INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Wegener's granulomatosis (WG) is a chronic inflammatory disorder of unknown etiology. It is characterized clinically by multiple organ involvement with a proclivity for respiratory tract and renal disease (1), serologically by a high prevalence of antineutrophil cytoplasmic antibodies (ANCA), mostly against proteinase 3 (PR3) (2), and histopathologically by vasculitic and granulomatous tissue lesions (3). Current concepts on the pathogenesis of WG focus on neutrophil and lymphocyte-dependent mechanisms. The neutrophils appear to play a dual role in WG and other ANCA-associated vasculitides in that they are targets of autoimmunity and are also involved in causing tissue injury (4). The main target of the autoimmune response in WG is PR3, a serine protease stored in the neutrophil azurophilic granules and secreted upon cellular activation (5). In vitro binding of ANCA to preactivated neutrophils results in functional activation with increased respiratory burst activity, expression of adhesion molecules, and degranulation (2, 4). According to the ANCA-cytokine sequence theory, inflammatory injury to the vessel wall is initiated by the attachment of neutrophils to the endothelium and the release of autoaggressive neutrophil constituents in response to engagement of ANCA with primed neutrophils that express the ANCA target antigen on their surface (2).

The concept of the ANCA-cytokine sequence was developed mainly on the basis of in vitro findings. It has received support from animal models attesting to an essential role of the neutrophils in the pathogenesis of small vessel vasculitides. Two models of ANCA-dependent renal disease showed that in addition to the induction of ANCA, renal perfusion with neutrophil granule extract and hydrogen peroxide or systemic injection of products released by activated neutrophils are required to induce glomerulonephritis, whereas immunization against myeloperoxidase (MPO) alone does not have this effect (6, 7). Moreover, the induction of ANCA capable of activating neutrophils enhances the pathogenic effect on the kidney of subnephritogenic doses of antiglomerular basement antibodies (8). Very recently, the role of the neutrophils was highlighted by the finding that treatment of animals with neutrophil-depleting antibodies substantially diminishes the severity of experimentally induced vasculitis (9).

Because none of these animal models ideally reflects WG, the relevance of the above findings to the situation in humans needs to be proven. The presence in the bloodstream of activated neutrophils has been demonstrated by several groups (10, 11), but the evidence of neutrophil degranulation at affected tissue sites is very limited. Local neutrophil degranulation can be verified by the immunohistologic demonstration of extracellular neutrophil products and this approach has disclosed extracellular neutrophil elastase in ophthalmic lesions of WG (12) and extracellular PR3, MPO, peroxidase (PEROX), and elastase in glomerular lesions caused by WG (13, 14). Moreover, the magnitude of local neutrophil activation correlated with the degree of renal functional impairment in WG glomerulonephritis (13).

