INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although the Registry of the International Society for Heart and Lung Transplantation (ISHLT) reported a significant improvement in long-term survival of patients after (heart-)lung transplantation [(H)LTX] in 1997 (1), overall prognosis of these patients is still unsatisfactory. The main causes of long-term mortality are obliterative bronchiolitis (OB) and/or infection (1, 2). Deterioration of lung function is also the major factor limiting the quality of life of these patients.

Although pathogenesis of OB is as yet unclear, it is generally acknowledged that it is a manifestation of chronic graft rejection (2). In agreement with this concept, in some patients augmentation of immunosuppression is at least temporarily successful (3). Previously described risk factors for OB include severe and frequent episodes of acute rejection and infection with cytomegalovirus (CMV) (2, 4). Because of the histologic appearance of OB as a patchy process, sensitivity of transbronchial biopsies for diagnosing OB is limited. Imaging techniques are commonly not diagnostic either (5). Thus, the diagnosis of OB is usually based on spirometry. A diminution of FEV₁ to less than 80% of a baseline value obtained 3 mo after transplantation in the absence of other causes for graft dysfunction is defined as bronchiolitis obliterans syndrome (BOS) (2, 6).

Neutrophilia was repeatedly observed in the bronchoalveolar lavage fluid (BALF) of patients after (H)LTX (7). In addition, infiltration of neutrophils into the bronchial epithelium has been detected in patients with higher degrees of active airway damage (7). Neutrophils are capable of causing cause severe damage to the lung tissue by releasing toxic proteases and reactive oxygen species if not counterbalanced by the antiprotease/antioxidant screen of the lung (8). Based on this background, a causal relationship between neutrophilia and the development of BOS has been proposed (9).

We analyzed cell counts in BALF samples of patients after (H)LTX and performed a correlation analysis to the presence of BOS and subsequent changes in FEV₁. To examine the possible role of alterations in antiprotease defense and oxidant activity for the pathogenesis of BOS we quantified: (1) the activity of BALF against neutrophil elastase (NE), the most important neutrophil protease (8); (2) the concentration of its complex with the principal antiprotease of the human lung, ₁-protease inhibitor (API) (8) in BALF; (3) the content of the latter in BALF and serum; (4) the concentration of secretory leukocyte protease inhibitor (SLPI), the major antiprotease of the bronchial system (8, 10, 11), in BALF; (5) the content in oxidized methionine [Met(O)] and the concentration of myeloperoxidase (MPO) as markers for oxidant activity.

All values were referenced to the volume of respiratory epithelial lining fluid (ELF) as determined by the urea method (12). As a comparison all tests were done in parallel with BALF samples obtained from healthy volunteers.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Samples

Seventy-two BALF samples obtained from 33 patients after (H)LTX and 17 BALF samples obtained from 17 healthy control subjects were evaluated. Pulmonary function testing including FEV₁ was performed in all control subjects and in most patients a few hours before bronchoscopy; otherwise test results were used that had been obtained shortly before. According to ISHLT criteria (6) three groups were established: 16 samples from patients with BOS (BOS-positive), 32 samples from patients without BOS (BOS-negative), and 24 samples from patients in the first 3 mo after transplantation. The staging of BOS into stage I-III as proposed by Cooper and coworkers (6) was not feasible owing to sample size. To examine the subsequent changes of graft function after BAL, the FEV₁ was determined 3 and 6 mo after bronchoscopy and referenced to the FEV₁ at the time of bronchoscopy (FEV_13m% and FEV_16m%).

Bronchoalveolar lavage (BAL) was performed using a flexible bronchoscope with instillation of 100 ml of isotonic saline each into three different lobes in the control group; in several patients a smaller volume (100 to 200 ml) had to be used because of reduced respiratory function. The volumes recovered did not allow us to perform all tests with all samples. In 47 cases, transbronchial biopsies were performed and graded according to international standards (7). Infection being a possible cause of neutrophilia after (H)LTX, BALF samples underwent thorough microbiological analysis (bacteria, fungi, viruses: CMV, [RSV], human herpes virus-6 [HHV-6], adenovirus, and influenza virus).

