We report here findings on the infiltration of inflammatory cells and expression of markers of humoral immune response in bronchial biopsies of patients with stable COPD with different degrees of lung functional decline. We evaluated three subgroups of COPD patients with rapid, slow or no lung functional decline monitored for an average of 5.8 years (all patients). The expression of secretory IgA was significantly reduced in bronchial epithelium and plasma cell numbers were significantly lower in the bronchial lamina propria of rapid decliners with stable COPD compared to non-decliners. No difference was found in the bronchial inflammatory cell infiltration due to CD4, CD8, CD68, CD20, NK, neutrophils, eosinophils, mast cells, polymeric immunoglobulin receptor (pIgR) in epithelium and lamina propria of rapid decliners compared to the other groups. Plasma cells in the lamina propria correlated positively with total IgA score in lamina propria of all patients. In vitro stimulation of 16HBE cells with LPS (10 μg/ml) and IL-8 (10 ng/ml) significantly increased pIgR expression in human bronchial epithelial cells while H2O2 (100 μM) significantly decreased it.

In the three subgroups (phenotypes) of COPD patients, follow-up duration for functional decline was similar (5.8 years) and ranged for all patients between 3 and 15 years. Similar FEV1% predicted values, RV%, DLCO/VA% and CT scored emphysema% were found, though the group with rapid functional decline showed a trend to lower DLCO/VA% and higher CT-scored emphysema%. DLCO% was slightly but significantly lower in rapid decliners compared to non-decliners (Table 1). It is likely that in a larger sample these lung functional parameters and CT-scored emphysema could would show statistically significant differences. However, this study was designed for different purposes, namely to analyze inflammatory cell infiltration and markers of humoral immune response, for which the sample size was sufficient, going by the literature [21, 22]. The subgroup of non-decliners improved their lung function over the 6.1 years (mean ± SD) study period. In an earlier study, 15% of patients assessed showed an improved lung function over the 3-year study period used for monitoring those patients [10], and the possibility that some patients might have improvements over time was also noted by Fletcher and Peto [9, 10] in historical studies on the natural history of COPD [9, 10]. Whether this is related to genetic differences or to response to treatment, numbers of exacerbations per year or smoking habit is not clear [10]. Interestingly, our rapid decliners showed a significantly higher pack/year consumption compared to slow and non-decliners (Table 1) while the percentage of frequent exacerbators in our rapid decliners (20%) was similar to the percentage amongst non-decliners (28%). These findings are in line with a previous study showing that current smokers had 21 ± 4 ml/year more decline compared to former smokers [10] while exacerbations during follow-up correlated to greater decline in FEV1 with a mean loss of 2 ± 0.5 ml/year per exacerbation [10], suggesting that smoking habit plays a major role in lung function decline.

In our study, the number of patients on the different therapies (mainly LAMA, LABA and ICS) before bronchoscopy and at discharge after bronchoscopy was similar in the three subgroups of COPD patients (Table 3), and after discharge therapy remained unvaried in the follow-up period in 80.9%, 78.6% and 100% of non-, slow, and rapid decliners, respectively (Table 3). Furthermore, ICS consumption at discharge tended to be higher in rapid decliners compared to the other groups (Kruskal–Wallis, p = 0.060). This seems in contrast with data from the literature, e.g. the TORCH study which indicated that decline in FEV1 may be reduced with regular treatment [7, 10]. Unfortunately our study cannot provide an answer to this point since it was designed with a limited number of patients to answer different questions. Our observation of a somewhat higher ICS consumption in rapid decliners at discharge was, in our opinion, mainly due to their severer outcome observed by clinicians who tended to increase the pharmacologic treatment for this subgroup of patients. We do not know whether a lower ICS treatment level than that observed in our rapid decliners would have even further aggravated the outcome. Based on the TORCH study [7, 10], we can surmise that it could be the case. Another aspect to consider is that the adherence to therapy of our patients was not precisely monitored prospectively, since our study was conducted retrospectively.