To what extent neutrophil degranulation is involved in WG lung disease has not been assessed previously. We therefore measured levels of extracellular neutrophil products in the BALF of a cohort of patients with active pulmonary WG. This approach provides easy access to study material and has a broader range of applicability than does lung biopsy. Moreover, it yields quantitative information that can be related to clinical data. Measurements included MPO, which has proved to be a useful indicator of neutrophil degranulation in other lung diseases, and PEROX activity, a compound measure of the oxidant burden in the lower respiratory tract. Measurements included also PR3, which is not only a potentially autoaggressive agent but is intimately involved in the immunopathogenesis of WG as a target of autoimmunity. Because WG lung disease commonly consists of a combination of vasculitic and granulomatous lesions composed of neutrophils, eosinophils, lymphocytes, and other mononuclear cells (3, 15), we also measured eosinophil cationic protein (ECP), which is a marker of eosinophil degranulation, and soluble interleukin-2 receptor (sIL-2R), a marker of lymphocyte activation. Control groups comprised 12 patients with pulmonary sarcoidosis, which is also a granulomatous disease but lacks vasculitic involvement, and nine control subjects devoid of pulmonary or systemic inflammatory disease.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Nineteen patients with highly active pulmonary disease were included in the WG group (Table 1). All 19 patients complied with the 1990 American College of Rheumatology criteria (16) and the 1992 Chapel- Hill definition for WG (17). Confirmatory biopsies were obtained from the upper or lower respiratory tracts in 12 patients and from kidney, skin, or muscle in the remaining seven patients. Eighteen patients were positive for ANCA, 17 of these had PR3-ANCA, and one had ANCA with perinuclear fluorescence pattern (pANCA) of undetermined specificity. Staging examinations included otorhinolaryngologic, ophthalmologic, and neurologic referral, cranial magnetic resonance imaging, chest radiography (CXR), bronchoscopy, BAL, and a laboratory profile, including testing for ANCA (18). Active progressive disease was diagnosed according to the European Vasculitis Study Group definition (19). Active pulmonary disease was diagnosed on the basis of ill-defined infiltrates or consolidation in 12 patients and solitary or multiple nodules in five patients. Two patients had ulcerative and granulomatous bronchitis in association with a lymphocytic BAL profile. An infectious etiology of these changes was ruled out by appropriate microbiologic studies. Extrapulmonary organ involvement is presented in Table 1. Twelve patients were devoid of any immunosuppressive medication at the time of examination, the others had a major relapse while receiving low-dose prednisolone (n = 2), low-dose methotrexate (n = 2), cotrimoxazole (n = 2), or cyclophosphamide (n = 1).

Six patients with WG underwent two examinations, the first during highly active pulmonary disease, the second after a median 5 mo (3 to 10 mo) of immunosuppressive treatment. At the first examination three patients had extensive infiltrates or consolidations, two had nodular lesions, and one had ulcerative and granulomatous bronchitis in association with a lymphocytic BAL cell profile. Four patients were treated daily with oral cyclophosphamide (100 to 150 mg/d) plus prednisolone (50 mg/d initially, tapered to 0 to 5 mg/d subsequently) and two with low-dose methotrexate (15 to 22.5 mg/wk) plus corticosteroid, which resulted in partial or complete remission in all six patients (19). The ESR dropped from a median 68 mm/h (40 to 100 mm/h) to 17 mm/h (14 to 22 mm/h) and the C-reactive protein from 9.5 mg/dl (2.1 to 15.8 mg/dl) to 0.5 mg/dl (0.4 to 0.6 mg/dl). At follow-up examination infiltrates had cleared, consolidations showed substantial regression, and bronchoscopy showed the airway lesion had resolved.

Two control groups were included. The sarcoidosis group comprised six women and six men with a median age of 33 yr (29 to 39 yr) with highly active, untreated pulmonary disease according to established criteria (20). Roentgenologically, nine patients had Stage I pulmonary disease, two patients had Stage II disease, and one patient had a normal CXR. Eight patients also had arthritis and erythema nodosum (Loefgren's syndrome), and four had pulmonary disease in association with arthritis. All 12 patients had a lymphocytic BAL profile, the median lymphocyte count being 47% (30 to 53) (Table 2). Immunotyping disclosed preferential elevation of the CD4+ cells and transbronchial biopsy disclosed a granulomatous histopathology in all 12 patients. The healthy control group comprised five women and four men with a mean age of 33 yr (29 to 39 yr). Six of these had been referred to our department for evaluation of suspected rheumatic disease and turned out to be free of inflammatory systemic or pulmonary disease; the remaining three were healthy laboratory personnel. All nine control subjects had a normal CXR, normal lung function, and a normal BAL cell profile. All patients and the control subjects were nonsmokers and free of clinical or microbiologic evidence of infection. Written consent was obtained from all normal volunteers, and the study protocol was approved by the institutional review board.