Estimation of ELF

The volume of ELF was determined as described by Rennard (12). To check for urea diffusion resulting from leakage of the alveolar-capillary barrier, correlation analysis between urea_BALF%serum and albumin_BALF%serum was performed, albumin diffusion being known to be enhanced in various settings with an increased permeability of this barrier (13). There was no significant correlation. Because increased albumin diffusion was suspected to occur chiefly in the BOS-positive group, the volumes of ELF in the four groups were compared by t test to exclude a possible bias (p = NS).

Cell counts were expressed per ml BALF. To test for bias, the number of neutrophils per ml BALF and per ml ELF of 48 samples were compared by correlation analysis (r = 0.97). Generally, BALF and ELF concentrations of all measured parameters were highly correlated.

Anti-NE activity in BALF. Anti-NE activity was measured by a titration assay as described by Ogushi and coworkers (14) using the NE-specific substrate N-methoxysuccinyl-ala-ala-pro-val-p-nitroanilide (Sigma, St. Louis, MO). Active NE in BALF resulted in negative values for the anti-NE activity.

API in BALF. The concentration of API in BALF was determined by use of an ELISA modified after Michalski and coworkers (15). Purified human API (Athens Research and Technology Inc., Athens, GA) was used as a standard. For coating, a goat polyclonal anti-API antiserum (Cappel Research Products, Durham, NC), as second antibody a rabbit anti-API antibody (Boehringer, Indianapolis, IN) was used.

NE-API complex in BALF. The concentration of NE-API complex was measured using a commercially available ELISA (PMN-Elastase-IMAC; E. Merck, Darmstadt, Germany) as described elsewhere (16).

SLPI in BALF. SLPI in BALF was measured with an ELISA kit (R&D Systems, Minneapolis, MN), using a monoclonal anti-SLPI antibody for coating and a polyclonal antiserum linked to horseradish peroxidase as detection system.

Met(O) in BALF. Met(O) was quantified using the cyanogen bromide (CNBr) reaction as published elsewhere (17, 18). Briefly, methionine is quantitatively transformed to homoserine and homoserinolactone by CNBr; the concentration of Met(O) is not influenced by this reaction. Met(O) is then reduced to methionine and the proteins already cleaved by CNBr are hydrolyzed in the presence of 5 mM dithiothreitol (DTT) and 6 M HCl at 110° C for 24 h. Finally, reduced methionine, homoserine, and homoserinolactone are quantified by amino acid analysis.

MPO in BALF. MPO concentration was measured by ELISA (Bioxytech, Bonneuil, France) using a monoclonal coating antibody and polyclonal goat antibodies coupled to biotin and an avidin-peroxidase complex for detection.

Albumin in BALF and serum, API in serum, and IgG in BALF and serum. For these parameters a nephelometer (Nephelometer 100; Behring, Marburg, Germany) was used; the concentrations of the complexes of the antigen with the added antibody serum (Behring OSAL 14/15, OSAZ 14/15 and OSAS 14/15, respectively) are measured by analyzing the light scattered from a beam of light passing through the solution.

Statistics

Statistics were calculated using SPSS for Windows 5.1 and 6.1 (SPSS Inc., Chicago, IL). Data are generally expressed as mean ± SEM. Data were compared using the t test, one-way analysis of variance (ANOVA), and correlation analysis according to Pearson. A p value  0.05 was considered significant.

