Further evidence of a type 2 inflammatory signature in chronic obstructive pulmonary disease or emphysema

Abstract

Background:

There is increasing recognition of a type 2 (T2) inflammatory pattern in a subset of patients with COPD and/or emphysema, characterized by blood and airway eosinophilia. The mechanism underlying this is not well established. The recognition that CD125 (IL-5Ra) is expressed on some lung neutrophils as well as eosinophils in patients with asthma led us to speculate that CD125 may also be expressed on lung neutrophils in patients with COPD/emphysema.

Objective:

We sought to interrogate the expression of CD125 on lung neutrophils (and, when present, eosinophils) in patients with COPD/emphysema and identify a meaningful biomarker to predict neutrophil CD125 expression, including other markers of T2 inflammation.

Methods:

We obtained blood and bronchoalveolar lavage (BAL) samples from patients with physician-diagnosed COPD/emphysema undergoing a clinically indicated bronchoscopy.

Results:

We found that a highly variable percentage of BAL neutrophils indeed expressed surface CD125 (0–78.7%), with obvious clustering of CD125high and CD125low patterns. No correlation was found with clinical characteristics, blood or BAL eosinophil or neutrophil counts, BAL cytokines, or BAL eosinophil CD125 expression.

Conclusion:

We conclude that, similar to asthma, lung neutrophils from patients with COPD display IL-5Ra (CD125) on their surface. This along with the frequent presence of IL-4 and IL-5 in airway fluid further suggests a possible role of the T2 pathway in contributing to COPD severity.

Introduction

Although most patients with chronic obstructive pulmonary disease (COPD) and emphysema have neutrophils and CD8⁺ T cells in their airways, there is increasing recognition of eosinophilic inflammation in 10–40% of patients with COPD¹. This COPD/emphysema endotype is characterized by increased blood and airway eosinophilia, with further increases in eosinophil counts during acute COPD exacerbations. Although the mechanism underlying this eosinophilic pattern in COPD/emphysema is not well established, it is thought to represent the presence of a T2^high process associated with elevated expression of interleukin (IL-5)^2–9. While some studies have demonstrated increased bronchial epithelial and sputum expression of T2 genes such as the eosinophilic membrane protein chloride channel accessory 1 (CLCA1), the eosinophil-specific chemotaxin CCL26 (eotaxin-3), and IL-13 in T2^high versus T2^low COPD/emphysema^{10, 11}, others have failed to identify significantly increased gene expression of IL-5 and IL-13 compared to non-obstructive airways disease controls¹². Clinically, the recognition that both T2^high and T2^low patterns of inflammation might be present has opened the possibility for use of inhaled corticosteroids¹⁰ or T2/IL-5-targeting biologics⁹ in patients with COPD or emphysema.

IL-5 is a T2 inflammatory cytokine whose receptor, IL-5R, is known to be expressed on eosinophils, along with B cells, airway epithelium, mast cells, and basophils^{13, 14}. In recently published work, we also confirmed expression of the IL-5 binding component of the receptor, IL-5Rα (CD125), on blood and bronchoalveolar lavage (BAL) neutrophils of humans with asthma and other atopic diseases¹⁵ as well as mice infected with influenza A¹⁶. This finding has been replicated by other groups including reports of CD125 expression on neutrophils in the lungs of dust mite sensitized mice¹⁷, in the lungs of horses with heaves (an asthma-like condition)¹⁸ and in the lungs of mice after severe hemorrhagic shock and tissue trauma¹⁹, as well as on human neutrophils in the setting of mycobacterial infection and sepsis^{20, 21}. Differential methylation of the IL-5RA gene on neutrophils from patients with asthma has also been reported²².

Importantly, we previously demonstrated that surface CD125 expression increases following neutrophil activation and – at least in individuals with asthma – is higher on lung compared to blood neutrophils¹⁵, suggesting potential upregulation upon exposure to an inflammatory milieu.

