INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The biological mediator nitric oxide (NO) has important physiological functions within the lung, especially the pulmonary vascular bed (1). It has also been implicated in the pathogenesis of airways disease (1, 2). NO is synthesized from L-arginine within many cells in the respiratory tract by nitric oxide synthases (NOS) in two isoforms, constitutive (cNOS) and inducible (iNOS) NO synthase (3). Constitutive NOS is constantly active and produces small amounts of NO (4). Inducible NOS is capable of producing large amounts of NO in response to proinflammatory cytokines such as interleukin 1 and tumor necrosis factor- (5).

In the gaseous phase, NO is relatively stable (6), so that NO produced in the respiratory tract can be measured in exhaled air (7). Exhaled NO (eNO) levels are increased in conditions associated with airway inflammation such as asthma (8, 9) and bronchiectasis (10) and during upper respiratory tract infections (11). Exhaled NO levels have been proposed as a noninvasive measure of respiratory system inflammatory disease (8). Nonetheless, until recently there has been little evidence to directly relate eNO levels with the severity of airway inflammation.

Lung transplantation (LT) is an established form of treatment for patients with severe lung or pulmonary vascular disease (12). The development of post-lung transplant obliterative bronchiolitis (OB) is the commonest cause of late graft failure (13) but may not be pathologically diagnosed until well established (14). Consequently, a clinical syndrome based on a fall in FEV₁ posttransplant in the absence of any other cause, the bronchiolitis obliterans syndrome (BOS) has been established. Patients are classified into BOS grade 1, 2, or 3 depending on the severity of airflow limitation relative to their best posttransplant FEV₁ (14).

BOS appears to be characterized initially by increasing airway wall and bronchoalveolar lavage (BAL) neutrophilia (15, 16), although this may actually precede the development of physiologically demonstrable airway obstruction or histologically evident airway fibrosis and obliteration (17). Indeed, several authors have found that even in stable lung transplant recipients (LTR), neutrophilic airway inflammation is common and is probably a hallmark of ongoing subclinical allogeneic immune stimulation even in the stable group (15, 16, 18).

We have recently shown that in stable LTR, eNO levels, although not elevated for the group, did directly correlate with both the degree of BAL neutrophilia and with the intensity and extent of iNOS expression in the bronchial epithelium (18). Interestingly, only bronchial epithelial iNOS (not iNOS in the lamina propria) correlated with eNO levels. Thus, in stable LTR, eNO levels appear to reflect bronchial epithelial iNOS, which in turn reflects airway neutrophilia, although cause/effect relationships could not be directly deduced from our cross-sectional data in this group. Exhaled NO levels are increased in BOS (19, 20) raising the possibility that eNO levels may be a surrogate marker of airway inflammation and may potentially precede the development of clinically obvious airway obstruction. However, whether the raised eNO levels seen in BOS really reflect worsening airway inflammation and up-regulation of bronchial epithelial iNOS has remained contentious. This is important, because if a relationship between eNO and airway inflammation is present in patients with BOS, as well as in stable LTR, then this would strengthen the plausibility of serial eNO measurements being useful in predicting the transition from stable LTR to the development of BOS and hopefully at a reversible stage.

This study will test the hypothesis that posttransplant BOS is characterized by frankly increased eNO levels, in direct proportion to the degree of BAL neutrophilia, and to the intensity and extent of iNOS expression in the bronchial epithelium.

Although we are primarily interested in factors involved in the pathogenesis of BOS, other causes of graft dysfunction such as acute vascular rejection (21) and pulmonary infection may effect eNO and be important confounders for interpretation of BOS data. Therefore, we have also studied patients with acute rejection and pulmonary infection, measuring the eNO levels, as well as the relationship between eNO and expression of NOS isoforms in these conditions.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population and Inclusion Criteria

The Alfred Hospital Ethics Committee approved the study and written informed consent was obtained from each patient. Forty lung transplant recipients (LTR) were recruited over a 12-mo period and divided into four groups according to clinical status as outlined below.