Normal distribution of the data was tested using the Kolmogorov-Smirnov test against a normal plot and assumed if p > 0.05. Otherwise, a log-normal distribution was assumed and the data were retested after transformation using the natural logarithm. The test of equality of variances was performed by the Levene test. All results were checked for possible bias by repeating them with only one sample per patient (19).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell Counts in BALF

The total cell count per ml BALF was significantly increased (p < 0.01) in all patient groups compared with control subjects. Patients with BOS had cell counts higher than all other groups chiefly due to a strikingly increased neutrophil count (Table 1, Figure 1). There was no significant correlation between cell counts and the time between transplantation and lavage.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Neutrophil count per ml BALF. All patient groups demonstrate significantly increased neutrophil counts compared with the control subjects (p < 0.001); patients with BOS show a significant increase in comparison to patients without BOS (p < 0.001).

Comparing the neutrophil count with the FEV_13m% and FEV_16m% (Figure 2) there was a significant correlation of r = 0.49 (n = 47; p < 0.001) and r = 0.56 (n = 39; p < 0.001), respectively. Repeating the correlation analysis with only one BALF sample per patient, correlations were similar with r = 0.46 (n = 19; p = 0.05) and r = 0.51 (n = 14; p = 0.06).

View larger version (20K): [in this window] [in a new window]   Figure 2.   FEV1 measured 3 mo (FEV13m%) and 6 mo (FEV16m%) after bronchoscopy in comparison to the neutrophil count in BALF. There is a significant inverse correlation of r = 0.49 (n = 47; p < 0.001) and r = 0.56 (n = 39; p < 0.001), respectively.

In 47 cases transbronchial biopsies (4.17 ± 0.39 biopsies per bronchoscopy; maximum n = 10, minimum n = 1) were performed. Biopsy samples were evaluated according to the revised ISHLT working formulation (7). Overall, 12 of 47 biopsy samples showed acute rejection ISHLT grade A. Only one of these patients was classified as suffering from BOS ( 3 mo after transplantation n = 4; BOS-negative, n = 7; BOS-positive, n = 1). Probably as a consequence of the limited number and size of the biopsies OB could not be detected in any of the biopsies.

Of the total of 72 BALF samples, 50 could be analyzed for microbes. The remaining 22 samples could not be evaluated because of limited material. In 22 of the 50 analyzed samples potentially relevant bacteria were found. Most frequent was colonization with Pseudomonas aeruginosa, Xanthomonas maltophilia, and Staphylococcus aureus. Neutrophil counts were significantly (p < 0.001) increased in BALF samples with positive cultures for bacteria. In patients with positive cultures, FEV_13m% was significantly (p < 0.001) diminished, whereas FEV_16m% was not significantly influenced. From the 16 BALF samples from BOS patients 13 showed positive culture results, whereas from the 17 samples from patients without BOS 12 had negative cultures. For fungi and viruses, statistical comparisons were not performed owing to the small sample size.

Anti-NE Activity in ELF

In the samples obtained from control subjects, anti-NE activity was uniformly positive with little variability (Table 2, Figure 3). There was a trend to increased anti-NE activity in samples obtained in the first 3 mo after transplantation with high variability. Two samples showed negative values, signifying an overwhelming elastase load in ELF. More than 3 mo after transplantation, anti-NE activity was close to normal in most patients without BOS, though some samples featured negative values. Free NE was frequently found in patients with BOS, who showed altogether a distinct drop in anti-NE activity (Table 2, Figure 3).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Anti-NE activity in ELF. Anti-NE activity was determined using a titration assay. Patients with BOS feature a distinct, although not statistically significant drop in anti-NE activity.

API

In the group of samples obtained in the first 3 mo after transplantation, all samples except one outlier with an increased concentration showed normal values (Table 2). Several patients without BOS had distinctly elevated API levels. In patients suffering from BOS there was a trend for higher API concentrations without reaching statistical significance.