Currently, biologic medications targeting the IL-5/IL-5R pathway are approved only for the treatment of severe eosinophilic asthma, defined typically as a blood absolute eosinophil count of greater than 150–300 cells/μl²³. Trials of IL-5/IL-5R-targeting biologics in COPD have demonstrated mixed results regarding benefit in reducing the rate of exacerbations and improving lung function¹ and the subgroup of patients with COPD that are most likely to respond to anti-IL-5 therapy is not defined. We hypothesize that those patients with COPD/emphysema with robust expression of CD125 on their lung neutrophils (along with eosinophils) may define a subgroup more responsive to IL-5/IL-5R targeting biologics. While defining a blood biomarker is attractive due to ease of obtaining patient samples in a clinical setting, our prior work highlights the importance of directly evaluating lung neutrophils as there is significant discordance between blood and airway neutrophil CD125 expression¹⁵. Therefore, in this work we interrogated the expression of CD125 on lung neutrophils (and, when present, eosinophils) from a cohort of patients with a history of smoking, physician-diagnosed COPD/chronic bronchitis, and either fixed airflow obstruction on spirometry or emphysema on chest CT who were undergoing a clinically indicated bronchoscopy/BAL. We next correlated CD125 expression with clinical characteristics and BAL cytokine levels.

Methods

Subjects.

20 subjects were enrolled between October, 2020 and October, 2021. 3 subjects were later excluded from some of the analyses due to incomplete laboratory data being available. The study population included subjects undergoing clinically indicated bronchoscopy for diagnostic evaluation of airspace disease that included an endobronchial ultrasound-guided fine needle aspiration of a lung nodule or lymph node. For research purposes, BAL fluid was obtained from the opposite lung. Subjects were required to have physician-diagnosed COPD or emphysema for at least 1 year based on the ATS-ERS 2004 definition²⁴, including a history of smoking along with presence of one or more of the following: FEV₁/FVC ratio ≤0.70, reduced DLCO (≤70% predicted), and/or chest CT demonstrating centrilobular emphysema. Patients with a current diagnosis of asthma or other medical condition that in the opinion of the investigator could influence presence of granulocytes in the airway, non-smokers (<10 pack-years) with a history of asthma, recent change in COPD controller medication regimen, positive pregnancy test at the time of bronchoscopy, or recent (within 1 year) administration of a biologic agent were excluded. Patients with a recent COPD exacerbation or symptoms of recent symptomatic infection (change in sputum volume or color) within 30 days of bronchoscopy were also excluded. All participants provided informed consent under protocols approved by the University of Virginia Investigational Review Board for Human Subjects Research (IRB-HSR #18975).

Flow cytometry.

BAL fluid was placed on ice immediately following collection and processed promptly. BAL fluid was centrifuged and cell pellets obtained. Cells were blocked with mouse IgG (Lampire Biological, Pipersville, PA) and Fc block (BD Pharmingen; Sparks, MD), washed and stained with Live/Dead fixable violet (ThermoFisher Scientific, Waltham, MA), PerCPCy5.5-conjugated anti-CD16 (BD Bioscience), APC conjugated anti-Siglec 8 (BioLegend, San Diego, CA) and PE-conjugated anti-CD125 (BD Bioscience) or mouse isotype control. Neutrophils were identified as live, SSC^highCD16⁺ Siglec8⁻ and eosinophils were identified as liveSSC^highCD16⁻ Siglec8⁺ as shown previously¹⁵. Fluorescence minus one (FMO) staining was used to set gates. BAL fluid granulocyte patterns were defined as isolated eosinophilia (≥1% eosinophils and <6% neutrophils), isolated neutrophilia (≥6% neutrophils and <1% eosinophils), mixed granulocytic (≥1% eosinophils and ≥6% neutrophils), or pauci-granulocytic (<1% eosinophils and <6% neutrophils) according to criteria previously established and validated in both children and adults^{25, 26}.

BAL cytokine levels.

Levels of IFN-γ, IL-2, IL-17A, IL-10, IL-4 and IL-5 in BAL fluid were measured with the 6-plex Milliplex MAP^® human T cell panel (Millipore, Ltd. Billerica, MA) using a Luminex^® MAGPIX system (Luminex, Austin, TX). Samples were assayed in duplicate, and the cytokine concentrations were calculated against the standard curve according to the manufacturer’s instructions. For samples in which measurements were below the lower limit of assay detection (LOD), a value was imputed equal to 0.5 * LOD.

Statistics.

A complete description of the statistical plan is provided in the online supplement (eAppendix 1). Briefly, categorical data were summarized by frequencies and relative frequencies as appropriate. Continuous scaled data were distributed in a non-Gaussian pattern and thus were summarized by the median, the interquartile range, and the range of the empirical distribution. The bivariate relationships between CD125 expression on BAL granulocytes, numbers of blood and BAL granulocytes and BAL cytokines were tested with the non-parametric Spearman rank correlation coefficient (r_s). A two-sided p<0.05 decision rule was used as the null hypothesis rejection criterion.