Bronchoscopy

Fibreoptic bronchoscopy was performed after local anaesthesia with lidocaine and premedication with atropine and a morphine antitussant. BAL was preferentially performed in roentgenologically abnormal lung segments in the patients with WG and in the middle lobe or the lingula in the patients with sarcoidosis and the control subjects. Twelve fractions of 20-ml sterile saline 0.9% were instilled and aspirated in a single segment. The fluid recovery was at least 60%, and all materials were placed immediately on ice. The first two fractions were separated and the third to twelfth fractions were pooled and further analyzed. Aliquots of the material were examined for conventional bacterial pathogens, acid-fast bacteria, Legionella species, Chlamydia species, Mycoplasma species, and Pneumocystis carinii, and any infected material was omitted from further study.

BAL Cell Analysis

The lavage material was filtered through surgical gauze, and centrifuged at 4° C, and the cell pellet was resuspended in RPMI 1640 medium (GIBCO, Eggenstein, FRG) to a density of 10⁶ cells/ml. Cytospin preparations were prepared with a Shandon II cytocentrifuge (Shandon Products, Cheshire, UK) and stained according to May-Giemsa-Gruenwald. The cell differential was assessed microscopically by counting 400 cells. Normal reference values in this laboratory are  15% lymphocytes,  5% neutrophils, and < 1% eosinophils.

Measurement of Soluble Cell Products in the BALF

Determinations of soluble cell products were done in concentrated BALF. The cell-free BALF was subjected to pressure filtration under a nitrogen athmosphere using Amicon YM10 membranes with a molecular weight cutoff of 10 kD (Millipore Corp., Eschborn, FRG). The mean concentration factor was 37, and solute concentrations in the original BALF were calculated by dividing measured concentrations by the individual concentration factor. MPO was measured using the Myeloperoxidase Capture ELISA from Calbiochem-Novabiochem Corp. (San Diego, CA) according to the recommendations of the manufacturer. The threshold of detection in the BALF concentrate was 1.2 ng/ml. ECP was measured using of the Unicap ECP-fluoroimmunassay method (Pharmacia, Freiburg, FRG) according to the recommendations of the manufacturer. The detection threshold in the BALF concentrate was 2 ng/ml. PEROX was measured using the K-blue assay as described by Segelmark and colleagues (21). Briefly, 50 µl of concentrated BALF (1:2 diluted in PBS) was mixed with 50 µl of the tetramethylbenzidine and hydrogen peroxidase-containing substrate/chromogen mixture (Neogen Corp., Lexington, KY) and the absorbance was measured at 660 nm after 30 min. The specificity of the method was verified by measuring progressive dilutions of BALF specimens producing values that paralleled the standard curve. Moreover, purified MPO (Calbiochem Corp., Bad Soden, FRG) was diluted serially and measured using the K-blue assay and the aforementioned capture ELISA for MPO. This gave a close correlation between PEROX and MPO measurements. The detection threshold of the PEROX assay was 0.026 U/ml.

Free PR3 (fPR3) was measured by a highly specific capture ELISA established in this laboratory according to the recommendations of Baslund and colleagues (22). In brief, a purified mixture of four monoclonal antibodies against PR3 (WGM1, 2, 3, and 12.8) 2 mg/ ml in carbonate buffer was coated on microtitre plates (Nunc immunoplate; Nunc, Roskilde, Denmark), followed by blocking with phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin (Serva, Heidelberg, FRG) and 0.05% Tween 20 (sample buffer). BALF samples were diluted in PBS as described above to a final dilution of 1:2 and incubated for 1 h at room temperature on the coated plates. After thorough rinsing, 100 ml affinity-purified rabbit anti-PR3 (kindly provided by Dr. J. Wieslander, Statens Seruminstitut, Copenhagen, Denmark) diluted 1:500 in sample buffer was added, and incubated for 1 h. This antibody binds fPR3, but not PR3, complexed with 1-antitrypsin (1-AT) or ANCA (22). Plates were developed by adding horseradish-peroxidase-labeled goat antirat IgG (Dako, Hamburg, FRG) diluted 1:1,000 to the sample buffer and the subsequent addition of peroxidase substrate consisting of 0.2 mg/ml o-phenylendiamine (Sigma, Deisenhofen, FRG) in 0.05 M phosphate buffer. The enzyme reaction was stopped by adding 0.5 M H₂SO₄, and the optical density was measured at 492 nm. The detection threshold in the BALF concentrate for fPR3 was 7.8 ng/ml.

