The Dynamics and Associations of Airway Neutrophilia Post Lung Transplantation

Authors


Abstract

Bronchoalveolar lavage (BAL) neutrophilia has been repeatedly observed in lung transplant recipients with established bronchiolitis obliterans syndrome (BOS). Little is known of the fluctuations in BAL and airway neutrophilic inflammation post-transplant. This prospective longitudinal study aimed to evaluate the dynamic changes of lung allograft neutrophils with time, immunosuppression, infection and BOS. A total of 28, initially healthy, BOS 0, lung transplant recipients underwent 134 bronchoscopic assessments, including BAL and endobronchial biopsies (EBB) (with immunohistochemistry) over 3-year follow up. Subsequently, 21 developed BOS 0p and 16 ultimately BOS. Compared to controls, there was early and persistent BAL neutrophilia (p < 0.05), contrasting with an initially normal EBB that shows a progressive increased airway wall neutrophil infiltrate. BAL neutrophilia (but not airway wall neutrophilia) was most striking when there was concomitant bronchopulmonary infection, particularly in the patients with BOS. Univariate and multivariate analyses suggested that BAL neutrophilia was linked to markers of infection while EBB neutrophilia was linked with coexistent inflammation with macrophages and lymphocytes. In conclusion: (i) BAL neutrophilia is predominantly associated with infection; (ii) Airway wall neutrophilia (as monitored by EBB) increases with time post-transplant and is not associated with infection; (iii) By itself, BOS is not the major contributor to BAL and EBB neutrophilia.

Introduction

Bronchiolitis obliterans syndrome (BOS), a form of chronic lung rejection, is the most important long-term complication of lung transplantation (LTx) and associated with a substantial mortality and morbidity. The mechanism by which chronic lung rejection leads to BOS remains poorly understood, but ongoing airway inflammation and inadequate tissue repair mechanisms are likely to play a pivotal role.

A number of human lung allograft studies have demonstrated increased neutrophil numbers with evidence of neutrophil activation in association with lung transplant BOS (1–6). Indeed, persistent bronchoalveolar lavage (BAL) neutrophilia can predict mortality after LTx (7). However, any temporal association between neutrophilia and BOS development must include consideration of the role of infection, which is known to contribute to both.

Most previous studies on airway inflammation in human lung allografts have focused on BAL (3–10) with minimal additional data available from airway wall biopsies (1,2,11). BAL findings are particularly useful in the diagnosis of post-LTx pulmonary infection (5), but give no direct insight into the nature of the cellular infiltrate within the airway walls. Endobronchial biopsies (EBB), by directly sampling the airway mucosa, have provided an enhanced understanding of the pathogenesis of airway diseases such as asthma and COPD (12–14).

Therefore, we conducted a prospective longitudinal study over several years, in initially clinically stable LTx recipients to evaluate the dynamic cellular profile of lung allograft airway neutrophils. The aim was to compare and contrast BAL and EBB results and associations, particularly attempting to define the potential roles of airway colonisation, infection and BOS on the development of airway neutrophilia.

Methods

Study populations

Stable LTx recipients, 3–9 months post-LTx, were recruited at The Alfred Hospital between January 1997 and December 1998. They were followed with surveillance bronchoscopic analyses that included BAL, EBB and transbronchial biopsy (TBB). For study inclusion, all patients needed to be clinically stable i.e. without evidence of acute rejection, clinical infection or BOS [i.e. BOS 0 (15)]. All enrolled patients received a standard immunosuppression regime, consisting of cyclosporine (CsA) [initially to achieve a trough blood level of 250–350 μg/L by EMIT assay (Syva, Los Angeles, CA) but reducing over time], azathioprine (1–2 mg/kg, maintaining white blood cells >5 × 109/L) and prednisolone reducing to a long-term maintenance dose of 7.5 mg/day. Routine clinical follow-up, including lung function testing and trough cyclosporine (CsA) measurements, was also performed at least every 2 months.

