INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cystic fibrosis (CF) lung disease is characterized by a chronic bacterial airway infection associated with a massive influx of neutrophils (1). The rapid turnover of airway neutrophils leads to the accumulation of large amounts of extracellular DNA which is thought to hinder the clearance of respiratory mucus (5). The cloning of the human gene encoding deoxyribonuclease I (DNase I) has allowed the development of a novel mucolytic approach in which recombinant human DNase I (rhDNase) is administered by aerosol to CF patients on a daily basis (6, 7). Aerosol rhDNase therapy is associated with an improvement in airflow obstruction and a decrease in the number of infectious respiratory exacerbations in some, but not all CF patients (8).

In addition to DNA, CF sputum contains high levels of neutrophil elastase. The burden of elastase overwhelms the local antielastase defenses in the CF airways resulting in the presence of active elastase at the respiratory epithelial surface (9). The potentially damaging effects of free elastase in the airways include increased airway compliance, opsonin receptor mismatch which decreases neutrophil phagocytosis of Pseudomonas aeruginosa, increased mucus gland secretion, induction of interleukin-8 gene expression and increased bronchial epithelial permeability (12). Interestingly, DNA is known to inhibit elastase activity in purulent sputum (17). Single-stranded sites in DNA have recently been shown to accelerate binding of secretory leukoprotease inhibitor (SLPI) to neutrophil elastase and slow the dissociation equilibrium of the enzyme-inhibitor complex (18). Treatment of DNA with DNase I decreases the association rate constant and enhances the dissociation of the elastase-SLPI complex, thus regenerating active neutrophil elastase (18).

The potential for rhDNase aerosol therapy to increase sputum elastase levels has recently been studied in CF patients (19). Sputum collected 18-20 h after rhDNase showed mildly increased elastase activity only on the first day after initiation of treatment, and subsequently active neutrophil elastase was not increased in sputum obtained 8 h post-rhDnase aerosolization up to 6 mo after initiating therapy. In addition, Costello and colleagues observed that not only did CF sputum elastase activity not increase, but it actually decreased after 12 wk of rhDNase therapy (20). Rochat and coworkers reported that rhDNase treatment of CF patients resulted in a modest although not statistically significant increase in elastase activity when results were expressed per ml of sputum (21). However, the daily acute effects of rhDNase therapy on CF sputum elastase immediately after aerosolization at a time when one would expect DNase I levels to be highest, remain unknown. The aims of the current study were therefore to determine first, whether CF sputum elastase activity increases immediately (within 1 h) after aerosol rhDNase therapy and, second, whether the DNase I-mediated increase in CF sputum elastase activity is sufficient to induce lung damage as assessed by a murine model of elastase-dependent hemorrhage.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Nine patients (four male, five female, age: 20.8 ± 1.8 yr) with cystic fibrosis were included in the study. All patients had moderate CF lung disease (FVC between 40% and 70% predicted) and produced sputum characterized by the presence of pathogenic bacteria including Pseudomonas aeruginosa, Staphylococcus aureus, and Hemophilus influenzae. At the time of sputum collection, no patients presented a respiratory exacerbation as defined by an increase in cough, sputum production and dyspnea, weight loss (> 2 kg) and/or fever (> 38° C). None were taking antibiotics at the time of the study. The study was approved by the institutional review board for human studies.

Study Design

To determine the effects of the in vitro addition of DNase I to airway secretions, sputum was collected from nine patients who had never been treated with DNase I. The sputum collected during chest physical therapy was weighed and diluted 1:1 (wt:vol) with phosphate-buffered saline (PBS), mixed and centrifuged at 25,000 × g for 20 min. The supernatants of the samples were stored at 80° C. The effects of aerosolized recombinant human DNase I (rhDNase; Genentech, San Francisco) were studied by collecting sputum on 4 separate days from each of six patients, 2 mo after they had initiated aerosolized Pulmozyme therapy at 2.5 mg a day. All patients were instructed to collect their sputum during 1 h preceeding rhDNase therapy and during 1 h following therapy. Sputum was treated as described above.

In Vitro Effects of DNase I on CF Sputum

Bovine pancreatic DNase I (Sigma Chemical Co., St. Louis, MO) 1 mg/ml, was added to sputum supernatants, in the presence or absence of 500 µM methoxysuccinyl alanyl-analyl-prolyl-valine chloromethylketone (MeOSAAPV-CMK; Sigma Chemical Co.), a human neutrophil elastase (HNE) inhibitor. The samples were incubated at 37° C for 3 h and subsequently used in assays of HNE and lung hemorrhage as described below. To determine the effects of varying DNase concentrations on sputum elastase activity in vitro, 0.1 mg/ml bovine DNase I was added to sputum and incubated for 3 h at 37° C. Human neutrophil elastase activity was subsequently assayed as described below.