PR3/1-AT complex levels were measured by adopting the ELISA method of Baslund and colleagues (22) with minor modifications. Microtitre plates (Nunc immunoplate) were coated overnight with the same four monoclonal anti-PR3 antibodies that were used in the fPR3 assay, followed by blocking with the aforementioned blocking/sample buffer. Samples were diluted as described above and incubated for 1 h at room temperature. Bound complexes were detected with alkaline-phosphatase-labeled antihuman 1-AT IgG (Merck, Darmstadt, FRG) diluted 1:1,000 in sample buffer and plates were developed by adding p-nitrophenyl-phosphate disodium 1 mg/ml in 1 M diethanolamine as substrate for the enzyme reaction. The optical density was read at 405 nm. A standard curve was established by use of serial dilutions of PR3/1-AT complexes prepared by mixing PR3 with 1-AT at a ratio of 1:10 in PBS with 0.05% Tween 20 (22). Complex formation was verified by immune electrophoresis against anti-PR3 and anti-1AT and was further confirmed by measuring the inhibition of PR3 enzymatic activity using the substrate MeO-Suc-Ala-Ala-Pro-Val-pNA (Sigma). Concentrations for the PR3/1-AT complex are expressed as the concentration of PR3 bound in the complex. The threshold of detection in the concentrated BALF was 5 ng/ml. sIL-2R was measured with a commercial assay (Milenia; DPI Biermann, Bad Nauheim, FRG) according to the instructions of the manufacturer. The detection threshold in the concentrated BALF was 1 U/ ml. All assays were run in duplicate.

Statistics

Data are presented as median and the 25th and 75th percentile of the median. Values for the BAL cell profile are expressed as percent of the total BAL cells. Concentrations of BALF solutes are presented as concentrations in the original lavage fluid. The Kruskall-Wallace test was used for testing differences between groups for statistical significance and the Mann-Whitney U test for testing pairwise differences. Correlations between cell numbers and solute concentrations were tested for statistical significance by Spearman's rank test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BAL Cell Profile in WG

An abnormal cell profile was found in 13 patients with WG (Table 2). Seven patients had BAL neutrophilia in the range of 6 to 18% (normal reference value  5%). Roentgenologically, six of these patients had ill-defined infiltrates or extensive consolidation, one had a nodular lesion. Six patients with WG had BAL lymphocytosis in the range of 18 to 42% (normal reference value  15%) concomitantly with normal BAL neutrophils. Four of these six patients had nodular lesions, one had a focal infiltrate, and one had ulcerative bronchitis concomitant with a normal CXR. The remaining six patients had a normal cell profile, three of these showing consolidation or nodular lesions. Low-grade elevation of the eosinophils in the range of 1 to 3% was found concomitantly with neutrophilia in three patients and in association with lymphocytosis in one patient.

MPO, PEROX, and ECP in the BALF

Compared with patients with sarcoidosis and control subjects, the patients with WG were characterized by elevated levels of granulocyte products in the BALF. The median PEROX activity in the BALF was 4.47 U/ml (1.34 to 19.55 U/ml) in WG, 0.48 U/ml (0.32 to 0.87 U/ml) in sarcoidosis, and 0.20 U/ml (0 to 0.37 U/ml) in control subjects. The higher values in the patients with WG differed significantly from the values in the two other groups (p < 0.001) (Figure 1). The same was true for MPO levels, which were 0.30 ng/ml (0.08 to 0.85 ng/ml) in patients with WG, 0.12 ng/ml (0.02 to 0.22 ng/ml) in patients with sarcoidosis, and 0.13 ng/ml (0.03 to 0.17 ng/ml) in control subjects (WG versus sarcoidosis, p = 0.012; WG versus controls, p = 0.018) (Figure 1). PEROX and MPO levels correlated significantly with the BAL neutrophil count in patients with WG (R = 0.571, p = 0.011 and R = 0.610, p = 0.005, respectively) (Figure 2), suggesting that the BAL neutrophils were an important source of PEROX and MPO in the BALF. A significant difference between the three groups was also found in the ECP levels (Figure 3). Twelve of the 19 patients with WG had measurable ECP, whereas the ECP was below the threshold of detection in all patients with sarcoidosis and control subjects (WG versus sarcoidosis, p = 0.003; WG versus control, p = 0.019).

