Lung function decline in asthma patients with elevated bronchial CD8, CD4 and CD3 cells

Discussion

The present study shows that the number of bronchial CD8⁺- and CD4⁺-cells at baseline predicts the subsequent decline in lung function in patients with mild-to-moderate atopic asthma during 14 years of follow-up. At the end of follow-up, the numbers of CD8⁺-, CD3⁺- and granzyme B⁺-cells were correlated with the preceding lung function decline. Our data indicate that, despite on average having normal decline, the long-term course of lung function is very heterogeneous among patients with mild-to-moderate asthma, and that subgroups of patients with high CD8 or high CD4 cells in bronchial biopsies might be at risk of more pronounced loss of lung function over many years.

To the best of our knowledge, the present study is the first to examine the relationship between inflammatory markers in bronchial biopsies and the long-term course of lung function in asthma. Previously our group has shown that CD8⁺ cells in bronchial biopsies are of prognostic value for subsequent lung function decline in patients with mild-to-moderate asthma during short-term follow-up (7.5 years) [11]. This has now been extended by doubling the follow-up period, and by a follow-up bronchoscopy at t=14 years. Both at baseline and t=14 elevated CD8 counts were associated with a more rapid decline of FEV₁, which demonstrates internal validity. Besides CD8 cells, we found that CD4 cells were also predictive for subsequent pre- and post-bronchodilator lung function decline as measured by FEV₁ % predicted. These differences between associations at t=7.5 and t=14 years might be caused by the fact that we followed these patients over a longer period of time. Our previous study at 7.5 years of follow-up might have missed the association with CD4 cells because the observation period was too short.

What could be the role of CD8 cells in asthma? In patients dying from acute asthma a higher number of activated CD8⁺ T-cells was observed in post-mortem lung tissue samples as compared with tissue from asthmatic and control subjects [12]. However, CD8 cells may also be involved outside acute episodes of asthma, particularly in patients with relatively severe disease. First, activation of circulating CD8⁺ T-cells was recently reported in patients with severe asthma using transcriptome analysis [13]. Secondly, an association between CD8 cells in sputum and asthma severity, as measured by bronchial hyperresponsiveness to histamine, was previously shown by Cho et al. [14]. In addition to production of T-helper (Th)2 cytokines by CD4⁺ T-cells in the asthmatic airways, CD8⁺ T-cells were also found to produce interleukin (IL)-4 and IL-5 in the airways in asthma [14, 15]. A recent study suggests an important role for a subset of peripheral blood effector memory IL-6Rα^high CD8 cells in the production of Th2 cytokines in asthma [16]. CD8 cells may be triggered by allergens as a result of cross-presentation of exogenous allergen by major histocompatibility complex class I on airway dendritic cells to CD8⁺ cells [17]. In mice the response of CD8 cells to aerosolised allergen was found to contribute to pulmonary eosinophilia and bronchial hyperresponsiveness [18]. This notion is supported by studies from Seneviratne et al. [19] who defined three CD8 T-cell epitopes for house dust mite allergen.

A link between CD8 cells and asthma might also be explained by respiratory viral infections and other intracellular pathogens, since cytotoxic T-cells play a major role in the defence against such infections. In young children wheezing episodes caused by rhinovirus infections are associated with the development of asthma [20]. Moreover, in more than 80% of asthma exacerbations viral infections are involved [21, 22], while respiratory viruses have also been observed in the airways in stable patients with asthma [23, 24]. In mice, CD8⁺ T-cells activated in response to viral infections can persist in the airways for several months [25], thereby causing lung damage [26]. This could contribute to airway remodelling leading to a decline in lung function.

Our study shows that CD8⁺-cell counts are relatively constant over time, whereas baseline CD4⁺- and CD3⁺-cell counts were not correlated with cell counts at 14 years of follow-up. We postulate that CD8⁺ cells might contribute to lung function decline as a result of their cytotoxic activity. A recent study in chronic obstructive pulmonary disease showed an association between granzyme B expression and small airway remodelling [27]. This is supported by our observation that granzyme B, as a marker of cytotoxic activity, was correlated with the preceding decline in lung function. An additional possibility is that CD8 cells are a major source of IL-13, which is an important cytokine in airway remodelling [28].