Serum values of API were significantly (p < 0.01) elevated in patients more than 3 mo after transplantation. There was no difference between patients with and without BOS and no significant correlation between serum API and API in ELF. Serum transudation caused by leakage of the alveolar-capillary barrier seemed the most likely explanation for this increase in concentration, because API has a similar molecular weight (52 kD) as albumin (66 kD), and because there are reports of increased API concentrations in other entities with damage to the alveolar-capillary barrier (20). To test this hypothesis, correlation analysis between API_ELF%serum and albumin_ELF%serum as well as IgG_ELF%serum was performed and showed significant correlations of r = 0.67 (n = 30; p < 0.001) and r = 0.70 (n = 29; p < 0.001), respectively. There was no significant correlation between albumin_ELF%serum and anti-NE-activity (n = 29) or Met(O) (n = 16).

NE-API Complex

In the first 3 mo after transplantation there was a 2-fold increase in the concentration of the NE-API complex without reaching statistical significance (Table 2). Samples obtained from patients more than 3 mo after transplantation showed significantly elevated levels with the highest values in the group of patients with BOS (p < 0.01 for all comparisons).

SLPI

Patients without BOS demonstrated essentially normal SLPI values compared with control subjects. In contrast, patients with BOS had markedly decreased SLPI concentrations (p < 0.05; Table 2, Figure 4).

View larger version (13K): [in this window] [in a new window]   Figure 4.   Concentration of SLPI in ELF. SLPI concentrations were measured by ELISA. All patients with BOS demonstrate decreased SLPI concentrations, the differences to all other groups being statistically significant (p < 0.05 for all comparisons).

Met(O)

The Met(O) content was elevated in all patient groups with the highest values being found in patients with BOS (p < 0.001 in comparison to patients without BOS and control subjects; Table 3, Figure 5). There was a significant correlation of r = 0.74 (n = 19; p < 0.001) between SLPI concentration and the percentage of Met(O).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Met(O) in ELF. Met(O) content was analyzed using the cyanogen bromide reaction. In samples obtained from patients with BOS, Met(O) was significantly elevated (p < 0.001 in comparison to patients without BOS and control subjects).

MPO

In the group of patients  3 mo after transplantation there was no significant difference in the MPO concentrations compared with control subjects (Table 3). In contrast, patients with and without BOS featured significantly increased MPO concentrations (p < 0.05 and p < 0.01 compared with control subjects).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study confirms that the influx of neutrophils observed after (H)LTX is most pronounced in patients with BOS. As a consequence we found considerable alterations in the protease-antiprotease balance and increased oxidant activity.

The neutrophil numbers in BALF obtained from patients classified as suffering from BOS are significantly correlated to the loss of FEV₁ over a 6-mo period. In the case that acute complications at the time of BAL-like episodes of acute rejection or infection are relevant for the observed neutrophilia and neutrophil activation, after resolution a subsequent improvement of lung function would be expected. Instead, the pulmonary function of our patients with neutrophilia went further downhill, suggesting an underlying progressive process. This is further supported by the fact that only in one patient considered to suffer from BOS histopathological changes compatible with acute rejection were found.

We found a significant correlation between the occurrence of positive cultures for bacteria and the progress of graft dysfunction as well as the number of neutrophils. Patients with OB do often have positive culture results (1). Whether infection is a cause for neutrophilia or just a sequel of BOS, e.g., due to the treatment of these patients with augmented immunosuppression or an increased susceptibility, cannot be concluded from the present data. However, even if neutrophilia should turn out to be a mere sequel of infection, the potential of neutrophil products to cause damage to the lung tissue must be taken into consideration as a contributing factor of its own.