Group 1. Stable LTR (n = 20). Patients in this group were clinically stable and at or near their best postoperative lung function (mean percent of best FEV₁ posttransplant 98.6 ± SEM 0.3%). Patients were studied at their first routine surveillance bronchoscopy after enrollment. The eNO level and bronchoscopic results from these patients have been described in detail in a previous publication from our group (18).

Group 2. BOS (n = 8). Patients in this group were diagnosed according to standardized criteria (14). In five patients there was biopsy evidence of obliterative bronchiolitis (17), and in all eight, there was a sustained fall in FEV₁ that could not be explained by an alternate cause. Two patients were classified as having BOS grade 1 (FEV₁ = 66-80% of baseline value), four patients had BOS grade 2 (FEV₁ = 51-65%), and two patients had BOS grade 3 (FEV₁  50%). Four patients were studied at a time of a bronchoscopy performed to investigate a greater than 10% decline in FEV₁ in the preceding 5 wk (diagnostic bronchoscopy), and four were studied at a time of stability (surveillance bronchoscopy). In two patients with stable BOS, a light growth of Pseudomonas aeruginosa was isolated after BAL culture. They did not receive any new antimicrobial therapy, and there were no adverse sequelae. Additional pathology (e.g., AR, or organizing pneumonia) was absent in all biopsies.

Group 3. BI (n = 6). Patients in this group were studied at a time of bronchoscopy performed for investigation of clinical, radiological, and/or physiological deterioration. In five patients, gram staining suggested the presence of significant numbers of bacterial organisms, and in all six patients, organisms were subsequently cultured (Streptococcus pneumoniae [2], Staphylococcus aureas [1], Ps. aeruginosa [2], Stenotrophomonas maltophilia [1]). Antibiotic therapy resulted in a full recovery and return to premorbid lung function. Additional pathology (e.g., AR, organizing pneumonia) was not present in any of the biopsies.

Group 4. AR (n = 6). Patients in this group were studied at a time of bronchoscopy performed for investigation of clinical, radiological, and/or physiological deterioration. In all six, transbronchial biopsies revealed evidence of at least mild AR (21), four had A2 grade rejection and two had A3 rejection. Patients were subsequently treated in a standard manner (21) with 3 d of intravenous methyl prednisolone. Two patients also received empirical antibiotic therapy pending culture results, although no organisms were seen on gram stain nor subsequently cultured. All six made a full recovery with return to premorbid lung function.

Individual patient data from the BOS, BI, and AR groups are summarized in Table 1. Similar data from the stable cohort have been previously published (18).

Bronchoscopic Procedures

At bronchoscopy, bronchoalveolar lavage (BAL), transbronchial lung biopsies (TBB), and endobronchial airway wall biopsies (EBB) were obtained. In single LTR, the bronchoscopy involved only the allograft.

The procedure for obtaining the BAL and its subsequent analysis has been previously described (18).

Immunohistochemistry

Immediately after bronchoscopy, EBB were embedded in OCT and snap frozen in liquid nitrogen-isopentane slurry. The subsequent preparation and analysis of the EBB have been previously described (18). The iNOS (type II NOS isoform) and cNOS (types I and III NOS isoforms) protein levels in the epithelial and subepithelial lamina propria were then determined by immunohistochemistry and scored according to previously described protocols (18, 22).

Briefly, five nonoverlapping high-power fields were counted from each of the two sections of each biopsy. The slides were coded and counted by one blinded observer using a computerized image analyzer (Image Pro Plus 3.0 for Windows; Media Cybernetics, Silver Spring, MD). For NOS protein levels in the lamina propria, previous work from our laboratory has shown that the between-section coefficient of variability is < 12% for all measurements (18). Slides from a subset of patients were recoded and counted on a second occasion. The mean within-observer coefficient of variability was < 10% for all measurements.

The scoring of NOS staining within the bronchial epithelium was based on a previously validated semiquantitative method (22). It was performed by two blinded observers. Previous work in our laboratory has shown that the mean difference ± 2 SDs between the two observers' scores (scale 0-5) for intensity and extent is less than 0.5 ± 0.9 and the mean within-observer coefficient of variation is less than 12% for all measurements (18).

