Laennec, Paris, France.

Blood was taken from the patients at the time of routine laboratory studies, and the neutrophils were isolated immediately, as described. Controls consisted of 10 healthy subjects recruited among blood donors at the Necker Hospital Blood Center.

Parental informed consent (all children) and the patient's consent (patients older than 7 yr of age) were obtained before the study. The study was approved by the Ethics Committee of the Necker Enfants Malades Hospital.

Preparation of Sputum Samples and P. aeruginosa Culture Supernatants

Sputum samples obtained after chest physiotherapy were diluted 1:2 with phosphate-buffered saline (PBS, pH 7.4), vortexed briefly, and centrifuged for 20 min at 10,000 × g. The supernatants were collected and stored at 80°C (3). To assess the enzymatic activity from the insoluble fraction of the sputum, the pellet was solubilized in an equal volume with 1% Triton-X100 (Sigma, St. Louis, MO) and then centrifuged 30 min at 15,000 × g. The supernatant containing the solubilized materials was stored at 80°C.

Three different strains of P. aeruginosa were used to produce culture supernatants according to previously described methods (22): One strain was from the American Type Culture Collection (ATCC) (reference 10145T) and the two other strains, NEM 14763 and NEM 128387, were isolated from CF patient sputum and were nonmucoid and mucoid, respectively. One hundred microliters of an overnight culture was added to 5 ml of Mueller-Hinton broth and incubated in a shaking water bath at 37°C for 24 h. This culture was then centrifuged at 8,000 × g for 20 min, and the supernatants were stored at 80°C until use.

Neutrophil Isolation and Purification of PR3 and HNE

Neutrophils were isolated from heparinized venous blood using density-gradient centrifugation in Ficoll-Hypaque (Pharmacia LKB Biotechnology Inc., Piscataway, NJ), as described previously (23). Neutrophils were disrupted by nitrogen cavitation, and azurophil granules were isolated using a percoll gradient. PR3 and HNE were purified from azurophil granules using an aprotinin-Sepharose column followed by Matrex gel orange A (Amicon Corp., Danvers, MA) chromatography and then immunoadsorption (24). The purity and the molecular weight of purified PR3 and HNE were determined by sodium dodecyl sulfate- polyacrylamide gel electrophoresis (SDS-PAGE), followed by silver staining (data not shown). Neutrophil lysates were obtained by sonication at the indicated concentration in the absence or in the presence of 1% Triton, followed by 30-min centrifugation at 15,000 × g to eliminate insoluble material.

Measurement of Antigenic PR3 and HNE by Enzyme-Linked Immunosorbent Assay

For measurement of PR3, enzyme-linked immunosorbent assay (ELISA) 96-well plates (Nunc Maxisorb; CML, Nemours, France) were coated with the anti-PR3 monoclonal antibody WGM2 (kind gift of Dr. Csernok, Borstel, Germany) (25) diluted 1:100 in carbonate buffer, pH 9.6, to achieve a final concentration of 2 µg/ml. After saturation with 1% casein, various dilutions of sputum samples (1:100 to 1:600) were added and incubated for 1 h at 37°C. The second antibody was a biotinylated PR3 monoclonal antibody (Clone CLB 12.8) detected with an alkaline phosphatase- streptavidin complex (Amersham, Buckinghamshire, UK) using p-nitrophenyl phosphate (Sigma) as the substrate and making measurements at optical density (OD) 405 nm.

For HNE ELISA, 96-well plates were first coated overnight with an antimouse Fc monoclonal antibody (Sigma) in a carbonate buffer, pH 9.6. An HNE mouse monoclonal antibody (Dako, Glostrub, Denmark) diluted 1:500 in PBS containing 1% casein was then added and incubated 1 h at 37°C. Diluted sputum samples were then added and incubated for 1 h at 37°C. The next antibody, a biotinylated HNE rabbit polyclonal antibody (Calbiochem, La Jolla, CA) diluted 1:200 in PBS/casein was incubated for 1 h at 37°C and was detected with an alkaline phosphatase-streptavidin complex. PR3 and HNE purified from neutrophil azurophil granules (as described previously) were used for ELISA standards, and their respective concentrations were determined in sputum samples and in isolated neutrophils. When mentioned, their concentrations were adjusted on the basis of total protein concentration measured with the BCA micromethod (Pierce, Rockford, IL), using bovine serum albumin as standard.