A total of 15 normal asymptomatic, nonsmoking volunteers with normal lung function (10 males, mean age of 35 years) were recruited as controls. They underwent fiberoptic bronchoscopy with BAL (n = 15) and EBB (n = 10) only on a single occasion. Patients and normal controls provided written informed consent. The study was approved by the institutional ethics committee.

BAL, EBB and TBB

The details of the bronchoscopy and sampling techniques have been previously described (1,2,11,16). Briefly, fibreoptic bronchoscopy was performed at 2, 4, 8, 12, 16, 39, 52, 78 and 104 weeks and thereafter, yearly post-transplant and as clinically indicated. Study bronchoscopies were ceased after the development of high grade BOS. After wedging the bronchoscope in the middle lobe, three 60 mL aliquots of phosphate buffered saline were introduced via syringe. The fluid was aspirated after each aliquot into a negative pressure (approximately −80 mmHg) vessel. Six EBB were taken from the origin of the lower lobe bronchi from patients and controls. Six TBB were taken from patients (but not controls) for the histological assessment of lung rejection (17). Fifteen milliliters of BAL was sent for bacteriological culture, fungal culture and polymerase chain reaction (PCR) for common respiratory viruses. BAL total cell and differential cell counts were also performed on cytospin preparations (1,2,11,16) and BAL supernatants were then stored at −80°C.

Immunohistochemistry

Immunohistological staining for 'panleukocytes', macrophages, lymphocytes and neutrophils was performed using monoclonal mouse anti-human antibodies against leukocyte common antigen (LCA), macrophage (CD68), lymphocytes (CD3, 4, 8) and neutrophil elastase (all from Dako, Denmark). Serial consecutive 3 m sections were cut from formalin fixed, paraffin embedded tissue (three sections for one cell marker stain) and were deparaffinised in xylene, rehydrated through graded ethanol solutions and washed with TBS. After pre-incubation with 20% normal horse serum for 20 min, the sections were stained with the primary antibodies to neutrophil elastase (diluted 1:600 in TBS), LCA (diluted 1:500 in TBS) or CD68 (diluted 1:100 in TBS) at 4°C in a moist chamber over night, and amplified by using the avidin-biotin-peroxidase complex (ABC). Isotype control immunoglobulins (IgG1 Dako, Denmark) were used as negative controls.

Stained sections were then quantified using a computer-assisted image analyzer system (Leica DMRB microscope, Germany, and Image Pro Plus, Media Cybernetics version 4.5, MD), at a final magnification of 400× and results were expressed as positive cells/mm2 of lamina propria (1,16).

BAL IL-8 assay

BAL IL-8 levels were measured in unconcentrated fluid by ELISA using a commercially available kit (Amersham, UK). The detection range was from 10 to 1000 pg/mL, with sensitivity of <2 pg/mL. The absorbency was measured with an ELISA reader (Model 450, Bio-Rad, Hercules, CA) at 450 nm. The amount of IL-8 present in the samples was calculated by reference to the wells containing dilutions of the standard for IL-8.

Lung function testing

Spirometry was performed according to American Thoracic Society Standards prior to each bronchoscopy. The determination of BOS status was according to the recent International Society of Heart and Lung Transplantation (ISHLT) guidelines (15). The treatment of BOS is described elsewhere (16).

Definitions of acute rejection and infection

Acute allograft rejection was defined as the presence of ISHLT grade A2 or above on TBB taken at the time of bronchoscopic assessment (18). The treatment of acute rejection is described elsewhere (19).

Microbiological results and clinical status were used to classify bronchoscopy results into three categories:

  • 1No infection i.e. no secretions on bronchoscopy, negative gram stain and culture and negative viral, fungal and mycobacterial culture or PCR;
  • 2Infection i.e. secretions on bronchoscopy and/or a symptomatic patient and/or positive bacterial, viral or fungal staining, culture or PCR with treatment initiated;
  • 3Indeterminate i.e. no secretions on bronchoscopy, asymptomatic patient but positive bacterial, viral or fungal staining, culture or PCR.