Results

20 subjects undergoing clinically indicated bronchoscopy were enrolled, with 17 subjects having sufficient samples available for all study procedures performed included in the final analyses (Table 1). The majority of the subjects were Caucasian (80%) and male (75%) with a median age of 68 years (range 53 – 76 years). All subjects were long-term smokers who had physician diagnosed COPD/chronic bronchitis, along with the presence of either radiologically-determined emphysema or fixed obstructive lung disease. 45.0% of subjects had a peripheral blood absolute eosinophil count (AEC) of 150 cells/μl or greater (Table 1, Figure 1A).

Table 1:

Subject Characteristics

Subject characteristics
Male, n (%)	15 (75.0)
Caucasian, n (%)	16 (80.0)
Non-Hispanic, n (%)	20 (100)
Age in years, median [IQR]	68.0 [62.0, 70.5]
FEV1 % predicted, median [IQR]	63.0 [49.5, 69.5]*
FEV1/FVC, median [IQR]	0.63 [0.53, 0.66]*
DLCO % predicted, median [IQR]	70.0 [50.0, 76.3]*
Peripheral blood AEC (cells/ul), median [IQR]	135 [83, 208]
AEC ≥150 cells/ul, n (%)	9 (45.0)
Peripheral blood ANC (cells/μl), median [IQR]	5375 [4515, 8365]
** SABA treatment, n (%)	14 (70%)
** LAMA treatment, n (%)	6 (30%)
** LABA treatment, n (%)	7 (35%)
** ICS treatment, n (%)	6 (30%)

^*Not all data were available on all subjects

^**Individual subjects may have been treated with combinations of inhaled medications

Abbreviations: AEC, absolute eosinophil count; ANC, absolute neutrophil count; DLCO, diffusing capacity of the lung for carbon monoxide; FEV1, forced expiratory volume, 1 second; FVC, forced vital capacity; ICS, inhaled corticosteroid; LABA, long acting bronchodilator; LAMA, long acting muscarinic antagonist; SABA, short acting bronchodilator

Figure 1. Blood and BAL granulocyte patterns.

Blood (A) and BAL (B) granulocyte patterns. BAL neutrophils (C) and eosinophils (D) were correlated with blood counts using the Spearman rank correlation coefficient (r_s). Dotted line in (A) represents upper limit of normal for eosinophil counts.

BAL fluid granulocyte patterns were defined as isolated eosinophilia, isolated neutrophilia, mixed granulocytic, or pauci-granulocytic as described above^{25, 26}. 2 subjects (11.8%) had a pauci-granulocytic pattern of inflammation, 6 (35.3%) had an isolated eosinophilic pattern, 1 (5.9%) had an isolated neutrophilic pattern and 8 (47.1%) had a mixed granulocytic pattern of inflammation. BAL fluid had a median of 8.4% neutrophils (IQR: [0.9, 16.5%]) and 12.8% eosinophils (IQR: [1.9, 66.5%]). Remarkably, in seven subjects (41.2%) 50% or greater of BAL granulocytes were eosinophils (Figure 1B). There was no correlation between blood and BAL eosinophil or neutrophil percentages (Figure 1C, D).

Surface CD125 expression on granulocytes.

CD125 expression was detected on a median of 9.5% of neutrophils (range 0–78.7%) and 21% of eosinophils (range 0–69.7%; Figure 2). There was no relationship between the percentage of neutrophils expressing CD125 and age, race, or gender (eTable 1). Neither the blood absolute eosinophil count (AEC), absolute neutrophil count (ANC), nor the BAL eosinophil percentage correlated with the percentage of CD125 positive BAL neutrophils (eTable 1). Similarly, the percentage of CD125 positive BAL eosinophils did not correlate with the percentage of CD125 positive BAL neutrophils (eTable 1).

Figure 2. BAL granulocyte CD125 expression.

(A) CD125 expression on gated neutrophil and eosinophil populations was determined via staining with PE-conjugated anti-CD125. (B) Cytokine levels in BAL fluid were measured using a 6-plex Milliplex MAP human T cell panel using a Luminex assay.

Cytokine Expression.

Among BAL cytokines, the highest concentration detected was IL-4, followed by IL-10, IFN-γ and IL-5. IFN-γ, IL-10 and IL-5 were detectable in 100% of samples, while IL-2 was detectable in only 5.9% (n=1) of samples and IL-4 and IL-17A in 58.8% and 35.3% of samples, respectively (Table 2 and Figure 2B). We calculated a partial Spearman correlation between the percentage of neutrophils expressing CD125 and several correlate variables including age, FEV₁, blood AEC, blood ANC, BAL cytokine levels (excluding IL-2 since most samples did not have detectable levels), BAL neutrophil and eosinophil percentage, and the percentage of eosinophils expressing CD125. After adjusting for all remaining partial correlate variables and multiple hypothesis testing, there was no significant correlation identified (Table 3).

