Reply to: Is the role of eosinophils in refractory chronic cough truly dismissed? Mepolizumab reduced blood and sputum eosinophils, but not eosinophil-derived neurotoxin

Extract

Reply to W. Zhan and co-workers: The recently published MUCOSA trial in the European Respiratory Journal, which evaluated the effects of mepolizumab on objective cough frequency in patients with refractory chronic cough (RCC), has prompted a re-evaluation of the role of eosinophils in cough pathophysiology [1].

Shareable abstract

Mepolizumab reduced blood and sputum eosinophils, but not eosinophil-derived neurotoxin https://bit.ly/4r8x9tb

Reply to W. Zhan and co-workers:

The recently published MUCOSA trial in the European Respiratory Journal, which evaluated the effects of mepolizumab on objective cough frequency in patients with refractory chronic cough (RCC), has prompted a re-evaluation of the role of eosinophils in cough pathophysiology [1]. The study examined whether mepolizumab improves cough in patients with asthma or non-asthmatic eosinophilic bronchitis (NAEB) and RCC with persistent airway eosinophilia despite inhaled corticosteroids (ICS), based on evidence that eosinophils co-localise with nerves [2] and are associated with heightened neuronal sensitivity following allergen exposure [3]. Contrary to expectations, mepolizumab did not improve objective or subjective coughs compared with placebo, despite reductions in circulating and airway eosinophils. This dissociation has prompted debate regarding the mechanistic relevance of eosinophils in RCC, or whether methodological factors explain the negative outcome.

W. Zhan and co-workers have raised concerns regarding airway target engagement and eosinophil phenotypic heterogeneity. We acknowledge that this is a small, single-centre proof-of-concept study (n=30), with modest median baseline airway eosinophilia (2.6%, range of up to 11.8%). Importantly, patients with asthma continued ICS therapy, which would be expected to partially suppress sputum eosinophil counts. Patients also underwent systematic assessment and treatment for other common cough drivers, including gastro-oesophageal reflux disease and upper airway cough syndrome, consistent with guideline-based RCC management.

A key point raised by W. Zhan and co-workers is that although mepolizumab rapidly and significantly reduced blood eosinophils, there was no statistically significant between-group difference in sputum eosinophils at week 14, despite significant reductions from baseline in the mepolizumab group. Sputum differential counts by microscopy lack the resolution to capture eosinophil heterogeneity or function; we therefore performed detailed flow cytometric immunophenotyping of eosinophils and measured sputum eosinophil degranulation products.

Blood and induced sputum were collected at pre-dosing (week 0) and post-dosing (week 14). Cells were measured for eosinophils by flow cytometry, and the sputum fluid phase was frozen for measurement of eosinophil degranulation products (eosinophil peroxidase (EPX), major basic protein (MBP), and eosinophil-derived neurotoxin (EDN)) by ELISA. Blood erythrocytes were lysed before staining. Blood and sputum leukocytes were stained for the enumeration of eosinophils (CD45⁺SSC^highCD16⁻CD15⁺; total eosinophils) with surface expression of CD123 (interleukin (IL)-3Ra) and CD62L (L-selectin) for categorisation into subsets of resident eosinophils (rEOS; CD62L⁺CD123⁻), and inflammatory eosinophils (iEOS; CD62L⁻CD123⁺), as previously described [4]. Eosinophils were also stained to measure surface expression of homing (CD49d, CD63, CD66b, CCR3 and CRTH2) and activation (CD125 and CD69) receptors. Blood and sputum total eosinophil levels were expressed as the frequency of the total CD45⁺ population, and as the number of cells per mL of blood and sputum. Eosinophil subsets categorised by surface markers were expressed as a percentage of the total eosinophil population. Pre-treatment-to-post-treatment delta values were compared between groups using the Mann–Whitney U-test.

Using flow cytometry at week 14, mepolizumab significantly reduced eosinophil levels (per mL and % of CD45 cells) in blood (p<0.0001 and p=0.0066, respectively) and in sputum (p=0.0002 and p=0.0021, respectively (figure 1a and e). Furthermore, when eosinophils were categorised as rEOS or iEOS based on the expression level of CD123 and CD62L, mepolizumab significantly reduced the frequency of blood iEOS from 0.79 to 0.038% (p=0.0045) and sputum iEOS from 21.16 to 9.85% (p=0.0048) (figure 1b and f), but not rEOS (figure 1c and g).

FIGURE 1.