Statistical analysis

Statistical analysis was performed using SAS Institute Inc. software (Cary, NC). Data were initially assessed for normality and log-transformed where appropriate. Nonparametric data are presented as median with an interquartile range, while normally distributed data are presented as means ± SD. Lung transplant recipients were divided into two major groups:

  • 1Those who have not ever developed BOS (including a subsequent minimum of 2 years post-study follow-up without BOS) i.e. hereafter designated ‘Never BOS’. Group 1 had a subgroup of patients that moved from BOS 0 to BOS 0p (but by definition never had BOS).
  • 2Those who have ever had BOS diagnosed during the 3 years of the study i.e. hereafter designated ‘Ever BOS’ (15). Group 2 was further subdivided for analysis into pre-BOS (i.e. healthy at that time, but subsequently develop BOS) and post-BOS (i.e. BOS present at that time).

Differences between groups were assessed using Student's t-tests and validated using Wilcoxon rank-sum tests. Univariate analysis for continuous outcomes was performed using Pearson's correlation coefficients. This initial analysis included many clinical variables (age, gender, type of LTx, date of LTx, time from LTx, underlying diagnosis, cytomegalovirus serostatus, infection data, ISHLT grade A and B rejection status, immunosuppressive medication doses or appropriate level and BOS status). Those with a p-value less than 0.1 were included in the subsequent multivariate analysis. Multivariate models were developed using both stepwise selection and backward elimination procedures, with final models adjusted for repeat measures and further validated for clinical and biological plausibility. A two-sided p-value of 0.05 was considered to be statistically significant.

Results

A total of 28 stable lung transplant recipients were recruited to the study at a median of 14 weeks post-LTx. (range 11–51 weeks). Their median age was 50 years (range 19–61 years), 14 were female. Thirteen patients had undergone single LTx (SLTx), 14 bilateral LTx (BLTx) and 1 heart-LTx (HLTx). The underlying indication for transplant included obstructive lung disease (n = 11+), cystic fibrosis (n = 6) and bronchiectasis (n = 6), interstitial lung disease (n = 3) and pulmonary hypertension (n = 2). A total of 136 bronchoscopic procedures were performed, with each patient undergoing a median of 3 bronchoscopic investigations (range 2–9) between 11 and 232 weeks post-LTx.

Overall clinical status

During the subsequent 3 years of follow-up after enrolment 21 patients developed BOS 0p at a median of 130 weeks post-LTx (range 11–236 weeks). Sixteen patients went on to develop BOS 1 or greater at a median of 103 weeks (range 11–221 weeks). Thirteen LTx recipients have died, 11 with BOS as the primary cause, one with lung cancer (BOS 2) and one with Parkinson's disease (BOS 0). Of those alive, five are BOS 0, 4 are BOS 0p, two are BOS 1, one is BOS 2 and four are BOS 3.

TBB features of rejection

A total of 129 TBB were given an ‘A’ grading (for acute rejection), with 10 being Grade A2 or above (18). And 123 TBB were given a ‘B’ grading (for airway rejection) with 19 being Grade B1 or above.

Immunosuppressive medication

Table 1 shows immunosuppressant drug serum levels and doses. There were no statistical differences between the Ever BOS and Never BOS LTx groups, except for azathioprine doses between 53–104 weeks post-LTx. However, there was a trend to lower immunosuppression in the early post-transplant period for those patients who ultimately developed BOS. As might be expected, CsA levels and the doses of prednisolone and azathioprine were gradually reduced over the study period in both groups (p < 0.05 compared to early transplant).

Table 1. Trough serum CsA levels and doses of immunosuppressives
Weeks post-LTx≤2526–5253–104≥105
  1. Data are shown as medians and interquartile ranges.

  2. *p ≤ 0.05 compared with ≤25 weeks post-LTx.

  3. **p ≤ 0.001 compared with ≤25 weeks post-LTx.