In our immunohistochemical analysis of bronchial biopsies from rapid, slow and non-decliners with COPD, we found no significant differences among groups concerning the bronchial inflammatory cell infiltration due to lymphocytes, NK cells, B cells, macrophages, mast cells, neutrophils and eosinophils, indicating that, at bronchial mucosal level, the increased systemic inflammation previously reported [11,12,13] is not evident. Interestingly, we observed decreased levels of secretory IgA in bronchial epithelium of rapid decliners compared to non-decliners (Fig. 1) and a decreased plasma cell count in the lamina propria of rapid decliners compared to non-decliners (Fig. 3). These data suggest that an impairment of the humoral immune response, inhibiting adherence of microorganisms to epithelium and their clearance, may be involved in the rapid lung function decline aggravating the disease state of these patients. In previous studies, a decreased level of secretory IgA or pIgR has been reported in association with severity of COPD. In severe COPD, a decreased pIgR bronchial expression correlated with airflow limitation and neutrophils [16, 17], while a reduced secretory component in both large and small airways correlated with the number of neutrophils in the glands in large airways, and with bronchial obstruction in small airways [16]. Areas of altered bronchial epithelium in COPD showed decreased secretory IgA and pIgR associated with increased inflammation [17]. Reduction of secretory IgA in small airways was associated with invasion of bacteria, NFkB activation, increased presence of macrophages and neutrophils and fibrotic remodeling of the small airways [23]. Specific IgA levels against P. aeruginosa were lower in stable non-colonized COPD compared to healthy subjects [24]. Another study found decreased pIgR in bronchial epithelium of severe COPD and decreased pIgR and IgA transcytosis in bronchial epithelium cell cultures of severe COPD vs. control subjects [25]. In our study, we did not observe a significant correlation between lung function decline level (ml/years) and post-β2 FEV1% predicted values (Spearman rank correlation, p = 0.249). In fact, the three groups of COPD patients had similar (mean ± SD) FEV1% predicted values notwithstanding their significantly different lung function decline (according to pre-defined cut-offs, see methods and Table 1). This suggests that, independently of the degree of severity of bronchial obstruction, the lower levels of secretory IgA in epithelium and plasma cells in lamina propria that we found in rapid decliners vs. non-decliners are a specific feature of functional decline and may be considered as markers of lung function decline in COPD patients.

In distal airways, IgA+ B cells numbers were increased in lymphoid follicles (LF) from severe COPD compared to control non-smokers [26] and the intra-LF IgA+ cells (%) were further increased in severe COPD compared to mild disease [26], presumably representing an adaptive immune response to microbial and/or self-antigens, particularly in severe disease [26]. However, in a different compartment of the lung, the IgA+ plasma cell numbers observed in bronchial mucosa of patients who died of COPD were lower compared to COPD patients who died from other causes [18]. More recently, plasma cells have been shown to be more numerous in the mucosal glands of patients with chronic bronchitis compared to asymptomatic smokers, but similar in number in the subepithelium [27]. In the bronchial mucosa (lamina propria) of our rapid decliners with COPD we observed a significant reduction of plasma cell numbers (Fig. 3) compared to non-decliners, which was also significantly and positively associated with the total IgA score in lamina propria and secretory IgA score in bronchial epithelium, confirming a relationship between the plasma cell numbers populating the bronchial mucosa and secreted IgA (Fig. 4). These findings coming from the large airways of COPD patients suggest an impairment of the humoral immune response developing not only in the more severe forms of COPD [9,10,11,12] or in patients who died from COPD [18] but also in the presence of a rapid lung function decline and deterioration of the disease state. Mechanisms of plasma cell maturation and differentiation need to be studied to further investigate the underlying molecular mechanisms [28,29,30] potentially involved in the reduction of plasma cells and secretory IgA in rapid decliners with COPD.

Our in vitro data, showing a decrease of polymeric Ig receptor both in supernatants and cell lysate samples after H2O2 treatment of 16HBE cells, a normal bronchial epithelial cell line, suggest speculatively, that oxidative stress caused by smoking habit, may play a role in reducing the IgA transcytosis from the normal bronchial mucosa to the airway lumen, starting in an early phase of bronchial obstruction.

In conclusion, we found an impairment of humoral immune response in COPD patients with rapid functional decline. This finding contributes to the phenotyping of COPD patients.