In the present study it was shown that CD11b expression on neutrophils is enhanced in association with their transition from the blood to the airways in smokers. This upregulation was observed irrespective of the presence of chronic airflow limitation, but was not shown in non-smokers. Furthermore, CD11b expression on circulating neutrophils was enhanced in patients with COPD compared to non-smokers. This finding demonstrates that even circulating neutrophils are activated in COPD and that smoking is associated with a further activation of airway neutrophils that is not observed in non-smokers. In addition it was shown that, while CD11b expression is enhanced on airway neutrophils from smokers, the expression of the adhesion proteins CD62L and CD162 on neutrophils almost vanishes when entering the airways of both smokers and non-smokers. We also showed that the levels of soluble ICAM-1 and ICAM-3 in serum are increased in COPD, confirming systemic inflammatory activity in this disease .
Although an increased CD11b expression on sputum neutrophils has previously been shown in COPD  this is, to our knowledge, the first time that CD11b expression in neutrophils has been studied in different compartments within the same COPD patient. The two groups of smokers are fairly well matched with regard to cumulative tobacco exposure and there is a substantial difference in lung function between the smokers with and without COPD, indicating that comparisons between the smoker groups may be related to the presence of airway obstruction. It is, however obvious that even smokers who did not display chronic airflow limitation and thus had not developed COPD according to current diagnosis criteria, were not healthy as diffusion capacity was clearly impaired in this group compared with healthy non-smokers. This finding indicates that pathological processes in the lungs, as a consequence of smoking, may develop along different lines in different individuals.
Our data showed an increased CD11b expression on sputum neutrophils from smokers without COPD and that CD11b expression on sputum neutrophils exceeded that of blood neutrophils in smokers, but not in non-smokers. This likely indicates that neutrophils from smokers, irrespective of airflow obstruction, become activated in connection with the transition from the circulation to the airways, a phenomenon that does not occur in non-smokers. In addition, sputum neutrophils from smokers without COPD are more activated than neutrophils from smokers with COPD. Smoke exposure of isolated neutrophils in vitro activates neutrophils as it causes increased expression of β2-integrin (CD18) and decreased CD62L expression . One may speculate that the activation of sputum neutrophils caused by smoking may be attenuated due to exhaustion of the immune system in subjects with COPD, a speculation that is supported by the finding of a correlation between CD11b expression on sputum neutrophils and disease severity, as assessed by FEV1 (% of predicted). In line with this, expression of CD11b is down-regulated during apoptosis and while the apoptosis rate of circulating neutrophils has been shown to be unaltered in COPD, an increased apoptosis rate has indeed been reported in sputum neutrophils from subjects with COPD [19, 20]. Further studies on neutrophil functions in this context have to be performed. Interestingly, FEV1/VC-ratio has previously been shown to be inversely correlated with CD11b expression on sputum neutrophils in subjects who were current or ex-smokers, irrespective of significant airflow obstruction . Our findings are in agreement with these data in smokers without COPD, but are not confirmed in the smokers with COPD. This may lend further support to the notion that sputum neutrophils from subjects with COPD might have an impaired function compared to neutrophils from smokers without COPD.
Neutrophil expression of CD11b does not only attract attention as an adhesion molecule but it is also considered a marker of neutrophil activation. In the current study we showed that CD11b expression is increased on circulating neutrophils from subjects with COPD as compared to healthy controls. This is in agreement with earlier studies [12, 21] and emphasises the systemic component of the inflammation in COPD as we demonstrated that even circulating neutrophils from patients with COPD displayed signs of activation. Whereas expression of CD11b increase on activated neutrophils, CD62L is shed upon activation . Although there was no difference in CD62L surface expression between the groups in the current study, CD62L was significantly lower on neutrophils from sputum as compared to neutrophils from blood. Thus the CD62L data support the CD11b data, and suggest that airway neutrophils are further activated in relation to circulating neutrophils.
While neutrophil numbers in BAL fluid and sputum are often increased in COPD, bronchial biopsies from subjects with COPD show varying results with regard to neutrophil numbers [1–3, 23, 24]. It has been hypothesised that this discrepancy between different compartments is due to the rapid migration of neutrophils through the tissue in order to reach their primary destination in the airway lumen. In addition, increased neutrophil numbers have been suggested to be more characteristic of the inflammation present in the small airways [25, 26]. In the current study neutrophil numbers in BAL fluid were increased and there was a similar trend in sputum.
Neutrophils expressing CD11b have previously been shown to be increased in the submucosa of subjects with COPD compared to control smokers . In this study, the presence of CD11b staining was confirmed and immunohistochemical staining of bronchial biopsies showed neutrophils expressing also CD62L and CD162.