  4. #p ≤ 0.05 comparison between Never- and Ever-BOS groups.

Never BOS:
 Serum levels CsA (μg/L)415326234**212*
(374–520)(228–477)(207–262)(142–246)
 Azathioprine (mg/kg)1.51.11.51.2
(0.3–1.8)(0.2–1.6)(0.8–1.7)(0.4–1.7)
 Prednisolone (mg/day)17.515.0**7.5**7.5**
(17.5–20)(11.9–15)(7.5–9.3)(7.5–7.5)
Ever BOS:
 Serum levels CsA, (μg/L)327#292281#250*
(269–367)(208–356)(223–294)(140–309)
 Azathioprine, (mg/kg)1.00.80.7#0.4*
(0.5–1.6)(0.5–1.5)(0.6–0.9)(0.3–0.8)
 Prednisolone, (mg/day)18.815.0**7.5**7.5**
(17.5–20)(10.0–17.5)(7.5–10.0)(7.5–10)

Study entry BAL and airway wall inflammatory cellular profiles

Table 2 summarizes the comparisons of baseline (study entry) BAL and airway wall cellular infiltrate between normal controls and the LTx patient groups. Compared with normal controls, the BAL total cell counts, neutrophil percentage and IL-8 levels were significantly elevated at study entry in both of the Never BOS and Ever BOS LTx patients (Table 2). In contrast, the airway total inflammatory cell counts (CD45-positive cells) and macrophage numbers (CD68-positive cells) were similar to normal controls in LTx patients early post-LTx. However, no significant differences were found for any study entry component between LTx patients who remained clinically stable in the following 3-year follow-up, and those who eventually developed BOS (Table 2).

Table 2. Study entry BAL and EBB cellular profiles
 Never BOSEver BOS Subsequently (n = 16)Controls (n = 15)
BOS 0 only (n = 7)BOS 0p subgroup (n = 5)
  1. Data are shown as medians and interquartile ranges.

  2. *p ≤ 0.05 versus controls.

  3. **p ≤ 0.01 versus controls.

BAL
 BAL return (mL)1139584**115
(66–126)(79–113)(57–105)(100–130)
 Total cells (×104/mL)27.5**48.1**27.9**12.5
(19.1–37.5)(16.2–51.0)(22.5–41.9)(11.3–15.0)
 Macrophages (%)76.975.672.781.1
(72.7–90.2)(63.5–92.3)(53.3–77.5)(70.4–91.4)
 Neutrophils (%)4.5**2.7*4.2**1.3
(1.9–5.2)(1.9–20)(2.2–13.9)(1.0–1.8)
 Lymphocytes (%)14.47.317.813.9
(3.6–23.7)(4.7–21.4)(9.1–21.8)(8.3–23.6)
 IL-8 (pg/mL)100**573**162**26
(37–735)(108–2349)(70–446)(19–34)
EBB
 CD45 (cells/mm2)709520848568
(475–1052)(394–979)(580–2300)(488–736)
 Neutrophils (cells/mm2)165295111124
(129–203)(130–551)(83–353)(94–194)
 CD68 (cells/mm2)154128262155
(101–300)(121–176)(153–399)(124–229)

Longitudinal changes in BAL and airway wall inflammatory cellular profiles

Longitudinally, the increases in BAL total cell count (cells/mL), neutrophils (%) and IL-8 levels (pg/ml) noted early post-LTx persisted unaltered with time (Figure 1A, Table 3). No significant longitudinal changes were found in either BAL macrophages (%) or BAL lymphocytes (%) (Table 3).

Figure 1.

Figure 1.

Changes of lung allograft neutrophilia with time post lung transplant in patients who remained clinically stable (Never BOS) versus those who develop BOS (Ever BOS). The subgroups of the bronchoscopic analyses of Never BOS patients who develop BOS 0p (but not BOS 1) are shown separately. The pre-BOS (open circle) and post-BOS (closed circle) subgroups of Ever BOS patients are also marked separately. (A) BAL neutrophilia; (B) Airway wall neutrophilia.