Illustrative examples of the immunohistochemical staining in stable LTR have been published previously (18).

Exhaled Nitric Oxide Measurements

Exhaled NO (eNO) was measured immediately prior to bronchoscopy using a rapidly responding and highly sensitive chemiluminescence analyzer (Siever 270 B; Boulder, CO) with a resolution of 0.3 parts per billion (ppb) of NO and response time of 0.7 s (0-95% rise time). The exhaled gas sampling rate was 250 ml/min for all measurements. Where relevant, all measurements were made before the commencement of treatment with intravenous methyl prednisolone. The analyzer was calibrated using medical grade NO at concentrations of 0 ppb (Nitrogen High Purity Gas; Air Liquide Australia, Melbourne, Australia) and 900 ppb (CIG Special Gases; Chatswood, Australia) diluted by a mass flow calibrator (Advanced Pollution Instrumentation, San Diego, CA). The measurement was made by an investigator who was blinded to the clinical context of each patient.

NO measurements were performed as previously described (23). Patients inhaled NO-free air (Medical Air; Air Liquide Australia). As NO concentrations are highly flow dependent (23), a constant flow rate of 5.95 L/min was maintained against a fixed resistance ensuring closure of the internal nasal route and thus eliminating nasal NO contamination of the exhaled gas (24).

The single breath test is characterized by a short NO peak followed by a longer NO plateau. The plateau concentration of NO when performed by the method described is reproducible and reflects lower airway NO release (25). The means of three technically acceptable measurements of plateau NO concentration were recorded. For comparison, exhaled NO levels performed in the same standardized method were also recorded on 50 healthy nonsmoking control subjects without a history of asthma or allergic rhinitis, as previously reported (18).

Lung Function

Spirometry was performed immediately prior to bronchoscopy. A computerized rolling-seal spirometer (Sensor Medics, Yorba Linda, CA) was used to measure flow volume loops.

Statistical Analysis

Data are expressed as means (± SEM) unless otherwise indicated. Comparisons between groups were made using the unpaired Student's t test for parametric data and the Mann-Whitney U test for nonparametric data. A correction for multiple testing (Tukey) was applied. To determine parameters that relate to the exhaled NO level (dependent variable) in the stable and BOS groups, univariate analysis was performed with the independent variables: iNOS epithelial score, iNOS and cNOS in the lamina propria, percentage neutrophils, and percentage lymphocytes in the BAL, age, days post-LT, cyclosporine level, and FEV₁ prebronchoscopy. Potentially significant parameters were then tested for possible interrelationship by multivariate analysis, which included the noncontinuous variables sex and type of transplant. All analyses were performed with SPSS software package.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The individual eNO levels in the four groups and 50 nonsmoking controls are shown in Figure 1. As discussed in a previous publication (18), mean ± SEM eNO levels for the stable LTR were not different from the controls (13.0 ± 0.7 ppb [range 6.4-18.3] compared with 14.2 ± 0.49 ppb [range 11.2-16.1]; p = 0.42). Mean ± SEM eNO levels for the BOS, BI, and AR groups are shown in Table 2. The results for the stable group, which have been previously reported, are represented to facilitate comparison. Exhaled NO levels were greater in both the BOS and BI groups when compared with stable LTR. There was no significant difference in eNO levels between the AR group and the stable LTR.

View larger version (7K): [in this window] [in a new window]   Figure 1.   Exhaled nitric oxide (NO) levels in 50 nonsmoking nontransplanted normal subjects (controls), stable lung transplant recipients (stable LTR), lung transplant recipients with acute rejection (AR), lung transplant recipients with bacterial infection (infection), and lung transplant recipients with bronchiolitis obliterans syndrome (BOS).

Within the BOS group, eNO levels tended to be higher in those (four of eight) with a recent clinical deterioration than those who were stable (mean ± SEM: 22.7 ± 0.7 versus 17.2 ± 0.9, p = 0.06).