Table 2:

BAL fluid cytokine expression

Cytokine (pg/ml)	Median	Min	Max	Measurements Above LOD, n (%)
IL-4	2.80	0.84*	276.83	10 (58.8)
IL-10	0.81	0.32	4.48	17 (100)
IFN-γ	0.40	0.22	1.72	17 (100)
IL-5	0.33	0.21	5.59	17 (100)
IL-17A	0.11	0.11*	0.77	6 (35.3)
IL-2	0.13	0.13*	0.26	1 (5.9)

^*For samples in which measurements were below the LOD, a value was imputed equal to 0.5*LOD

Abbreviations: BAL, bronchoalveolar lavage; IL, interleukin; LOD, lower limit of assay detection

Table 3:

Correlation of clinical features and BAL cytokines with BAL neutrophil CD125 expression

Correlate	Spearman Partial Correlation (rsp) [95% CI]	P-value
Age	−0.36 [−0.78, 0.29]	0.274
FEV1	0.38 [−0.21, 0.77]	0.208
Blood AEC	0.07 [−0.32, 0.43]	0.742
Blood ANC	−0.23 [−0.59, 0.22]	0.311
IFN-g	0.08 [−0.29, 0.43]	0.675
IL-10	−0.02 [−0.43, 0.39]	0.911
IL-17A	0.15 [−0.26, 0.52]	0.479
IL-5	−0.02 [−0.39, 0.35]	0.916
IL-4	−0.42 [−0.78, 0.12]	0.119
% BAL neutrophils	−0.15 [−0.60, 0.36]	0.567
% BAL eosinophils	0.11 [−0.3 7, 0.55]	0.661
% CD125+ BAL eosinophils	0.26 [−0.30, 0.69]	0.688

Abbreviations: AEC, absolute eosinophil count; ANC, absolute neutrophil count; BAL, bronchoalveolar lavage; FEV1, forced expiratory volume, 1 second; IL, interleukin

Discussion

Our previous work, as well as that of others, has established high levels of expression of the IL-5R subunit CD125 on the lung neutrophils of individuals with asthma^{15, 16}. The expression on airway neutrophils of a T2 cytokine receptor that is classically found on eosinophils (the so-called “eosinophil-ification” of the neutrophil), may be another example of the inflammation driving the mixed T1/T2 phenotypes that are increasingly being recognized in airway disease. This is seen both in the subset of patients with COPD that have eosinophilia (T2^high COPD) and the subset of patients with severe asthma that have neutrophilia (T2^low asthma), both disease states which are mechanistically poorly understood and can be difficult to treat^{6, 27, 28}. This “eosinophil-ification can include the upregulated expression on neutrophils of other surface molecules generally ascribed to the eosinophil in allergic diseases and asthma including the adhesion molecular vascular cell adhesion moledule-1 (VLA-4/CD49d)²⁹ and the cysteinyl leukotriene receptors³⁰. This led us to speculate that the IL-5R may be expressed on lung neutrophils in patients with COPD. Our previous studies demonstrated that CD125 receptors on neutrophils are functional⁸, suggesting that patients with COPD/emphysema with robust expression of CD125 may define a subgroup in whom clinical trials addressing the efficacy of IL-5/IL-5R targeting biologics would be appropriate.

There are several limitations of our study that should be noted. Reflecting the difficulty in subject recruitment for this kind of study, our cohort size was relatively modest (n=20) and BAL fluid could only be fully analyzed in 17. This is unlikely to have impacted our conclusions insofar as the study was more than sufficiently powered to demonstrate our primary goals regarding expression of CD125 on neutrophils. And given the wide variability of CD125 observed on lung neutrophils it is implausible that screening more subjects could have identified more than the stochastic pattern already displayed in Figure 1A. Since subjects were referred specifically for only a single visit for the clinical bronchoscopy, we typically had limited clinical information that might have been useful in performing further correlations, e.g., annual exacerbation rate, presence of allergic sensitization on skin prick or serum IgE testing, and inhaled/oral corticosteroid equivalent dose. As a final limitation, it should be appreciated that all of our subjects were life-long smokers with either airflow obstruction or emphysema diagnosed by CT and thus met criteria for COPD (chronic bronchitis and/or emphysema). However, because of widespread safety restrictions put in place during the SARS-CoV2 pandemic we were unable to assess maximum airflow reversibility. It is thus possible that at least some of our subjects, especially those with the higher eosinophil counts, might be considered as having the asthma-COPD overlap syndrome (ACOS).