Figure 1.

Figure 1.

Changes of lung allograft neutrophilia with time post lung transplant in patients who remained clinically stable (Never BOS) versus those who develop BOS (Ever BOS). The subgroups of the bronchoscopic analyses of Never BOS patients who develop BOS 0p (but not BOS 1) are shown separately. The pre-BOS (open circle) and post-BOS (closed circle) subgroups of Ever BOS patients are also marked separately. (A) BAL neutrophilia; (B) Airway wall neutrophilia.

Table 3. Longitudinal changes in BAL cellular profiles
Weeks post-LTx≤2526–5253–104≥105
  1. Data are shown as medians and interquartile ranges.

  2. *p ≤ 0.05 versus ≤25 weeks post-transplant.

  3. #p ≤ 0.01 comparison between Never- and Ever-BOS groups.

  4. **Data used only from this group before BOS developed.

Never BOS:
 BAL return (mL)11310910775
(90–126)(79–121)(95–111)(68–110)
 Total cells (×104/mL)28.319.6*22.535.5
(25.4–37.5)(8.9–23.45)(16.9–32.35)(28.4–55.9)
 Macrophages (%)73.981.388.279.8
(72.7–88.0)68.7–91.2)(78.9–91.5)(40.5–89.1)
 Neutrophils (%)4.55.15.94.2
(1.9–9.9)(2.2–8.9)(3.1–9.5)(1.4–8.3)
 Lymphocytes (%)14.45.85.29.4
(6.7–23.7)(3.8–18.9)(3.4–11.5)(4.8–37.5)
 IL-8 (pg/mL)65121100366
(43–918)(30–255)(63–144)(153–886)
Ever BOS**:
 BAL return (mL)829195102
(52–112)(70–106)(50–109)(65–111)
 Total cells (×104/mL)29.627.4#29.028.1
(20.3–45.3)(21.4–42.5)(13.1–48.1)(19.0–43.5)
 Macrophages (%)75.874.865.2#72.4
(71.3–87.8)(55.5–81.3)(37.3–75.9)(48.4–83.8)
 Neutrophils (%)3.18.612.57.8
(1.7–14.7)(3.2–16.5)(4.3–45.8)(3.6–10.5)
 Lymphocytes (%)10.314.99.313.4
(5.9–16.9)(7.6–19.2)(6.1–18.9)(4.3–20.7)
 Il-8 (pg/mL)160260490*436
(90–431)(93–1030)(162–2070)(102–1176)

By contrast, the airway EBB total inflammatory cell counts (CD45-positive cells) and macrophage numbers (CD68-positive cells), although similar to normal controls in LTx patients early post-LTx, subsequently increased with time post-LTx (Figure 2). Airway wall neutrophil counts similarly started at normal levels and increased only in those with a diagnosis of BOS (Figure 1B).

Figure 2.

Changes of airway wall leucocytes (CD45+) with time post lung transplant in patients who remained clinically stable (Never BOS) versus those who develop BOS (Ever BOS). The subgroups of the bronchoscopic analyses of Never BOS patients who develop BOS 0p (but not BOS 1) are shown separately. The pre-BOS (open circle) and post-BOS (closed circle) subgroups of Ever BOS patients are also marked separately.

Subanalysis comparing pre-BOS with post-BOS groups revealed a prominent airway wall and BAL neutrophilia and higher levels of BAL IL-8 after the establishment of BOS (Table 4). In addition, BAL return decreased significantly after BOS (Table 4). BAL IL-8 levels were associated with BAL neutrophil percentage (r= 0.68, p < 0.001) and EBB neutrophil numbers (r= 0.26, p = 0.01) in regression analysis.

Table 4. BAL and EBB inflammatory cell profiles pre- and post-BOS development
 Pre-BOSPost-BOSp-value
  1. Data are shown as medians and interquartile ranges.

