Authors: Christine M. Freeman, Jeffrey L. Curtis, Annette T. Hastie
Categories: Editorials
Source: American Journal of Respiratory and Critical Care Medicine
Asthma practice guidelines recommend considering eosinophil counts from blood, induced sputum, or both for treatment decisions (1). However, controversy persists on the relevance of eosinophils to chronic obstructive pulmonary disease (COPD) outcomes. Association of eosinophils with greater risk of COPD exacerbations has been either claimed for blood (2, 3) or refuted for both blood and sputum (4). Although 15–37% of patients with COPD persistently have >2% blood eosinophils (5), low rather than high blood eosinophils were associated with both greater emphysema and lung function decline in Global Initiative for Chronic Obstructive Lung Disease group D disease (6). A recent systematic review concluded that blood eosinophil counts cannot predict COPD clinical outcomes (7).
Studies of lung samples also do not support a clearly deleterious role of eosinophils in COPD. Numbers of eosinophils in lungs resected from smokers did not differ in those patients without versus with COPD (8). Eosinophils in BAL were significantly increased in current (but not former) smokers with COPD relative to ever-smokers without COPD or never-smokers (9).
However, because these studies assayed eosinophils as a homogeneous cell type, the recognition of eosinophil subsets (10) necessitates their reevaluation. In both mice and humans, eosinophils exist either as inflammatory eosinophils (iEos) having high expression of the IL-5 receptor α-chain (IL-5Rα, CD125) or as tissue-resident eosinophils (rEos), which have regulatory properties (11, 12). Distinguishable by surface phenotype, these subsets differ in developmental pathways, anatomic localization, and roles (10).
In this issue of the Journal, Cabrera López and colleagues (pp. 155–162) examined phenotypes and proportions of these eosinophil subsets in peripheral blood in four discovery cohorts (n = 10 each): participants with allergic asthma not receiving biologicals, ever smokers with or without COPD, and healthy volunteers (13). Using a published flow cytometric approach (11), they distinguished iEos (Siglec-8^+^ , CD62L^low^, CD123 [IL-3 receptor]^high^) from an alternative phenotype (Siglec-8^+^ , CD62L^high^, CD123^low^) termed rEos, although detected in blood, for consistency with previous literature. Their results showed significantly higher iEos proportions in participants with asthma than in participants with COPD (25% vs. 0.5%) matched for age, sex, and FEV1 percent predicted. Disparate eosinophil subset proportions were confirmed in validation asthma (n = 17) and COPD (n = 59) cohorts. In the COPD validation cohort, proportions of iEos versus rEos did not differ across Global Initiative for Chronic Obstructive Lung Disease stages or by inhaled corticosteroid use. These findings were interpreted to suggest that in COPD, eosinophils (which are primarily rEos) play a lesser role in disease progression than that of iEos in asthma (13). Alternatively, peripheral blood might not accurately reflect differences in airway eosinophil subsets, which were not investigated.
Importantly, in the asthma discovery cohort, the mean fluorescence intensity of IL-5Rα was significantly greater on iEos than on rEos. This difference was not seen in the COPD discovery cohort, though their very low number of iEos may have compromised receptor detection. Conversely, although expression of IL-5Rα by rEos was significantly higher in COPD than in asthma in the validation cohorts, it was still lower than that of iEos in both COPD and asthma (13). IL-5Rα expression differences between eosinophil subsets in the two diseases are noteworthy, given the availability of biologicals that therapeutically target IL-5 directly or via its receptor (14).
In patients with asthma having elevated blood eosinophils, mepolizumab, reslizumab (anti–IL-5), and benralizumab (anti-IL-5Rα) have demonstrated clinical benefits, including reduced exacerbations (15–17), improved quality of life (15, 17), and increased lung function (16, 17). In large trials of eosinophilic COPD, in contrast, mepolizumab and benralizumab significantly reduced blood eosinophils but had minimal effects on exacerbation rates, health status, or lung function (18, 19). Such divergent responses in asthma versus COPD to therapies targeting the IL-5 axis may be partially explained by the high proportion of blood iEos in asthma versus the low proportion in COPD observed by Cabrera López and colleagues (13). Their findings also extend previous data demonstrating significantly increased eosinophils in bronchial biopsies from patients with asthma and in exacerbations of chronic bronchitis, but increased numbers of IL-5^+^ eosinophils only in asthma (20). In addition, they shed light on the finding that although current smokers with COPD in the SPIROMICS (SubPopulations and InteRmediate Outcome Measures In COPD Study) cohort had the most BAL eosinophils, compared with former smokers with COPD, ever smokers without COPD, and never smokers, they had the lowest IL-5Rα expression (9). That finding is compatible with abundant rEos in the lungs of some current smokers. However, as correctly stated (13), expression of a single receptor cannot accurately identify rEos, which were not recognized in humans when the SPIROMICS immunophenotyping protocol was developed (21).
Interestingly, an alternative approach that more broadly targets type 2 inflammation appears effective in both asthma and COPD. Dupilumab binds the shared component of the receptors for IL-4 and IL-13, inhibiting both signaling pathways (22, 23). In asthma, dupilumab reduced exacerbations and improved lung function and quality of life, regardless of baseline eosinophils (24). A recent preliminary report found similarly improved outcomes in patients with COPD with blood eosinophils >300 cells/μl (25). In the asthma trial (24), blood eosinophil counts transiently increased in 13% of dupilumab recipients, without clinical adverse events or consequences. This increase is consistent with the known role of IL-4 and IL-13 to promote eosinophil trafficking into tissues (5). Whether a similar increase in eosinophil counts occurs in COPD has not yet been reported, and whether iEos and rEos differ in expression of the shared IL-4/IL-13 receptor α-chain is unknown. It seems clear that eosinophils in asthma and COPD have distinct differences but potentially overlapping characteristics.
The Cabrera López and colleagues study (13) unsurprisingly has limitations. Despite including validation cohorts, a significant strength, overall participant numbers remain small. The less rigorously matched validation cohorts did not replicate all significant observations from the discovery cohort, such as increased expression by rEos of CD11b or of CD123, a signature receptor of that subset (11). Most cogently, whether peripheral blood cells having the surface phenotype of rEos exhibit the regulatory properties of lung rEos (11, 14) is unclear. Second, as observed in SPIROMICS, eosinophils in sputum (but not in blood) showed significant associations with lower lung function (FEV1 in liters); higher quantitative computed tomography measures of emphysema, air trapping, and functional small airway disease; and greater exacerbations in the year before enrollment (26). Third, that current smoking alters eosinophil proportions in BAL but not in matched peripheral blood (9) is relevant to the significant differences in current smoking between participants with COPD and participants with asthma in the Cabrera López and colleagues study, an unexplored potential confounding variable. Collectively, those published data imply marked differences between airway and peripheral blood eosinophils.
In summary, this paper provides important new insights into differences in the roles of eosinophils between COPD and asthma, but it predictably raises new questions, particularly regarding anatomic compartments. Rather than simply counting eosinophils, future studies should characterize subsets in both circulating and respiratory compartments. Clinical trials of existing or novel biologics would benefit from in-depth eosinophil phenotyping to identify the best participants for personalized therapies.