The findings from the immunohistochemical staining of the EBB from the BOS, BI, and AR groups, as well as the previously reported stable group are shown in Table 2. Inducible NOS was present in both the bronchial epithelium and within macrophages and neutrophils within the lamina propria of all LTR. Constitutive NOS staining was present in both the lamina propria cellular population and endothelium but was absent from the bronchial epithelium.

Compared with stable LTR in the BOS and BI groups, iNOS protein expression was increased in the lamina propria and epithelium. In patients with AR, the bronchial epithelial iNOS score was not increased compared with stable LTR. In AR, however, iNOS staining was increased in the lamina propria, but this was noted to be predominantly in a perivascular distribution. In contrast, constitutive NOS staining was reduced in patients with BOS but not in the BI or AR groups.

The percentage of neutrophils in BAL aspirate was higher in the BOS group than stable LTR. Mean ± SEM percentage BAL neutrophilia was 26.7 ± 4.8 in the BOS group compared with 11.5 ± 3.2 in the stable LTR (p < 0.01).

To determine which factors related to eNO levels in the BOS group, regression analyses were performed with exhaled NO as the dependent variable. On univariate analysis, only the epithelial iNOS score and percentage BAL neutrophilia were associated with exhaled NO levels. The relationship between epithelial iNOS score and eNO is shown (Figure 2; r² = 0.67, p < 0.001). The relationship between percentage BAL neutrophilia and eNO appeared to be semilogarithmic (Figure 3; r² = 0.72, p < 0.0001). These relationships were very similar to those previously found in the stable cohort (black line in Figures 2 and 3).

View larger version (16K): [in this window] [in a new window]   Figure 2.   Relationship between exhaled nitric oxide (NO) level and bronchial epithelial inducible (iNOS) nitric oxide patients with obliterative bronchiolitis (OB) (open diamonds) including solid regression line and regression coefficient. The dotted line shows the relationship previously published (18) in stable lung transplant recipients.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Relationship between exhaled nitric oxide (NO) level and log percentage neutrophils in the bronchoalveolar lavage (BAL) in patients with OB (open diamonds) including solid regression line and regression coefficient. The dotted line shows the relationship previously published (18) in stable lung transplant recipients.

Exhaled NO levels were not significantly related to iNOS in the lamina propria nor cNOS, immunosuppressive medication, age, sex, FEV₁, time after transplantation, or transplant type.

All potentially significant variables on univariate analysis were examined for interrelationships by multivariate stepwise analysis. Both epithelial iNOS and percentage BAL neutrophilia remained strongly associated with eNO levels in both the BOS and stable groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although LT is an established and effective treatment for patients with severe lung and pulmonary vascular disease (12), its long-term benefit is limited by the development of BOS (21) and its pathological hallmark obliterative bronchiolitis. Progressive fibrosis and obliteration of airways with progressive airflow limitation characterize this (17), although these processes are likely to affect the whole of the airway tract in the allograft (15). BOS is often not diagnosed until well established by which time BOS is usually irreversible (21). However, early in its course, BOS appears to be characterized by increasing airway wall and BAL neutrophilia (15, 16) pathological findings present, albeit to a lesser degree, in stable LTR (15, 16, 18). We have previously shown that in stable LTR, eNO levels, although not elevated in this group as a whole, correlate with both the degree of BAL neutrophilia and with the intensity and extent of iNOS immunostaining in the bronchial epithelium adjacent to the airway lumen (18). We were interested in the current study to determine if this relationship held up in patients who have developed BOS. If so, eNO levels may be useful in predicting increasing airway inflammation that may characterize early, potentially reversible OB/BOS, although longitudinal studies will be needed to confirm that.

In this cross-sectional study, we have confirmed previous reports (19, 20) that eNO levels are increased in BOS patients when compared with stable LTR. Interestingly, we also found that while eNO levels were increased in the BOS group as a whole, there was a trend to suggest that the highest levels are found in those with recent clinical deterioration. This is consistent with the data of Fisher and colleagues (19), who found that eNO levels were higher in patients with early BOS (grade 1) compared with more established BOS (grades 2 and 3). This raises the possibility that in BOS, the highest eNO levels are found when active airways inflammation is occurring.