View larger version (27K): [in this window] [in a new window]   Figure 1.   Myeloperoxidase level and peroxidase activity in the BALF in Wegener's granulomatosis, sarcoidosis, and controls.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Correlation between the myeloperoxidase level in the BALF and the BAL neutrophil count in Wegener's granulomatosis.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Levels of eosinophil cationic protein and soluble interleukin-2 receptor in the BALF in Wegener's granulomatosis.

fPR3 and PR3/1-AT Complexes in the BALF

fPR3 and PR3/1-AT levels in BALF did not differ between the three groups. Measurable fPR3 was found in seven of the 19 patients with WG, five of the 12 patients with sarcoidosis, and one of nine control subjects; the intergroup differences falling clearly short of statistical significance (Figure 4). Measurable concentrations of PR3/1-AT were detected in all subjects. Medians were 2.47 ng/ml (1.47 to 3.69 ng/ml) in the patients with WG, 2.56 ng/ml (1.72 to 3.53 ng/ml) in the patients with sarcoidosis, and 1.69 ng/ml (1.02 to 2.30 ng/ml) in the control subjects and did not differ significantly between the study groups (Figure 4). Moreover, no correlation was found between the BAL neutrophil count and fPR3 or PR3/1-AT levels in the patients with WG.

View larger version (21K): [in this window] [in a new window]   Figure 4.   Levels of free proteinase 3 and proteinase 3 complexed with 1-antitrypsin (PR3/1-AT) in the BALF in Wegener's granulomatosis.

sIL-2R in the BALF

The highest sIL-2R levels were found in the patients with sarcoidosis. The median was 3.05 U/ml (1.35 to 11.55 U/ml) and this exceeded significantly the level in the control subjects (p < 0.001), only four of which had measurable sIL-2R (Figure 3). The level in the patients with WG was 1.76 U/ml (0.16 to 3.89 U/ml) and this was also higher than the level in the control subjects (p = 0.021) but not significantly different from the level in the patients with sarcoidosis. A significant correlation was found between the BAL lymphocyte count and the sIL-2R level in the patients with sarcoidosis (R = 0.759, p = 0.004) but not in the patients with WG.

Effect of Treatment on Cells and Cell Products in the BALF

Three of the six patients with WG followed prospectively had elevated BAL neutrophils during highly active disease. In all three patients the neutrophil count normalized during treatment, but concomitantly the BAL lymphocytes increased (Figure 5). Two of the remaining three patients had a lymphocytic BAL cell pattern and one had a normal cell profile during highly active disease, these three patients showing either minor increases or decreases in the lymphocyte count in response to the treatment. Clinical improvement was associated with a sharp drop in PEROX and MPO in the BALF in five of the six patients (Figure 6). The one patient with low PEROX and MPO levels before treatment showed small increases in these values in response to treatment, but these changes were unimpressive compared with the changes in the other five patients. The sIL-2R levels in BALF showed no consistent pattern of response. Increases as well as decreases or persistently undetectable values were observed and no correlation was found between changes in the BAL lymphocyte count and sIL-2R levels.

View larger version (13K): [in this window] [in a new window]   Figure 5.   Effect of clinically effective immunosuppressive treatment on the BAL cell profile in Wegener's granulomatosis.