Western Blot Analysis of PR3 and HNE

PR3 Western blot analysis was performed under reducing conditions, using a polyclonal rabbit anti-PR3 (gift of Dr. Joelle Gabay, Columbia University, New York, NY) as previously described (26); the secondary antibody was the F(ab')₂ of donkey immunoglobulin (Ig)G antirabbit IgG conjugated to alkaline phosphatase (Jackson Immunoresearch Laboratories, West Grove, PA).

Measurement of PR3 and HNE Enzymatic Activity

The enzymatic activity of PR3 and HNE were evaluated by measuring the hydrolysis of the tripeptide thiobenzyl ester (Boc-Ala-Pro-Nva-SBzl; Sigma), which has been described to be a specific substrate for HNE and not for cathepsin G (27, 28). Enzymatic hydrolysis of Boc-Ala-Pro-Nva-SBzl was measured in the presence of 5.5'-dithiobis-2-nitrobenzoic acid (DTNB). The complete system consisted of 10 µl DTNB (16 mM); 10 µl of Boc-Ala-Pro-Nva-SBzl (6.5 mM); 160 µl N-2-hydroxyethylpiperazine- N'-ethane sulfonic acid 100 mM/NaCl 0.5 mM, pH 7.2; and 10 µl of sample or purified serine proteinase (HNE or PR3 at 30 µg/ml). Elastase-like enzymatic activity was measured at OD 405 nm. PR3 and HNE purified from human neutrophils were used as standards. To quantify specifically the enzymatic activity of either PR3 or HNE, we used two inhibitors: (1) ICI 200,355 (17), an HNE inhibitor (kindly provided by Zeneca Pharmaceuticals, Wilmington, DE) that inhibits both HNE and PR3; and (2) SLPI (kindly provided by Dr. Michel Chignard, Institut Pasteur, Paris, and by Dr. John Hoidal, Pulmonary Medicine, University of Utah Medical Center, Salt Lake City, UT), which inhibits only HNE (18, 19). The respective activities of PR3 and HNE were evaluated in neutrophil lysates (20 × 10⁶ polymorphonuclear neutrophils [PMN]/ml, diluted 1/50) or in sputum (diluted 1/25), both in the soluble fraction and in the fraction solubilized from the pellet. The samples were incubated for 30 min at 37°C with either 100 µM ICI 200,355 or 100 µM SLPI or 100 µM phosphoramidon (Sigma), a specific inhibitor of P. aeruginosa (18) or buffer before addition of the substrate, Boc-Ala-Pro-Nva-SBzl. As a result, the total elastase-like activity (HNE + PR3) was defined as the enzymatic activity toward Boc-Ala-Pro-Nva-SBzl that is inhibited by ICI 200,355. The HNE-derived activity was defined as the fraction that is inhibited by SLPI. The PR3-derived activity was defined as the fraction that is insensitive to SLPI.

The enzymatic activity of P. aeruginosa culture supernatant was assessed using hydrolysis of the substrate Boc-Ala-Pro-Nva-Sbzl and measurement of the OD at 405 nm as described. It was compared to purified PR3 or HNE in the presence or absence of either ICI 200,355 or phosphoramidon.

Taurine Assay

The taurine concentration in sputum was assayed by ion-exchange chromatography and ninhydrin coloration on a Beckman 6300 analyzer (Beckman Instruments, Palo Alto, CA) (3).

Measurement of Gland Serous Cell Secretion Induced by PR3 and HNE

Bovine tracheal gland serous cells, which maintain the characteristics of differentiated serous cells, were cultured as described previously (16). The cells are capable of incorporating radiolabeled precursors into macromolecules (chondroitin sulfate proteoglycan) and secreting them in response to secretory stimulation, without toxicity for the cells (16).