We have confirmed that in the LT population, raised eNO levels are not specific for BOS/OB and are increased also in the presence of respiratory infection. Although it is possible that bacterial or fungal airway infection contributed to the raised eNO levels found in the BOS group, this would appear to be unlikely in that no organisms were seen on gram stain, and organisms were cultured from only two patients who were, and remained clinically stable.

The immunohistochemical findings presented suggest that in both BOS and BI, there is up-regulation of iNOS in the bronchial epithelium and lamina propria. Other authors have found that iNOS is up-regulated in posttransplant BOS (26, 27) and in asthma (5), although in these studies eNO levels were not measured. Our findings support the view that it is up-regulated iNOS, rather than cNOS, which is responsible for the raised eNO levels seen in conditions characterized by airway inflammation, presumably in response to proinflammatory cytokines (5).

We have confirmed Fisher's findings that eNO levels are not increased in biopsy-proven AR. Silkoff and coworkers (28) found that eNO levels may be increased in patients with clinically suspected AR, but they did not confirm their finding with biopsy evidence of AR. Several authors have shown that in animal models, increased levels of circulating nitrates are seen in association with AR of organ allografts (29), which would seem paradoxical in view of our eNO data. However, our immunohistochemical findings may explain the discrepancy. In the group with AR, we found up-regulation of iNOS only in the lamina propria and only then in a perivascular distribution. Inducible NOS was not up-regulated in the bronchial epithelium. Due to its unpaired electron, NO is highly reactive (2). It seems probable that NO produced by up-regulated iNOS in the lamina propria is less able to diffuse into the airway lumen, thereby contributing far less to eNO levels, than NO produced in the airway epithelium. In contrast, up-regulated perivascular iNOS has ready access to the circulation and may be identified by increased levels of nitrate in the blood.

In this study, cNOS was down-regulated in the group with BOS but not in those with bacterial infection. Watkins and coworkers found that iNOS but not cNOS was expressed in association with inflammation in the airway epithelium of six patients undergoing surgical resection for lung cancer (32). In most studies of normal subjects, free of inflammation (5, 26, 27, 33), cNOS rather than iNOS was predominant in the airway wall. Guo and coworkers (34), among others however, found that iNOS is the predominant isoform in the airway epithelium. Our findings suggest that in chronic inflammation (but perhaps not acute inflammation such as seen with bacterial infection or AR), not only is iNOS up-regulated but, additionally, cNOS is down-regulated. Reduced perivascular total NOS (iNOS + cNOS) activity may be proinflammatory, favoring increased neutrophil adhesion and migration through the epithelium (35).

We are particularly interested in the factors that influence the transition from stable LTR to a development of BOS. We have confirmed previous reports from our group (15) and others (16) that when compared to stable LTR, posttransplant BOS is associated with increasing BAL neutrophilia. Further, on regression analysis, we found that the increased eNO levels found in BOS are significantly correlated with the degree of BAL neutrophilia. This suggests that eNO levels may represent a noninvasive measurement that can detect the development of airway inflammation, which characterizes early BOS and thus aid in its recognition, perhaps while it is still reversible.

We also found that in both stable LTR and transplant recipients with BOS, eNO levels are associated with the expression of iNOS in the bronchial epithelium. Saleh and colleagues found a similar relationship in people with asthma (36). Exhaled NO levels were not related to iNOS expression in the lamina propria, again presumably because the NO that is produced is less able to diffuse into the airway lumen or is produced by cNOS in much smaller amounts. This may explain why eNO levels are not raised in transplant recipients with biopsy proven AR in the absence of airway inflammation.

In summary, we have found that following lung transplantation, eNO levels are increased in bacterial infection and BOS. We have found a close relationship between eNO and BAL neutrophilia and bronchial epithelial iNOS in both stable LTR and transplant recipients with BOS. This suggests that serial eNO measurements may be able to predict the early development of OB, an hypothesis that urgently needs to be tested in a longitudinal cohort.