View larger version (18K): [in this window] [in a new window]   Figure 6.   Effect of clinically effective immunosuppressive treatment on the myeloperoxidase level, peroxidase activity, and the soluble interleukin-2 receptor level in the BALF in Wegener's granulomatosis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MPO is a secretory enzyme of neutrophils involved in the extracellular generation of highly reactive halide derivatives (23). Together with other cytoplasmic constituents it is released in response to various proinflammatory stimuli. Comparison of MPO levels in blood and BALF suggests that the bulk of the MPO at inflammatory tissue sites is of local origin (24). Elevated levels of MPO in BALF have been found in a number of lung diseases with prominent neutrophil involvement, including bacterial pneumonia (25), chronic bronchitis (26, 27), idiopathic pulmonary fibrosis (28, 29), and interstitial lung disease caused by systemic sclerosis (29). The present study found elevated MPO levels in the BALF also in active pulmonary WG (Figure 1), a finding that reflects the presence of substantial numbers of degranulating neutrophils in the lower respiratory tract in this disorder. The correlation between the BAL neutrophil count and MPO levels in BALF suggests that the neutrophils present on the epithelial surfaces are an important source of MPO in BALF. However, elevated MPO levels were also found in patients with elevated lymphocytes but normal BAL neutrophils. This implies that the MPO in BALF originates not only from neutrophils on the epithelial surfaces but also from neutrophils located subepithelially.

The BALF was also found to have high PEROX activity (Figure 1). PEROX levels correlated significantly with the neutrophil count in BALF, which suggests that the major part of the PEROX activity in the BALF does indeed originate from the neutrophils. In support of this conclusion Brouwer and colleagues (13) found substantial numbers of peroxidase-producing neutrophils in cryostat sections of renal tissue affected by glomerulonephritis caused by WG and reported that neutrophils in inflammatory tissue but not unprimed neutrophils from control subjects exhibited PEROX activity. On the other hand, no correlation was found between PEROX and MPO levels in the BALF. One explanation for the lack of a correlation here could be that the MPO was subject to in vivo or in vitro alteration resulting in disproportionate loss of its enzymatic activity and antigenicity. An alternative explanation would be that peroxidases other than MPO contribute to PEROX activity in BALF, one likely source being the eosinophils. Upon stimulation eosinophils release a spectrum of cytoplasmic constituents, among which are ECP, other toxic peptides, and various enzymes, one of which is eosinophil peroxidase (30). ECP as well as eosinophil peroxidase are stored in the eosinophil granules and are physiologically released together (30). Given the elevated ECP levels in the present patients with WG (Figure 3), it is likely that an excess of eosinophil peroxidase also contributed to the PEROX activity measured in BALF.

PR3 is a further secretory product of neutrophils that is stored together with MPO in azurophilic granules (23). In vitro studies have shown that degranulation of neutrophils results in the concomitant release of MPO and PR3 (10). Immunohistologic studies have demonstrated that extracellular PR3 in inflammatory tissues colocalizes with extracellular MPO and other neutrophil cytoplasmic constituents (13, 14), suggesting that PR3 and MPO are also secreted concomitantly in vivo. The present study, however, reveals a discordant behavior of PR3 and MPO in the BALF in active pulmonary WG in that only trace amounts of fPR3 were detected in the presence of high levels of MPO (Figure 4) and no correlation existed between fPR3 and MPO levels.

A possible reason for the low levels of fPR3 is rapid degradation in the tissue, but this is an unlikely explanation given the ready detection of extracellular PR3 in immunohistologic studies (13, 14). An alternative reason can be binding of the strongly cationic PR3 to tissue structures. Immunohistologic studies have demonstrated preferential deposition of extracellular PR3 along basement membranes and on the surface of vascular endothelium (13). It is conceivable that the affinity of PR3 for negatively charged structures is strong enough to curtail its elution from lung tissue by the BAL procedure. A further explanation is the formation of PR3 complexes in the extracellular fluid, resulting in the detection of only small amounts of the free compound, and the present findings support this view. The most important physiologic inhibitor of serine proteinases, which binds and inactivates also PR3, is 1-AT (31). Indeed, the amount of PR3 forming complexes with 1-AT in the BALF of these patients exceeded by far the amount of fPR3 (Figure 4), which indicates that sufficient amounts of 1-AT are available in the BALF to effectively reduce the level of fPR3. In addition to binding by physiologic inhibitors, binding by ANCA may also be an important reason for the low fPR3 levels in these patients. Unfortunately, because of the limited supply of material, ANCA and ANCA immune complexes could not be measured in these patients. An earlier study, however, found that patients with active pulmonary WG do have ANCA in the BALF, but in that study measurement of complexes out of PR3 and ANCA was not attempted (32). Finally, there is the possibility that PR3 is lost during the concentration procedure because of adsorption to the filter or other materials.