Cells were studied between passages 10 and 25. Cells were seeded at a density of 2 × 10⁴ cells/cm² on tissue-culture plastic coated with human placental collagen in medium containing 40% Dulbecco's modified Eagle's H21 medium, 40% Ham's F12 medium, 20% fetal calf serum, and 50 µg/ml gentamycin.

Confluent monolayers of serous cells were incubated with 2 ml medium containing 7.5 µCi/ml Na235SO4 (ICN Radiochemicals Inc., Irvine, CA). After 24 h, the medium containing the radiolabel was removed, the cells were washed with PBS, and serum- and antibiotic-free medium was added to the flask. The medium was renewed every 30 min for 210 min. At 210 min, the medium was collected and either replaced with fresh medium alone (baseline control), or with medium containing either purified HNE or PR3. At 240 min, the medium was collected again, and the spent medium from the 180 to 210-min and the 210 to 240-min incubation periods was dialyzed exhaustively against distilled water to remove unincorporated radiolabeled sulfate. Nondialysable ³⁵S-labeled macromolecules were counted after addition of scintillation fluid by scintillation spectroscopy. Secretion was expressed as the percentage increase in release of ³⁵S-labeled macromolecules during incubation with the serine proteinase (either HNE or PR3) over the release during the immediately preceding time period. The calculation was corrected for the slightly declining baseline, which was determined by controls incubated with medium alone.

Statistical Analysis

Data were analyzed with the Statistica Software package (Statsoft, Tulsa, OK). Results are means ± SEM and are compared using nonparametric tests (Wilcoxon or Mann- Whitney U test for paired samples, as indicated). Differences were considered significant when the P value was less than 0.05. Standard regression analysis and Pearson R correlation coefficients were used to determine the relationships among PR3, HNE, taurine, and FEV₁%.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Enzymatically Active PR3 Is Present at High Levels in CF Sputum

Initially, we determined the concentrations of HNE and PR3 in CF sputum by using specific ELISA, which detects both free plus bound (total) enzyme (data not shown). Hence, the levels measured may include enzyme bound to inhibitors. In CF sputum, the concentration of PR3 was higher, but not significantly, than the concentration of HNE (146.7 ± 20.5 versus 99.7 ± 18.6 µg/ml, n = 41, P = 0.063; Figure 1A). In contrast, in sputum from patients with non-CF chronic bronchitis, the mean level of PR3 was significantly lower than that of HNE (34.7 ± 1.5 versus 59.8 ± 6.1 µg/ml, n = 10, P = 0.013). Of note, both PR3 and HNE levels were lower in non-CF chronic bronchitis than levels observed in sputum from CF patients. Similar results were obtained when the amounts of PR3 and HNE were related to protein concentrations: in CF sputum, 30.06 ± 4.6 versus 18.68 ± 2.82 µg/mg protein, and in non-CF chronic bronchitis patients, 7.39 ± 1.40 versus 12.23 ± 1.55 µg/mg protein, for PR3 and HNE, respectively. In contrast, lysates from purified blood neutrophils at 10 × 10⁶ cells/ml, isolated from controls or from CF patients, contained approximately five times less PR3 than HNE. A similar proportion of PR3 and HNE was found in azurophil granules isolated from control neutrophils.

View larger version (51K): [in this window] [in a new window]   Figure 1.   Measurement of PR3 and HNE in sputum and in neutrophil lysates using specific antibodies. (A) PR3 and HNE were measured with specific ELISA, and concentrations of each enzyme are given in micrograms per milliliter. PR3 and HNE were measured in sputum from CF patients (n = 49), from non-CF chronic bronchitic patients (n = 10), in neutrophils (20 × 106 PMN/ml) from CF patients or controls, and in purified azurophil granules from control neutrophils. Results are expressed as means ± SEM. (B) Western blot analysis of proteinase 3 in cystic fibrosis sputum. Samples (20 µl of each of three different sputum samples: S1, S2, and S3 in sample buffer containing -mercaptoethanol) were run on a 12.5% SDS-PAGE gel and analyzed by Western blot using a rabbit polyclonal anti-PR3. Samples of azurophil granules (PMN AZ), purified PR3, and a combination of PR3 + 1-antitrypsin were run in the right lanes for comparison. Arrows indicate 29 and 80 kD, the molecular weights of free PR3 and control PR3/1-antitrypsin complex formed in vitro in the presence of an excess 1-antitrypsin, respectively.