In the current study, neutrophil expression of CD162 did not differ between the groups, but there was a lower expression on sputum neutrophils compared to blood neutrophils. Contradictory to this finding, a small study of subjects with varying stages of COPD (GOLD I-IV) found increased CD162 expression on circulating neutrophils in the COPD group . The ligand of CD162, P-selectin, has received attention as a marker of platelet activation and systemic inflammation. In one study it was found that platelets were activated in COPD, but this was not reflected by differences in soluble P-selectin . This is in line with the current results where no differences in serum P-selectin were detected between the groups. However, other studies support a role for P-selectin as a marker of systemic inflammation in COPD and P-selectin has been associated with impaired lung function in a large cross-sectional study . Also, Ferroni et al. found increased plasma levels of P-selectin in subjects with COPD as well as an inverse relation between P-selectin and PaO2. It is possible that a larger sample size is needed to demonstrate a difference between groups although this would also limit its usefulness as a marker of systemic inflammation.
E-selectin is only expressed on activated endothelium and together with soluble ICAM-1 soluble E-selectin is sometimes considered a sign of endothelial activation. The percentage of E-selectin positive vessels in bronchial biopsies as well as serum E-selectin has been shown to be increased in COPD [32, 33]. However, the current study found no significant difference in serum levels and it was below the detection limit in both sputum and BAL fluid samples.
Soluble ICAM-1 and sICAM-3 in serum were increased in smokers with COPD as compared to healthy non-smoking subjects. This finding is in agreement with those of Riise et al.  who found increased levels of circulating ICAM-1 in COPD patients with FEV1 between 60 and 70% of predicted value, i.e. similar to the patients in this study. Walter et al. also found increased levels of soluble ICAM-1 in serum from patients with moderate COPD, FEV1 60 – 70% of predicted value . Walter et al. also found a relationship between circulating ICAM-1 levels and disease severity, as assessed by FEV1, but this association was not demonstrated by Riise et al. . In contrast to those findings Noguera et al. showed lower serum levels of ICAM-1 in COPD than in healthy non-smokers . In that study patients with severe COPD, with FEV1 on average 33% of predicted value, were studied. The main sources of ICAM-1 are bronchial epithelial cells and endothelial cells [33, 34]. Endothelial cells respond to pro-inflammatory stimuli by an increased production of sICAM-1 [34, 35]. Elevated levels of TNF and other pro-inflammatory cytokines is a common feature of COPD which likely is of importance for the increased sICAM-1 levels observed in COPD. It could thus be hypothesised that there is an association between disease severity and the levels of ICAM-1 in mild and moderate COPD and that severe COPD, characterised by loss of functional pulmonary capillaries and airway mucosal cells, is associated with decreased production and release of adhesion molecules into the circulation.
Previously, Riise et al. showed an increase of sICAM-1 in bronchial lavage fluid . In agreement with this observation, a trend towards higher sICAM-1 levels was observed in the sputum supernatants from the COPD group. Considering the 50-fold difference in volume between our BAL technique and their small volume bronchial lavage it is reasonable to relate the small volume bronchial lavage to sputum supernatants as they both represent the central airways. The role of sICAM-1 in the lung is not fully understood but both human studies and animal models indicate that alveolar epithelial cells are an important source of the sICAM-1 found in BAL fluid . Moreover, it has been shown that sICAM-1 can induce neutrophil-mediated cytotoxicity  and activate lung macrophages . It is therefore possible that the decreased sICAM-1 levels in BAL fluid, observed in the COPD group, could be related to an increased binding of sICAM-1 to different effector cells. In addition, it appears that ICAM-1 is differentially regulated in the bronchial epithelial cells and alveolar epithelial cells, with a more marked increase of ICAM-1 in response to stimuli observed in bronchial epithelial cells [36, 39].
Expression of PECAM-1 on the endothelium has been shown to be of importance for transendothelial migration of neutrophils as the blocking of PECAM-1 abolishes transendothelial migration . Soluble PECAM-1 can bind endothelial PECAM-1 and thus prevent neutrophils from transmigrating . However, other adhesion molecules have been shown to be involved in transmigration and the adhesion molecules involved may change depending on stimulus and disease . In the current study a tendency towards lower serum sPECAM-1 in smokers was found irrespective of airflow obstruction, with a significantly lower level in smokers without COPD as compared to the COPD group. It could be speculated that the lower levels in the smoker group are caused by sPECAM-1 binding to endothelial PECAM-1 as part of a protective mechanism, a mechanism that has failed in the COPD group. In BAL fluid sPECAM-1 was higher in smokers without COPD as compared to both other groups.
The available literature on the role of soluble adhesion molecules in COPD is contradictory. Most studies have investigated serum levels; although some authors have studied BAL fluid. There is large variation in severity of disease between the studies and there is also a difference in the techniques used to analyse the samples which makes direct comparisons difficult.
In conclusion, our data show that neutrophils, from COPD patients, retain, and further enhance, the state of activation that is observed in circulation, even after migration into the lungs. While the general role of soluble adhesion molecules in COPD still requires further investigations, some molecules, such as serum ICAM-1, appear to be reliable markers of the systemic inflammation in COPD.