BAL
 BAL return (mL)101 (80.0–114)61.5 (40–79.6)<0.001
 Total cells (×104/mL)27.8 (20.5–40.5)31.3 (13.3–51.1)0.82
 Macrophages (%)74.4 (55.4–82.6)64.6 (20.1–72.3)0.025
 Neutrophils (%)5.5 (2.9–12.9)12.5 (6.2–60.0)0.007
 Lymphocytes (%)13.6 (6.9–21.8)6.9 (3.4–16.5)0.054
 BAL IL-8 (pg/mL)212 (94–480)1099 (597–2355)<0.001
EBB
 CD45 (cells/mm2)896 (447–2051)788 (483–2405)0.51
 CD68 (cells/mm2)295 (178–481)512 (254–1107)0.007
 Neutrophils (cells/mm2)240 (114–404)423 (269–585)<0.001

Effect of infections on BAL and airway wall neutrophilia

In combining clinical and bronchoscopic results from 134 procedures, there were 21 (15.6%) diagnoses of ‘infection’ (as defined above) in 16 patients. Forty-three (32.0%) of the bronchoscopies were assessed as ‘no infection’. The remainder were of ‘intermediate infection’ status. The predominant bacteria identified were Ps. aeruginosa and S. aureus. Viral culture noted CMV on 24 occasions and respiratory RNA viral PCR noted other viruses in two specimens. Aspergillus was the only fungus, cultured on four occasions.

Subanalysis for BAL and airway wall neutrophilia was performed, in the absence and presence of infections and with or without simultaneous BOS (Figure 3A,B). Importantly, BAL neutrophil percentage (Figure 3A), but not EBB neutrophil numbers (Figure 3B), increased significantly with the presence of bronchopulmonary infection in LTx patients who developed BOS. Additionally, compared with pre-BOS, in the presence of concomitant infections, BAL neutrophilia was more marked after BOS developed (Figure 3A, p = 0.01).

Figure 3.

Figure 3.

Effect of bronchopulmonary infections on lung allograft neutrophilia in patients who remained clinically stable (Never BOS) versus those who developed BOS (Ever BOS, pre-BOS: open circle; post-BOS: closed circle): (A) BAL neutrophilia; (B) Airway wall neutrophilia.

Figure 3.

Figure 3.

Effect of bronchopulmonary infections on lung allograft neutrophilia in patients who remained clinically stable (Never BOS) versus those who developed BOS (Ever BOS, pre-BOS: open circle; post-BOS: closed circle): (A) BAL neutrophilia; (B) Airway wall neutrophilia.

Multivariate analysis

Univariate and subsequent stepwise multivariate analyses, developed models for the variables that influenced the BAL neutrophil percentage (Table 5) and EBB neutrophil numbers (Table 6). Bacterial gram stain (as a marker of bacterial load) accounted for 18% of the total 38% of explainable variability of BAL neutrophil percentage, while EBB macrophage counts accounted for 26% of the total 32.9% of explainable variability of EBB neutrophil numbers. Importantly, the presence of BOS per se had relatively little effect on the BAL neutrophil percentage and EBB neutrophil numbers, and infection markers did not affect EBB neutrophil numbers.