Western blot analysis of all sputum samples showed that PR3 appeared in sputum at a molecular mass of 29 kD, similar to that obtained in control neutrophil azurophil granules and in purified PR3 (Figure 1B). Additionally, complexed PR3 appeared as two or three bands at molecular mass in the range of 65 to 80 kD; the latter might be complexed to inhibitors, as shown by incubation of purified PR3 with 1-antitrypsin, which leads to two additional bands at approximately 60 and 80 kD.

To quantify the respective activity of either PR3 or HNE, serine proteinase activities were evaluated in neutrophil lysates (20 × 10⁶ PMN/ml, diluted 1/50) and in sputum (diluted 1/ 25) in the presence or absence of either 100 µM ICI 200,355 or 100 µM SLPI (Figure 2). As shown in Figure 2A, enzymatic activity of purified PR3 at 1 µM (30 µg/ml) was inhibited completely by 100 µM ICI 200,355 and was not affected by 100 µM SLPI. In contrast, the enzymatic activity of HNE (30 µg/ml) was completely inhibited either by ICI 200,355 or by SLPI. The enzymatic activity of CF and chronic bronchitis sputum, as well as of PMN lysates, was decreased by ICI 200,355, and to a lesser extent by SLPI, demonstrating an ICI 200,355-inhibitable but SLPI-resistant elastase-like activity (characteristic of PR3).

View larger version (45K): [in this window] [in a new window]   Figure 2.   Measurement of the respective enzymatic activities of PR3 and HNE in sputum and in neutrophil lysates using ICI 200,355 and SLPI as inhibitors. (A) Enzymatic activity was evaluated using the tripeptide thiobenzyl ester Boc-Ala-Pro-Nva-SBzl, which is a substrate for both PR3 and HNE. Purified enzymes (PR3 or HNE at 30 µg/ml), sputum (diluted 1/25 in PBS), or neutrophil lysates (20 × 106 PMN/ml, diluted 1/ 50) were preincubated with 100 µM ICI 200,355, 100 µM SLPI, or PBS for 30 min at 37°C, and then tested for enzymatic activity with Boc-Ala-Pro-Nva-SBzl. Data are OD measured at 405 nm, 1 h after addition of the substrate. (B) Calculation of the respective enzymatic activity attributable specifically to PR3 and HNE was based on the use of the inhibitors ICI 200,355 and SLPI. The activity that was inhibited by ICI 200,355 was considered the combination of HNE and PR3 enzymatic activities. The fraction that was inhibited by SLPI was considered HNE-derived, and the difference was attributable to PR3. (C ) Correlation between PR3 concentration in CF sputum measured by ELISA (expressed in µg/ml, as shown in Figure 1A) and the specific PR3-derived activity (expressed in OD at 405 nm multiplied by the dilution of the sample).

When the combination of both inhibitors was used, it was possible to calculate the respective activity of PR3 and HNE (Figure 2B). In sputum from non-CF chronic bronchitis patients and in total neutrophil lysates from control or CF patients, the elastase-like activity was largely due to HNE, and the PR3-derived enzymatic activity was only 31 ± 5.1%, 31.4 ± 6%, and 33.9 ± 5.2%, respectively. In contrast, in sputum from CF patients, the elastase-like activity was almost equally distributed between PR3 and HNE, 46.2 ± 4.9% versus 53.1 ± 4.8%. Interestingly, the PR3-specific enzymatic activity (defined as the ICI 200,355-inhibitable but SLPI-resistant fraction) correlated with PR3 concentrations measured by ELISA in CF sputum (R = 0.873, P < 0.001, n = 26; Figure 2C). In contrast, no correlation was observed between the HNE-derived enzymatic activity and HNE concentration measured by ELISA (R = 0.093, P = 0.652, n = 26).