Sputum (a–c) and blood (e–g) eosinophils are measured by flow cytometry. (a and e) Eosinophils were identified as CD45⁺SSC^HighCD1⁻CD15⁺ cells and expressed as a percentage of total CD45⁺ cells (leukocytes). (b and f) Inflammatory eosinophils (iEOS) are defined as CD45⁺SSC^HighCD1⁻CD15⁺CD62L⁻CD123⁺ cells. (c and g) Resident eosinophils (rEOS) are defined as CD45⁺SSC^HighCD1⁻CD15⁺CD62L⁺CD123⁻ cells. rEOS and iEOS are expressed as a percentage of total eosinophils. d) Induced sputum supernatant was used for measurements of eosinophil-derived neurotoxin (EDN). h) A Spearman correlation between the frequency of sputum iEOS and the level of EDN in sputum supernatant. Mean±sd of data are presented in squares, and individual data points are presented in circles. The Mann–Whitney U-test was used to compare pretreatment and post-treatment delta values between treatment groups.

Despite eosinophil reductions, sputum EDN levels remained unchanged in both the mepolizumab and placebo groups, and with no between-group difference (p>0.05) (figure 1d). Sputum EPX and MBP levels were below the level of detection. Although the study was not powered for exploratory end-points, mepolizumab consistently reduced total eosinophils and iEOS in blood and sputum without altering airway eosinophil degranulation, as indicated by sustained sputum EDN levels. A similar dissociation between eosinophil depletion and degranulation after mepolizumab treatment has been reported previously [5, 6] and is consistent with elevated EDN levels in other eosinophilic conditions [7, 8]. This dissociation suggests incomplete suppression of local eosinophil subsets during anti-IL-5 therapy. While an IL-5-dependent population is present in circulation, accumulates in inflamed tissue, and responds to treatment, a tissue-resident population may mature and persist via IL-5-independent pathways [5], continuing to release eosinophil granule proteins that contribute to inflammation and possibly sensitise nerves to cause cough [9]. This may explain why anti-IL-5 therapy depletes circulating eosinophils but only reduces tissue eosinophils by ∼50% [6]. However, whether or not EDN is still causing cough by sensitising nerves is uncertain, because objective chronic cough frequency declined over time in both the placebo and mepolizumab arms, despite stable EDN concentrations and undetectable EPX or MBP.

Mechanistically, the reduction in iEOS frequency that we observed in blood and sputum after mepolizumab treatment was driven primarily by a reduced frequency of eosinophils expressing CD123, while the frequency of eosinophils expressing CD62L remained unchanged. These findings are consistent with a prior report [5] demonstrating that CD123 expression on circulating eosinophils is reduced after mepolizumab treatment. Although CD62L and CD123 have been proposed as markers to define inflammatory eosinophils, these markers alone are unlikely to fully capture the functional heterogeneity of eosinophils across compartments, and these subsets have not been confirmed to exhibit functionally distinct features to date. Furthermore, in this study, we found no relationship between sputum iEOS and EDN levels (figure 1h). In contrast to our findings on CD123 expression, we demonstrate that the frequency of eosinophils expressing CD125 (IL-5Ra) significantly increases in both blood and sputum following mepolizumab treatment (p=0.0054 and p=0.041, respectively), consistent with receptor stabilisation in the context of ligand neutralisation [10].

While mepolizumab treatment also reduced the frequency of sputum eosinophils expressing the activation marker CD69 (p=0.0008) and showed a trend towards a reduction in the frequency of blood eosinophils expressing CCR3 and CD66b (p=0.054 and p=0.052, respectively), the functional significance of these changes is currently unknown. The proportion of eosinophils expressing CRTH2, CD49d, Siglec-8, and CD63 remained unchanged after mepolizumab treatment (p>0.05; data not shown).

In summary, mepolizumab significantly reduced blood and sputum eosinophils (total and iEOS) in patients with RCC, without a corresponding change in airway EDN levels. EDN is a relatively stable granule protein that reflects prior eosinophil activation and may persist in airway secretions despite reductions in eosinophil numbers [11]. Clarifying whether eosinophils contribute to RCC through degranulation-mediated effects on airway sensory nerves will require interventions that more directly modulate this pathway. Alternative peripheral and central neuro-inflammatory mechanisms that drive cough in these patients need to be considered. While these findings do not exclude a broader role for eosinophils in cough pathogenesis, they suggest that suppression of IL-5-dependent eosinophils alone is insufficient to modify the dominant cough-generating mechanisms in long-standing RCC. Patients with chronic cough and elevated blood or sputum eosinophils should continue ICS therapy, as cough may improve in this subgroup [12]. However, for patients with long-standing RCC despite ICS, targeting IL-5-dependent eosinophils is unlikely to produce meaningful improvements in cough outcomes.