A diminished anti-NE activity could be found in patients with BOS with elastase activity overwhelming the antiprotease defense in 50% of the cases (5 of 10 samples). Free elastase activity has been shown to be present also in other pathologic conditions with increased neutrophil numbers such as API deficiency (14), cystic fibrosis (21), or after ozone challenge of healthy subjects (11). A part of the increased NE load can be inhibited by API as shown by elevated concentrations of API and the NE-API complex, a phenomenon we showed to be the result of increased leakage of the alveolar capillary barrier. Patients with cystic fibrosis have normal API concentrations in BALF in spite of similar neutrophil counts and an even higher NE activity (21). Neutrophilia and increased NE activity could also be found in BALF of healthy volunteers with acute ozone-induced airway inflammation as well as in patients with bacterial pneumonia; under these circumstances a parallel rise of API in BALF was observed (11, 20). In addition, in a study by Kindt and coworkers BALF samples obtained from patients with idiopathic bronchiolitis showed a dramatic increase of the neutrophil number (54 ± 10% of total cells). In patients treated successfully with prednisolone a significant decrease of the BALF neutrophilia was observed (22). Nunley and coworkers evaluated BALF samples from 52 lung-transplanted cystic fibrosis patients. The reported results support our findings: from 14 patients showing free elastase activity in BALF 10 were suffering from BOS, whereas only eight of 38 patients without BOS had free elastase in BALF. In addition, there was a remarkable association between elastase activity and infection with P. aeruginosa (23).

The leakage of the alveolar capillary barrier is apparently not solely caused by neutrophil products. Because the correlation of neutrophil counts with the decrement of FEV₁ over time abrogates a predominance of samples with reversible acute complications, plasma leakage is probably not exclusively an acute process. The hypothesis arises that intercellular junctions may be hampered as a consequence of graft dysfunction. In support of this assumption signs of fibroproliferative remodeling were found in OB, as indicated by an increased production of transforming growth factor 1 and 2 as well as platelet-derived growth factor (24, 25).

The discrepancy between API concentration and anti-NE activity (Table 2) indicates a high percentage of inactive API. The elevated percentage of Met(O) in these patients suggests a major role for oxidative inactivation of API as a cause. Furthermore, the presence of normal or even elevated API concentrations in ELF in our study does not seem to be inversely correlated to the loss of pulmonary function. This could indicate a more important role for other protease inhibitors, primarily SLPI, than previously thought. SLPI is considered to be the major antiprotease of the bronchial tree.

We observed an 85% reduction of SLPI concentrations in patients with BOS. Interestingly, the average concentration of SLPI in ELF of patients with cystic fibrosis, who show a considerably higher NE activity in ELF and higher neutrophil counts, is approximately 4-fold higher than the average SLPI concentration in our study (21). The factors contributing to the "loss" of SLPI in OB may be: (1) Inhibition of SLPI production as a consequence of epithelial damage caused by perfusion deficits and/or OB. The factors hindering SLPI production obviously overwhelm the stimuli of SLPI secretion known to be present in patients after (H)LTX, including increased NE activity, steroid medication (26), and probably increased tumor necrosis factor-alpha secretion by macrophages (27). (2) An increased consumption of SLPI by proteolytic and oxidative inactivation (10, 28, 29). A major role for reactive oxygen metabolites is suggested by the inverse correlation between SLPI concentration and Met(O) content.

In all patients after (H)LTX, a significant increase of oxidative stress as shown by an elevated Met(O) content could be found, values in patients with BOS being even higher. An increased oxidant burden in patients with BOS is also suggested by the elevated concentrations of MPO. The relative weakness of the correlation between Met(O) and neutrophil count in ELF (r = 0.57, n = 15; p < 0.05) is probably the result of an increase in antioxidant capacity caused by the influx of oxidant scavenging proteins such albumin (30) and API. An increased influx of other antioxidants such as transferrin, ceruloplasmin, catalase, and superoxide dismutase (31) may also occur (30). Possible effects of oxidant damage include cytotoxicity, increased permeability of cellular barriers and inactivation of surfactant (32). In addition, both API and SLPI can be inactivated by oxidation of the methionine residue in their active center (8).

In conclusion, our data suggest that neutrophils and their released products may be involved in the development of BOS. The most important finding in our view is the dramatically reduced SLPI concentration, considering the role of SLPI for the defense of the bronchial epithelium. Further studies are needed to elucidate the exact role of neutrophils and alterations of the physiologic defense mechanisms for the pathogenesis of OB.