Sputum Elastase Activity

Elastase activity was determined using the chromogenic substrate methoxysuccinyl alanyl-analyl-prolyl-valine p-nitroanilide (MeOSAAPVpNA; Sigma Chemical Co.) (22). A 5 µl sample of sputum supernatant was added to 485 ml PBS and 10 µl of a 5 mM solution of MeOSAAPVpNA diluted in 10% DMSO, 0.5 M NaCl, 0.1 M HEPES, 0.1% Brij-35 (30%) at pH 7.5. After mixing, the change in absorbance was recorded at 410 nm over 2 min at RT in a spectrophotometer (Model DU-7; Beckman Instruments [Canada] Inc., Mississauga, ON, Canada). Activity is reported as µM elastase based on a standard curve derived from purified HNE (Elastin Products Corporation, Pacific, MO).

Lung Hemorrhage Model

To determine whether DNase I could enhance the capacity of CF sputum to induce lung injury, we developed a murine model of lung hemorrhage modified from a similar model described in the hamster (23). Mice (C57BL/6) were anesthetized with inhaled halothane followed by the application at the nares of 50 µl sputum sample prepared as described above. One hour after sputum instillation, the mice were anesthetized (inhaled halothane and 50 mg/kg IP pentobarbitol), the trachea was canulated through a tracheostomy using a 20 gauge needle and the lungs were lavaged with 3 aliquots of 900 µl PBS. The recovered bronchoalveolar lavage fluid was sonicated (Sonifier; Ultrasonics Inc., Plainview, NY) for 90 seconds at 4° C, and hemoglobin concentration was calculated by recording the absorbance of the Soret band at a wavelength of 414 nm (24).

Statistical Analysis

Results are expressed as the mean ± SE of the mean (SEM). Data in Figures 1, 3, and 5 are analyzed by the paired t test. All other data are analyzed by the unpaired t test. A p value < 0.05 was considered significant.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Comparison of CF sputum elastase activity and its potential to induce lung hemorrhage in the presence and absence of bovine DNase I. Sputum collected from nine CF patients who had never been treated with rhDNase, was supplemented with saline (open bars), 1 mg/ml bovine DNase I (closed bars), or 1 mg/ml bovine DNase I and 500 µM MeOSAAPV-CMK for 3 h at 37° C (hatched bar). (A) Elastase activity was determined with MeOSAAPV-pNA. (B) The potential of sputum to induce lung hemorrhage was assessed by instilling 50 µl sputum supernatant intranasally in C57BL/6 mice, followed 1 h later by bronchoalveolar lavage for hemoglobin measurement. Each result represents the mean of three to four determinations.

View larger version (22K): [in this window] [in a new window]   Figure 3.   In vivo effect of aerosolized rhDNase on sputum elastase activity in CF patients within 1 h of therapy. Patients (six patients, each evaluated on 3-4 successive days; n = 22 paired samples) were instructed to collect their sputum during the hour preceeding and the hour following aerosolization of 2.5 mg rhDNase. Elastase activity was measured as described in the text.

View larger version (20K): [in this window] [in a new window]   Figure 5.   The effect of aerosolized rhDNase on the potential of CF sputum to induce lung hemorrhage in the murine lung. Patients (six patients, each evaluated on 2-4 successive days; n = 21 paired samples) were instructed to collect their sputum during the hour preceeding and the hour succeeding 2.5 mg rhDNase aerosolization. Anesthetized C57BL/6 mice received 50 µl intranasal installations of CF sputum supernatants followed 1 h later by bronchoalveolar lavage to determine the degree of lung hemorrhage.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In Vitro DNase I Addition to CF Sputum

Human neutrophil elastase was detectable in all CF sputum samples before the addition of DNase I with a mean activity of 3.91 ± 0.62 µM (Figure 1A). In the presence of 1 mg/ml DNase, HNE activity increased to 7.97 ± 1.56 µM (p < 0.01) and was decreased by the addition of MeOSAAPV-CMK to 0.12 ± 0.10 µM (p < 0.001). Sputum not treated with DNase I induced very little hemorrhage when instilled in the murine respiratory tract (Figure 1B). The mean hemoglobin concentration of BAL in mice exposed to sputum before in vitro DNase I addition was 44.5 ± 12.0 µg/ml. However, in the presence of 1 mg/ml DNase the sputum induced marked hemorrhage as determined by an increase in BAL hemoglobin to 192.8 ± 40.7 µg/ml (p < 0.01). Lung hemorrhage was completely inhibited by the addition of MeOSAAPV-CMK to the DNase I-treated sputum as reflected by BAL hemoglobin levels of 36.5 ± 2.8 µg/ml (p < 0.005 compared with DNase I-treated sputum).