We then investigated whether P. aeruginosa proteinases could interact with the serine proteinase assay using Boc-Ala-Pro-Nva-Sbzl as substrate. As shown in Figure 3, serine proteinase activity of sputum from CF patients infected with P. aeruginosa was not modified in the presence of phosphoramidon, a specific inhibitor of P. aeruginosa elastase, whereas ICI 200,355 inhibited both PR3 and HNE present in the sputum. In addition, culture supernatants obtained from P. aeruginosa cultures did not hydrolyze the Boc-Ala-Pro-Nva-Sbzl subtrate, because ICI 200,355 or phosphoramidon were without effect on the OD measured at 405 nm. However, we verified that P. aeruginosa supernatants from the three different strains did show proteolytic activities toward purified fibronectin (data not shown).

View larger version (38K): [in this window] [in a new window]   Figure 3.   Measurement of serine proteinase activities in P. aeruginosa culture supernatants and in sputum from CF patients chronically infected by P. aeruginosa: effect of ICI 200,355 and phosphoramidon. Enzymatic activity was evaluated using the tripeptide thiobenzyl ester Boc-Ala-Pro-Nva-SBzl. Purified enzymes (PR3 or HNE at 30 µg/ml), sputum from CF patients chronically infected with P. aeruginosa (diluted 1/25 in PBS), or supernatants of P. aeruginosa cultures from three different strains (ATCC, mucoid, or nonmucoid) were preincubated with 100 µM ICI 200,355, or with 100 µM or 200 µM phosphoramidon for 30 min at 37°C and then tested for enzymatic activity with Boc-Ala-Pro-Nva-SBzl. Data are OD measured at 405 nm, 1 h after addition of the substrate. Data are means ± SEM of five independent experiments.

HNE and PR3 are very cationic and thus bind to anionic DNA contained in the insoluble pellet of the sputum not evaluated in our assay. To test whether HNE and PR3 activities might partition differently between the pellet and the soluble fraction of the sputum, we solubilized the sputum pellets with 1% Triton to assess the respective enzymatic activities of HNE and PR3 associated with the pellet and with the soluble fraction. As shown in Figure 4, the total activity associated with the sputum pellet was roughly the same as that measured in the soluble fraction. Both fractions were similarly inhibited with ICI 200,355. However, the pellet-associated activity was almost completely inhibited with SLPI, indicating that this serine proteinase activity is mainly due to HNE.

View larger version (31K): [in this window] [in a new window]   Figure 4.   Comparison between the soluble and the pellet-associated serine proteinase activity in CF sputum. Enzymatic activity was evaluated using the tripeptide thiobenzyl ester Boc-Ala-Pro-Nva-SBzl in CF sputum in both the soluble and the pellet-associated fractions obtained after triton solubilization. Both fractions were measured in the presence or absence of inhibitors (100 µM ICI 200,355 or 100 µM SLPI) as described in Figure 2. The soluble fraction was almost completely inhibited by ICI 200,355 but contained an SLPI-resistant activity characteristic of PR3. In constrast, the pellet-associated activity was inhibited by both ICI 200,355 and SLPI, which indicated HNE-dependent enzymatic activity. Purified enzymes (30 µg/ml) were tested in parallel as positive controls for inhibitors. Data are OD measured at 405 nm and are the means of six independent experiments using six different sputum samples.

Neutrophil-Purified PR3 Induces Secretory Activity in Bovine Serous Cells

PR3 (10⁷ M) stimulated secretion in cultured bovine airway gland serous cells (Figure 5). Athough PR3 caused slightly less than half the secretion caused by HNE, this amount of secretion is still much more than "classic" secretagogues such as histamine and bethanechol, whose thresholds are approximately 10⁵ M (16). The secretagogue activity of PR3 and HNE was prevented by the elastase inhibitor ICI 200,355. 

View larger version (11K): [in this window] [in a new window]   Figure 5.   Effect of purified PR3 and HNE on secretion from cultured bovine airway gland cells. Release of 35S-labeled macromolecules in the culture medium was measured after treatment with agonists: 105 M bethanacol (BCH), 107 M PR3, 107 M HNE as compared with untreated control. Both PR3- and HNE-induced macromolecule secretion was inhibited when cells were preincubated with 105 M ICI 200,355. Secretion is expressed as percent increase in release of 35S-labeled macromolecules during incubation with the agonists over the release during the immediately preceding time period. Histograms are means ± SEM (n = 4).