The finding of apparently efficient binding of PR3 by 1-AT in the BALF appears to contradict the concept that 1-AT is subject to oxidation at tissue sites with highly active inflammation and an excess of oxidant capacity, resulting in reduced binding capacity and an increase in the amount of unbound proteinases (33). A potential explanation for the apparently efficient binding of PR3 to 1-AT in the present patients is the focal distribution of inflammatory tissue lesions in WG lung disease (3). Given the large tissue volume sampled by segmental BAL, the BALF is likely to contain a mixture of material extracted from severely affected tissue as well as only marginally affected tissue, the latter containing sufficient amounts of 1-AT to bind most of the PR3. With respect to the quantitative relationships between PR3 and 1-AT, therefore, BALF need not necessarily reflect the situation in the tissue.

Lymphocytes respond to a spectrum of stimuli by upregulating their membrane receptor for IL-2. This is associated with shedding of the soluble form of the receptor (sIL-2R) (34). The serum level of sIL-2R correlates with clinical activity in lung diseases with strong lymphocyte involvement, a prominent example of which is sarcoidosis (35). Elevated serum levels of sIL-2R have also been found in WG and were shown to vary with clinical disease activity (36). Complementary to this were high sIL-2R levels in the BALF in lung diseases with massive lymphocyte involvement. In a comparative study sarcoidosis and hypersensitivity pneumonitis, both of which feature a prominent lymphocyte component in BAL and tissue, were found to have the highest levels of sIL-2R (37). Tissue infiltrates in chronic types of lung lesions caused by WG also include a substantial lymphocytic component (3, 15). Moreover, BAL lymphocytosis is a common finding in WG (38), which prompted us to examine the diagnostic information contained in sIL-2R measurements in BALF in this disorder. The sIL-2R levels in these patients were significantly higher than levels in control subjects, but they fell substantially below levels in patients with sarcoidosis (Figure 3). A correlation between BAL lymphocyte counts and sIL-2R levels was found in sarcoidosis but not in WG. Moreover, clinically effective immunosuppressive treatment of WG resulted in no consistent response from sIL-2R levels and no correlation was found between treatment-induced changes in sIL-2R levels and the BAL cell profile. These findings imply that activated lymphocytes are present in the lower respiratory tract in WG. However, the relationship between sIL-2R levels in BALF and the local disease status appears to be more complex than the relationship between blood sIL-2R levels and systemic disease activity.

In conclusion, the presence of extracellular neutrophil products at affected tissue sites and the correlation between levels of these products in the extracellular fluid and clinical disease activity support the view that neutrophils are indeed an important effector cell population in WG lung disease. The present data also suggest that oxidative injury is an important aspect of neutrophil-mediated lung injury, whereas they do not allow a firm assessment as to the role of proteolytic enzymes in this setting. The PR3 in the BALF of these patients was found to be mostly bound with 1-AT, but because of the focal distribution of inflammation in WG lung disease and the large surplus of 1-AT in the extracellular fluid, the balance between 1-AT and PR3 in the BALF may not adequately reflect the situation at highly inflammatory tissue sites. BALF levels of MPO and PEROX, but not of PR3 and sIL-2R, changed in parallel with the clinical status and may thereby help in assessing local disease activity. Whether this can be exploited clinically in the monitoring of treatment effects needs to be addressed in a prospective study.