Table 5. Significant associations with BAL neutrophil percentage (a) univariate and (b) multivariate analyses
 Pearson correlation coefficientp-value
(a) Univariate variable
 BAL bacterial gram stain0.430.0001
 Presence of clinical infection0.380.0001
 BOS development within 6 months0.330.0001
 ‘B’ ISHLT TBB rejection grade0.300.001
(b) Multivariate variableProportion of variability explainedp-value
 BAL bacterial gram stain18%0.0001
 Any recipient diagnosis except
 Pulmonary hypertension8%0.0007
 ‘B’ ISHLT TBB rejection grade5%0.006
 CMV positive serology5%0.005
 BOS development within 6 months2%0.04
 Total variability explained38% 
Table 6. Significant associations with EBB neutrophil number (a) univariate and (b) multivariate analyses
(a) Univariate variablePearson correlation coefficientp-value
 EBB CD68-positive cell number0.57<0.0001
 Time post-transplant0.45<0.0001
 EBB CD3-positive cell number0.43<0.0001
 EBB CD8-positive cell number0.40<0.0001
 BAL neutrophil percentage0.35<0.0001
 Prednisolone daily dose−0.34<0.0001
 Azathioprine daily dose−0.310.0007
 BOS development within 6 months0.260.004
(b) Multivariate variableProportion of variability explainedp-value
 EBB CD68-positive cell number26%<0.0001
 Azathioprine daily dose3.2%0.002
 BAL neutrophil percentage3.7%0.04
 Total variability explained32.9% 

Discussion

This study has investigated lung allograft airway neutrophil profiles prospectively in 134 sequential EBB and BAL samples from 28 LTx recipients over a 3-year follow-up. The results reveal the novel observation that there was early (≤25 weeks post-LTx) and persistent BAL neutrophilia. In contrast, airway wall neutrophils started off normal but progressively increased. Samples obtained after the diagnosis of BOS showed a further significant elevation of BAL/airway neutrophils and BAL IL-8 levels, compared to pre-BOS, although both the longitudinal BAL neutrophil percentage and airway wall neutrophil density showed no significant differences between the Never and Ever BOS groups at baseline. Subanalysis suggests that BAL neutrophilia (but not airway wall neutrophilia) was most striking when there was concomitant bronchopulmonary infection, particularly in the patients with BOS. Univariate and multivariate analyses reveal that BAL neutrophilia was linked primarily to markers of infection while EBB neutrophilia seems to be most linked to coexistent infiltration with macrophages and lymphocytes. Surprisingly, the multivariate analyses appear to down play a strong primary association of LTx airway neutrophilia with BOS.

BAL neutrophilia has been repeatedly observed in lung transplant recipients, a situation most pronounced in patients with established BOS (3–6,8–10). In comparison, little is known of airway wall neutrophilia (1,2,11). The present longitudinal study confirms our previous cross-sectional study findings of increased airway wall neutrophil density in LTx patients (1) and also provides new evidence of kinetic changes in both BAL and airway wall neutrophil infiltrates with time, infection, immunosuppressant medication (i.e. azathioprine) and the development of BOS.

The distinct and discordant longitudinal profiles of BAL and airway wall neutrophilia revealed by this study provide better understanding of the diagnostic value and utility of BAL and EBB in monitoring LTx recipients. Both techniques are useful and complementary. BAL neutrophil percentages have been suggested as an early, easy to measure, predictor of BOS (3–6,9,10). However, it can be seen that BAL neutrophils are elevated well before the development of BOS (indeed, even elevated in those transplant recipients who never develop BOS), remain constant at high levels over a long-term, and are increased significantly further with bronchopulmonary colonisation and infection. This raises the interesting question as to whether all LTxs always have subtle infection/colonisation or inflammation (allo or ischemic or some other factor related) all along. Airway wall neutrophil infiltrates occur progressively and relates to immunosuppression doses and coexistent cellular infiltration. It is likely this airway inflammation represents uncontrolled alloimmune phenomena driven by cytokines and is seemingly less responsive to allograft infections (3,20). Whatever its genesis, this additional airway inflammation will add further to the high baseline BAL neutrophil load. An inflammed functionally abnormal large airway may well then increase the prospects of infection which will increase the BAL neutrophil load further and result in further large and small airway damage and dysfunction.