Relation between Sputum Levels of PR3 and Clinical Status

The wide range of PR3 and HNE concentrations in CF sputum led us to examine the possible relationships to the clinical status of the patient. The concentration of taurine was measured in the sputum to evaluate the inflammatory state in CF patients (21). We found a significant correlation between taurine concentrations and PR3 measured by ELISA (R = 0.807; P < 0.001; Figure 6A) or PR3 enzymatic activity (R = 0.808; P < 0.001), whereas no such correlation was found with HNE measured by ELISA (R = 0.412; P = 0.11) or HNE enzymatic activity (R = 0.423; P = 0.06). Similarly, a significant correlation was obtained between PR3 concentrations and the FEV₁% (R = 0.45; P = 0.028; Figure 6B), whereas no correlation was found with HNE (R = 0.13; P = 0.54; data not shown).

View larger version (17K): [in this window] [in a new window]   Figure 6.   Correlation between PR3 concentrations in sputum with clinical status of CF patients. (A) Correlation between antigenic PR3 and taurine concentrations. (B) Correlation between antigenic PR3 and FEV1%.

We investigated the influence of bacterial infection on both PR3 and HNE levels. As shown in Table 1, significantly higher levels of taurine and lower respiratory scores were observed in the group of patients chronically colonized with P. aeruginosa than in the group with no P. aeruginosa infection. In the group of CF patients chronically infected with P. aeruginosa, PR3 levels measured by ELISA and PR3 enzymatic activity were significantly higher in the infected CF patients than in the group of CF patients without P. aeruginosa infection. In contrast, no difference in HNE concentration or HNE-derived activity was observed between the two groups of patients. The proportion of HNE in the total elastase-like activity was significantly decreased in the group of CF patients chronically infected with P. aeruginosa as compared with non-P. aeruginosa- infected CF patients.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Since the first description of PR3 as a new serine proteinase contained in azurophil granules (6) and the demonstration that, like HNE, PR3 could induce experimental emphysema when instilled in hamsters (15), no published study has addressed the clinical implications of PR3 in human airway inflammation.

We found that soluble PR3 was present at higher levels than HNE in CF sputum, whereas in non-CF chronic bronchitis patients much lower levels of serine proteinase with comparable amounts of HNE and PR3 were found. Western blot analysis revealed that soluble PR3 appeared in sputum as a free enzyme or complexed with inhibitors (e.g., 1-antitrypsin) and thus appears at various molecular weights (29). Indeed, in the case of HNE, previous studies have shown that HNE-1-antitrypsin complexes are heterogeneous in size in sputum because 1-antitrypsin is proteolytically degraded (4). We demonstrated that PR3 contained in CF sputum is enzymatically active and that this PR3-derived activity, which is resistant to SLPI, correlated with the antigenic measurement of PR3. In contrast, no such correlation between HNE-derived enzymatic activity and HNE concentration measured by ELISA was found. Thus, it appears that PR3 should be considered as an active part of the elastase-derived enzymatic activity in the soluble fraction of CF sputum (4, 22, 30). What is called "active HNE" measurement as opposed to "immunogenic HNE" measured by specific ELISA is a combination of PR3 and HNE activities. Detergent solubilization of the pellet revealed that sputum pellets contain a serine protease activity equal to that in the soluble fraction, the majority of which appears to be due to HNE. PR3 and HNE thus partition differently between the soluble and the pellet-associated fractions, the pellet being a reservoir of HNE activity. It should be noted that the treatment of sputum pellet with 1% Triton might not be sufficient to solubilize fully the total serine proteinase activity. However, more drastic treatment might affect the protein conformation and/or enzymatic activity.

Interestingly, the clinical relevance of the presence of a high amount of PR3 in the soluble fraction of the sputum is illustrated by the correlation with the clinical status of CF patients. PR3 levels correlated closely with taurine, which is thought to evaluate airway inflammation accurately. Moreover, PR3 contained in the soluble fraction of the sputum appears to be available to mediate deleterious effect on airways and as such correlated highly with respiratory scores (e.g., FEV₁%).