The effect of different DNase I concentrations on sputum elastase activity is shown in Figure 2. As little as 50 µg/ml DNase I was sufficient to increase sputum elastase activity 2.6-fold and a maximum increase of 3.3-fold was observed at 500 µg/ml DNase (no DNase I, elastase activity = 3.4 ± 1.2 µM; 500 µg/ml DNase I, elastase activity = 11.4 ± 0.9 µM, p < 0.05, n = 4 independent experiments).

View larger version (13K): [in this window] [in a new window]   Figure 2.   Concentration-dependent effect of bovine DNase I on CF sputum elastase activity. Sputum from a CF patient who had never received rhDNase was prepared as described in METHODS section. The sputum supernatant was subsequently supplemented with 0-1 mg/ml bovine DNase I for 3 h at 37° C, followed by measurement of neutrophil elastase activity. Each point represents the mean ± SEM of four experiments.

In Vivo Effects of rhDNase on Sputum Elastase Activity

Sputum collected 24 h after rhDNase had elastase activity levels similar to the levels found in sputum before initiation of rhDNase therapy (sputum elastase activity 24 h post-rhDNase = 3.2 ± 0.4 µM versus no rhDNase = 2.6 ± 0.6 µM, p > 0.4, data not shown). However, elastase activity of sputum collected within the first hour after aerosol rhDNase treatment was consistently higher than in sputum collected 1 h before rhNDase therapy (5.9 ± 0.7 versus 2.7 ± 0.3 µM, p < 0.0001, Figure 3).

Murine Elastase Mediated Lung Hemorrhage Model

Purified human neutrophil elastase at concentrations below 6 µM did not induce lung hemorrhage in mice. However, bronchoalveolar lavage fluid hemoglobin was clearly increased in all mice receiving intranasal human neutrophil elastase at concentrations greater than 8 µM (Figure 4).

View larger version (14K): [in this window] [in a new window]   Figure 4.   Human neutrophil elastase-mediated lung hemorrhage in C57BL/6 mice. HNE was diluted in PBS at 0-12 µM. Intranasal instillation of 50 µl HNE solution was followed 1 h later by bronchoalveolar lavage. The BALF was sonicated and hemoglobin content was determined by measuring light absorbance of the Soret band at 414 nm. Each data point represents the mean ± SEM of three to seven animals.

CF-Sputum-Mediated Murine Lung Hemorrhage

Sputum samples obtained before rhDNase, induced little hemorrhage as evidenced by low levels of BALF hemoglobin in the mice. However, the post-rhDNase sputum from six patients each tested on between 2 and 4 consecutive days, induced lung hemorrhage when instilled intranasally in mice (Figure 5; BALF Hb 1 h before rhDNase: 36.5 ± 7.3 versus 1 h after rhDNase: 91.4 ± 22.3 µg/ml, p < 0.05, n = 21 paired samples obtained from 6 patients each on separate days). Only two patients had post-rhDNase elastase greater than 8 µM and their sputum induced the highest levels of BALF hemoglobin (identified by closed circles and closed squares in Figure 5).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The potential for DNase to increase elastase activity in purulent sputum has been recognized for many years (17). Since rhDNase is commonly used as a mucolytic agent in patients with CF, understanding the extent of the interactions between DNase I and CF sputum elastase is desirable. The results of the current study provide two novel observations concerning these interactions. First, the therapeutic administration of aerosolized rhDNase at the recommended dose of 2.5 mg daily was sufficient to increase sputum elastase activity within 1 h of therapy in all CF patients. This effect was observed repeatedly on successive days of rhDNase therapy. Second, DNase was found to increase elastase-mediated lung damage by CF sputum in mice, as reflected by lung hemorrhage. DNase I-mediated increase in sputum elastase activity has been reported in several previous studies (17), however the biological consequences of this increase on the lung remain unknown. Our study provides evidence that CF sputum treated with DNase I not only increases its elastase activity but also significantly increases its potential to induce in vivo toxicity as determined by lung hemorrhage.

CF sputum-induced lung hemorrhage was clearly mediated by neutrophil elastase, since hemorrhage was completely suppressed by the serine protease inhibitor, MeOSAAPV-CMK, which specifically blocks human neutrophil elastase. Bovine DNase I alone was unable to induce lung hemorrhage and had no detectable protease activity. Ying and Simon have recently demonstrated that the rate of association of HNE with its natural inhibitor SLPI, is increased 44-fold in the presence of 23 µg/ml DNA, a concentration much lower than that which is generally observed in CF sputum (18). Although other mechanisms such as charge interactions between highly cationic elastase and anionic DNA may exist, the data from Ying and Simon strongly suggest that the progressive degradation of DNA by DNase I markedly accelerates SLPI-elastase dissociation, thus increasing the levels of elastase activity observed in CF sputum after DNase I treatment (18).