While interesting, the association of BAL neutrophilia and airway infection does not allow a conclusion as to whether they both represents a cause or a consequence of BOS. Conflicting evidence is presented in the lung transplant literature (1–8). These variable results represent the considerable difficulties in defining infection versus colonization and the limitations of cross-sectional versus longitudinal studies. Notwithstanding, there is a solid body of evidence linking the onset of an obliterative bronchiolitis/BOS picture with infection. Community acquired viral infections have been best shown to increase the incidence of BOS (21) and cytomegaloviral reactivation is also linked (22). There is similar emerging evidence of the role of intracellular bacteria such as Chlamydia pneumoniae (23). In our sequential 134 bronchoscopy assessments for 28 subjects over 3 years, there were 21 (15.6%) diagnosed as having infection, of which 16 were in the BOS group, and only 2 in the non-BOS group. There is a real question as to whether airway colonization with bacteria is actually a benign condition or a marker and predictor of problems to come. Indeed, it may be pathological as there is evidence to support the adverse effects of airway ‘colonization’ in the etiology of airflow obstruction in the COPD (24) and pediatric cystic fibrosis (25) and the pathogenesis of BOS is likely to have a similar association.

Having been attracted to the airway, whether as a result of initial infection and an innate immune response, or as part of the adaptive immune reaction to alloantigens, the neutrophil is likely to be an effector cell for the development of increasing airway damage and the progression of clinical BOS. Neutrophil products, such as proteases, matrix metalloproteinases, neutrophil myeloperoxidase, and reactive oxygen metabolites, can potentially induce tissue injury at the site of neutrophil-dominated inflammation. High neutrophil numbers in the airways correlate with the degree of airflow limitation in patients with COPD (26,27). Neutrophilic inflammation and neutrophil activation observed in lung allografts, therefore, may provide a possible novel target for new therapeutic strategies in the management of BOS. In this context, the observations that azathioprine doses appeared initially lower in those who developed BOS, and that there was an inverse relation between prednisolone and azathioprine doses and EBB neutrophilic infiltration, may be important. Interestingly, the importance of azathioprine therapy as an agent to prevent BOS progression was first suggested over a decade ago (28).

In the recruitment of neutrophils to the lung, chemoattractants serve as a homing mechanism to precisely target neutrophils to the sites of inflammation. High levels of BAL IL-8, one of the most potent neutrophil chemokines produced by monocytes/macrophages, and other cells, including bronchial epithelium and pulmonary endothelial cells (18,29) has been also repeatedly reported in BOS patients (1,3,4). The longitudinal changes of BAL IL-8 levels post-transplant appeared to mirror the pattern of BAL neutrophilia. We similarly found a significant association of BAL IL8 and EBB neutrophilia. Our study also reveals a significant association between EBB neutrophilia and macrophage and lymphocytic infiltration. Laan et al. showed that lymphocyte derived IL-17 is a potent inducer of IL-8 release from lung endothelial and epithelial cells in order to create a chemotactic gradient toward the airway lumen (20). We have previously reported an early and persistent increase in BAL CD8+ T-cells, and a progressively developing CD8+ cell infiltrate in the airway walls with the development of BOS in the same group of lung transplant patients (17). Potentially, IL-17 may be the link between alloantigen-dependent innate immune responses (such as T-cell-mediated insults) and alloantigen-independent (such as infection and neutrophil induced tissue injury) in the pathogenesis of BOS. Although these processes have defied traditional immunosuppression to date, the neutrophil potentially represents a novel therapeutic common pathway target. Further studies are needed.

In conclusion, this study has characterized the features of longitudinal BAL and airway wall neutrophilic inflammation occurring in lung allografts. The presence of luminal infection and the nature of the immunosuppression account for the major attributable variability in airway neutrophilia. Utilizing a longitudinal, multivariate approach, BOS appears to have less of a role than has been recently proposed. With mounting circumstantial evidence that infective agents may have more of a role than previously suspected, future lung allograft neutrophilia studies must include even more details of infection status. Monitoring airway wall neutrophilia by EBB provides complementary information over simple BAL alone and may also prove useful in monitoring the development of BOS post-transplant since it appears less affected by airway sepsis. Novel therapeutic strategies to prevent neutrophil sequestration and activation must now be considered in the management or prevention of BOS in lung transplant recipients.

Ancillary