Aside from the predictive value of CD8 cells, we also found that CD4 cells were predictive of lung function decline and cross-sectionally related to lung function. However, these cells were not correlated to preceding lung function decline when measured after 14 years of follow-up, and were not constant over time. Since CD4 cells may contribute to activation of naïve CD8 cells, this finding may reflect the role of CD4 T-cells in (the early stages of) CD8 T-cell activation. However, it needs to be noted that the correlation was lost when decline in absolute FEV₁ was analysed.

CD3 counts were correlated to preceding lung function decline after 14 years of follow-up. No correlation between CD8⁺ cells and granzyme B could be found, but granzyme B was correlated with CD3⁺ cells (data not shown). This might indicate a role for CD3⁺CD4⁺ natural killer T-cells (iNKT cells), which were found to be increased in bronchoalveolar lavage-fluid from allergic asthmatic patients when compared with healthy controls [29]. Moreover, it was shown that granzyme B was increased in iNKT cells in peripheral blood in allergic asthmatic patients when compared with healthy controls [30]. This suggests that cytotoxic iNKT cells might be involved in the processes that lead to lung function decline in asthma.

The strength of our study is that it describes 14 years of prospective follow-up, using standardised methodologies. Moreover, this study focuses on the heterogeneity in lung function decline and describes subgroups that show a phenotype with a smaller or larger decline in lung function. The heterogeneity in lung function is considerable and might be affected by many factors. Therefore, we considered the possibility that the associations between inflammatory markers and lung function decline were explained by differences in medication use. However, the results of our regression analysis were not affected by the use of inhaled corticosteroids as demonstrated when including corticosteroid use as a covariate (table S2). Nevertheless, there are several potential limitations. First, this study was performed in a relatively small cohort, as several patients were lost to follow-up after 14 years or were no longer willing to participate. However, it is unlikely that this has markedly affected our conclusions since: 1) the sample size was adequate for repeated measures of bronchial biopsy cell counts; 2) no differences were observed in lung function and corticosteroid use between patients who participated in the follow-up study and those who did not (p>0.10); and 3) we did find correlations between long-term lung function decline and inflammatory markers in the airways. Secondly, in view of developments in immunohistochemistry and data on predominant inflammatory pathways in asthma since the early 1990s, the methods of staining at t=0 and t=14 years were not identical. Thirdly, during the follow-up period patients were receiving asthma treatment from their own general practitioner, and for some patients, a chest physician. After 14 years, three of the 19 patients were not suffering from any asthma symptoms at all and were not receiving any asthma medication, while two-thirds of the patients were receiving inhaled corticosteroids. Fourthly, while CD8 counts at baseline were correlated with CD8 counts after 14 years of follow-up, it cannot be concluded that the elevated CD8 counts at baseline and the elevated CD8 counts at the end of the study are a reflection of a continuous process and/or are the results of similar triggers. Finally, at baseline none of the patients were current smokers. However, after 14 years of follow-up, three of the 19 patients had become current smokers (16%), of which two patients had smoked 2 pack-years and one patient had smoked 5 pack-years. Because of the limited prevalence of smoking in this cohort, we cannot draw conclusions on the influence of smoking on the parameters and associations studied. However, it is relevant to note that the results of our regression analysis were not affected by smoking (despite its low prevalence) as shown when using smoking as a covariate in the analysis (table S2).

Despite the substantial variability of lung function decline in our cohort, the average lung function decline in this group of patients with mild-to-moderate asthma was still within normal limits. Lange et al. [1] found declines of 38 mL per year in self-reported asthma and 22 mL per year in healthy subjects, with large variability. In our study, pre- and post-bronchodilator FEV₁ declined by 21 and 31 mL per year, respectively. The decline we found is less pronounced, however, the decline in post-bronchodilator FEV₁ was substantial when compared with the decline in healthy individuals. An explanation for the limited lung function decline might be that the patients in our cohort suffer from relatively mild disease. Nevertheless, given the results of Cho et al. [14] and Tsitsiou et al. [13] and our observation that more pronounced lung function decline is associated with more airway inflammation it cannot be excluded that our observations are also of relevance for more severe cases of asthma.

Identification and characterisation of phenotypic subgroups of asthma is relevant in relation to the clinical course of the disease [2] and responses to treatment [7]. Even though clusters of patients based on clinical, cellular and gene expression data have been identified in mild-to-moderate asthma [31–33], such clusters have not yet been linked to long-term disease outcome. Therefore, phenotyping is likely to contribute to the development of targeted therapies, which could have benefits for the treatment of both acute and long-term asthma symptoms.