The results of our study are consistent with those reported by Shah and coworkers in which CF sputum collected either 8-12 h or 18-20 h after the last dose of rhDNase did not demonstrate increased elastase activity (19). In four CF patients for whom data was available, we did not observe an increase in sputum elastase activity 24 h after rhDNase compared to levels in sputum collected before initiating rhDNase therapy (data not shown). However, the present study provides data that strongly support the suggestion of Shah and coworkers that higher levels of protease activity are likely to be observed if sputum samples are collected soon after aerosolization of rhDNase (19).

Within the first hour following rhDNase, not only did the CF sputum's elastase activity markedly increase, but its potential to induce hemorrhage in murine lungs was also clearly increased. Several investigators have previously demonstrated that elastase induces lung hemorrhage within minutes of intratracheal instillation, and the maximal amount of hemorrhage is observed at 1 h (23, 25, 26). We were unable to identify morphologic evidence of lung damage by light microscopy in animals with elastase and sputum-induced lung hemorrhage (data not shown). Ultrastructural studies by Morris and colleagues revealed that most of the elastase-treated lung tissue is indistinguishable from saline controls during the first 3 h after intratracheal elastase instillation (27). However, these investigators also demonstrated that some areas of the elastase-treated lung show definite signs of type I epithelial damage with exposed basal lamina. Type I cells adjacent to these denuded areas appeared to be separated from the alveolar walls. It is therefore likely that the acute lung hemorrhage observed in the current study originated, at least in part, from the alveolar space.

An intriguing aspect of the current study is that for the same elastase activity as measured by the low molecular weight substrate, MeOSAAPVpNA, purified human neutrophil elastase induced more lung hemorrhage than did sputum. In contrast to purified elastase, sputum elastase is surrounded by several proteins which, while allowing low molecular weight substrates to interact with the active site of elastase, may limit the access of elastase to lung tissues. ₂-macroglobulin is an example of such a protein which prevents access of elastase to high molecular weight substrates such as elastin, but does not prevent elastase from hydrolysing a low molecular weight substrate such as MeOSAAPVpNA (28). However, the current study did not allow us to determine whether such interactions occur in sputum.

Although the effect of rhDNase aerosol therapy on CF sputum elastase activity was transient, it was observed consistently on each day of therapy. Since aerosolized rhDNase is usually taken daily for indefinite periods by patients with CF, one should be concerned about the potential long term adverse effects of daily, albeit temporary, spikes in neutrophil elastase activity at the airway epithelial surface. It is notable that each of the subjects in the current study had been receiving rhDNase therapy for at least 2 mo prior to sputum collection. The acute increases in post-rhDNase sputum elastase activity therefore were not restricted to the first days following initiation of rhDNase therapy as previously observed at the 8- 12 h post-rhDNase time period (19).

Adverse effects of aerosolized rhDNase were carefully recorded in a randomized, double-blind, placebo-controlled study of 968 adults and children with CF, over a 24-wk period (8). Remarkably, few adverse effects were observed in the treatment group, and, of particular relevance to the current study, rhDNase did not increase the incidence of hemoptysis. A number of possibilities may explain the absence of observable elastase-related adverse effects with aerosolized rhDNase in CF patients included in clinical studies. First, the daily increase in elastase activity may be too transient to produce detectable adverse effects. Elastase induces an acute increase in epithelial permeability allowing plasma proteins to cross the epithelial barrier (16, 23). Among these proteins are protease inhibitors such as alpha1-proteinase inhibitor, which could rapidly inactivate elastase and limit any further damage. Second, although sputum elastase activity increases acutely with rhDNase therapy, the volume of secretions within the lower respiratory tract may be sufficiently reduced by the mucolytic action of rhDNase to decrease the total amount of elastase in CF sputum, thus minimizing the adverse effects of daily increases in elastase activity immediately after rhDNase aerosol. This possibility is supported by data from Costello and coworkers (20) demonstrating a reduction in sputum elastase activity after 12 wk of rhDNase therapy. Third, the duration of available clinical studies may have been too short to detect long-term adverse effects of daily rhDNase-mediated elastase increases.

In summary, the DNase I-mediated increase in CF sputum elastase activity is associated with an increase in the potential of the airway secretions to induce lung hemorrhage. The effect of DNase I on elastase activity in airway secretions is relevant to CF patients receiving aerosolized rhDNase since a marked increase in sputum elastase activity was consistently observed within 1 h of treatment in all of the study subjects. Mucolytic therapy of selected CF patients with rhDNase has proven to be safe and effective in short-term studies. However, based on observations from the current study it would appear reasonable to suggest that elastase-mediated lung hemorrhage should be considered as a potential contributing factor in patients with CF presenting with hemoptysis while receiving rhDNase therapy. In addition, studies of the long-term potential adverse effects of aerosolized rhDNase in patients with CF should take into account DNase-I and elastase interactions.