In our cohort of patients with physician-defined COPD undergoing clinically indicated bronchoscopy, we found, similar to previously published work, that 47.1% of subjects had a peripheral blood AEC of 150 cells/μl or greater and 82.4% of subjects had either isolated eosinophilic or mixed granulocytic patterns of inflammation in their BAL fluid (Figure 1A, B). We show, again similar to previously published work, that the blood AEC and ANC do not predict the BAL granulocyte pattern (Figure 1C, D)^{8, 31}. This highlights the importance of directly evaluating lung neutrophils when possible and challenges the common practice of reliance on the presence of blood eosinophilia to select patients/subjects for inclusion in studies and/or treatments. It is imperative to appreciate, however, that bronchoscopy will never be an acceptable screening procedure as a basis for initiating endotype-specific personalized treatments and – as is also well appreciated in asthma – there is a pressing need for identifying utilitarian biomarkers that can predict responsiveness to our increasing armamentarium of biologics.

Similar to our asthma studies¹⁵, a high prevalence of neutrophils express surface CD125 in COPD. This expression was widely variable (0–78.7%) with obvious clustering of CD125^high and CD125^low patterns (Figure 2A). We were eager to identify specific features which predicted a higher level of BAL neutrophil CD125 expression. However, no association with age, race, or gender, blood AEC, blood ANC, BAL granulocyte percentages, or BAL eosinophil CD125 expression was observed (eTable 1). Interestingly, BAL eosinophils similarly dichotomized into CD125^high and CD125^low expressing subjects (Figure 2B), but there was no association of BAL eosinophil CD125 expression with blood or airway eosinophil counts. Similarly, eosinophil CD125 expression was totally unpredictive of neutrophil CD125 expression. While blood eosinophils are defined by their expression of CD125, the loss of this receptor on lung eosinophils is well appreciated in asthma, reflecting in part the action of matrix metalloproteinases, and the current studies suggest similar effects in COPD³².

Finally, we explored the expression of BAL cytokines as a possible predictor of neutrophil CD125 expression. Not surprisingly we identified a fair representation of type 1 (IFN-γ) and type 3 (IL-17) cytokines. However, T2 cytokines (IL-4 and IL-5) were both highly expressed and, indeed, were identified in most samples (Figure 2B). The co-expression of IL-5 with IL-5 receptor (IL-5R)-expressing neutrophils supports its having a pro-inflammatory role. We did not identify any correlation of IL-5 cytokine levels with IL-5Rα (CD125) expression which perhaps is not surprising as IL-5 measurements would be confounded by the sequestration of this cytokine after binding to its receptor. While IL-4 levels were amongst the highest cytokine concentrations we observed, these numbers did not correlate with CD125 expression (Table 3). While we have previously published data that IL-4 has pro-inflammatory activities with respect to neutrophils³³, IL-4 is also recognized as having anti-inflammatory activities especially in regards to expression of innate inflammatory cytokines (e.g., IL-1, TNF, IL-6).

In summary, these studies demonstrate that, similar to asthma (as well as to mycobacterial infections, trauma, and sepsis), lung neutrophils from patients with COPD may display IL-5Rα (CD125) on their surface. The frequent presence of IL-4 and IL-5 in airway fluid further suggests a possible role of the T2 pathway in contributing to COPD severity. However, neither the number or pattern of blood/airway eosinophils or neutrophils, nor the BAL cytokine pattern was able to predict neutrophil CD125 expression in COPD with a T2 inflammatory milieu. Future studies are needed to identify a meaningful and practical biomarker that would predict which COPD patients could potentially benefit through the targeting of this mechanism.

Abbreviations:

AEC

absolute eosinophil count

ANC

absolute neutrophil count

BAL

bronchoalveolar lavage

COPD

chronic obstructive pulmonary disease

CT

computed tomography

DLCO

diffusing capacity of the lung for carbon monoxide

FEV1

forced expiratory volume, 1 second

FMO

fluorescence minus one

FVC

forced vital capacity

IFN

interferon

IgG

immunoglobulin G

IL

Interleukin

IL-5R

IL-5 receptor

LOD

lower limit of assay detection

SSC

side scatter

T2

type 2