We examined the possible role of PR3 in the secretagogue activity of neutrophil-derived serine proteinases. Our results show that PR3, like HNE, triggers the release of macromolecules from serous cells, which is inhibited by ICI 200,355. PR3-induced secretagogue activity was much higher than that obtained with classical secretagogue (e.g., the parasympathomimetic drug, bethanechol). The molecular mechanisms of HNE- or PR3-induced secretory responses are still not elucidated (16). Although this novel biological activity of PR3 is not surprising in view of the enzymatic specificity shared by HNE and PR3 (9, 20, 26), our data demonstrate that PR3 is involved in the mechanisms leading to airway secretion. Sputum-induced secretion is probably mediated by both HNE and PR3 (18).

Another issue is the respective importance of PR3 and HNE in neutrophil serine-protease activity: Neutrophil lysates or purified azurophil granules contain five times more HNE than PR3. Interestingly, in chronically P. aeruginosa-infected patients, the prominent elastase-like activity appears to be due to PR3, whereas in the absence of P. aeruginosa infection, although the contribution of PR3 in the total enzymatic activity is significant, HNE appears to play the major role. Our data strongly suggest that the deleterious effects of PR3 and PR3-derived secretagogue activity are exacerbated by chronic P. aeruginosa colonization. Perhaps PR3 is more resistant to degradation and proteolytic attack than HNE.

With regard to a potential interaction between P. aeruginosa elastase and neutrophil serine proteinases, we have ascertained that P. aeruginosa elastase, which is a zinc metalloprotease with the ability to hydrolyze elastin or fibronectin, did not hydrolyze the Boc-Ala-Pro-Nval-Sbzl substrate. In addition, previous studies have shown that it is not inhibited by ICI 200,355 (17). Moreover, a specific inhibitor of P. aeruginosa, phosphoramidon, did not interfere with the serine protease activity measured in CF sputum. However, it should be noted that CF sputum samples contain a substantial residual hydrolysis of chromogenic substrate in the presence of ICI 200,355, which is not related to serine proteinase activity and which might be due to P. aeruginosa products, independent of proteinase activity.

PR3 has some biochemical features that are not shared by HNE, in particular its plasma membrane localization (31). The hydrophobicity of PR3 and its subsequent membrane association might play a role in the resistance to proteolysis in sputum. An alternative hypothesis is that neutrophils might not be the only cellular source of PR3 in CF sputum. For instance, monocytes that contain PR3 (our personal data) might represent a potential source of PR3 in the present study.

One important issue is that PR3, in contrast to HNE, is not inhibited by SLPI and thus might play a significant role in protease-mediated airway damage. SLPI is present in large amounts in respiratory epithelial lining fluid and may play a significant role in protecting the normal respiratory epithelium, especially in the upper airways where the levels of SLPI are highest (32). Studies on the interaction of SLPI with PR3 show that SLPI does not bind PR3 and that PR3 degrades SLPI (20), thus potentiating the deleterious role of other proteinases sensitive to SLPI such as HNE. It has also been reported recently that SLPI is a major leukocyte elastase inhibitor in human neutrophils (33). In fact, in neutrophils lysed in 1% Triton, the amount of PR3 measured by ELISA is about five times less than that of HNE, whereas the specific PR3-derived activity is one third of the total activity toward the tripeptide thiobenzylester Boc-Ala-Pro-Nva-Sbzl (our personal data).

Restoration of the proteinase/antiproteinase balance is of crucial importance in CF, because the antiproteinase protection is overwhelmed in CF airways by the presence of HNE on the respiratory epithelial surface and by the inactivation of serine proteinase inhibitors by neutrophil- derived chlorinated oxidants and by proteases (34). In this context, the presence of PR3 has to be taken into account in the balance of proteinase/antiproteinase and in the evaluation of the potential usefulness of SLPI as a therapeutic agent in managing inflammation in CF (35).

In conclusion, our data show that PR3 has to be considered when evaluating proteinase-mediated airway damage and suggest that PR3 plays an important role in the pathophysiology of airway inflammation and in the hypersecretory state found in CF.