Beyond Blood Eosinophils: Target-Tier Heterogeneity and Patient Enrichment in COPD Type 2 Inflammation

Abstract

Chronic obstructive pulmonary disease (COPD) exhibits considerable clinical and biological heterogeneity, and frameworks centered on symptom burden, airflow limitation, and exacerbation risk do not fully explain variation in inflammatory pathways or responses to targeted therapies. Blood eosinophils (blood EOS) provide an accessible clinical biomarker for estimating inhaled corticosteroid (ICS) benefit, stratifying exacerbation risk, and enriching biologic trials in COPD. We position blood EOS as an entry biomarker within a target- and context-dependent framework because peripheral counts incompletely represent airway epithelial activation, local chemotaxis, mucus pathology, tissue EOS activation, or mixed inflammatory microenvironments. This focused narrative review integrates biologic trials, biomarker analyses, and human airway and lung-tissue studies to map type 2-related signals across target tiers. IL-5/IL-5R targeting tests the EOS effector-cell compartment. IL-4Rα blockade supports the therapeutic relevance of an IL-4/IL-13-associated epithelial–chemokine–tissue response program in selected enriched populations. IL-33/ST2 evidence supports therapeutic relevance in COPD, although efficacy remains intervention- and context-dependent, whereas TSLP evidence remains exploratory. Collectively, the evidence supports interpreting COPD type 2 inflammation through target tier, tissue pathway, and patient context. Future studies should test target-directed composite enrichment and prospectively link target perturbation, pathway readouts, and clinical responses.

Evolving Understanding of Type 2 Inflammation in COPD

Chronic obstructive pulmonary disease represents a major source of chronic respiratory disease burden, leading to persistent symptoms, recurrent exacerbations, increased hospitalization risk, and elevated mortality.^1,2 Patients typically present with chronic cough, sputum production, exertional dyspnea, recurrent acute exacerbations, and progressive lung function decline, with a pathological basis involving small airway disease, mucus retention, parenchymal destruction, sustained inflammatory responses, and multisystem comorbidities.³ Management frameworks centered on lung function, symptom burden, and exacerbation risk have provided a foundation for COPD treatment,¹ yet they still cannot fully account for inter-patient differences in disease trajectory, inflammatory status, and treatment response.^3–5 Accordingly, COPD is more appropriately understood as a heterogeneous disease involving multiple clinical phenotypes, biological pathways, and treatable traits, rather than a single airflow-limitation entity.^6,7

Earlier understanding of COPD inflammation predominantly emphasized neutrophils, macrophages, CD8+ T cells, and Th1/Th17 responses.⁸ Evidence from blood EOS, induced sputum EOS, ICS responsiveness, and eosinophilic inflammation as a treatable trait has supported recognition of type 2-related inflammation as a clinically relevant dimension in a subset of patients with COPD.^9,10 Type 2 inflammation-associated transcriptional signatures in COPD and associations between blood EOS, fractional exhaled nitric oxide (FeNO), and lung function decline in chronic airway disease populations provide further support for the clinical relevance of these signals.^11,12 In this setting, blood EOS is a practical and low-cost biomarker for exacerbation risk stratification, assessment of likely ICS benefit, and enrichment of biologic trials.^9,13 Rather than defining a discrete eosinophilic COPD subtype, blood EOS should prompt pathway-level interpretation in relation to the clinical question and treatment target, while its relationship to airway and tissue inflammation requires separate consideration.^9,10,14,15

Literature Search and Evidence Selection

This focused narrative review was based on targeted searches of PubMed/MEDLINE, ClinicalTrials.gov, and relevant reference lists. The literature search was initially conducted before submission and was updated for revision to include records available through July 10, 2026. The main search terms included combinations of “COPD” or “chronic obstructive pulmonary disease” with “type 2 inflammation”, “blood eosinophils”, “sputum eosinophils”, “FeNO”, “biomarker”, “biologic therapy”, “patient enrichment”, “IL-5”, “IL-5R”, “IL-4Rα”, “IL-13”, “dupilumab”, “IL-33”, “ST2”, “TSLP”, “mucus plugs”, “chronic bronchitis”, “smoking”, and “asthma-COPD overlap”. Individual drug and trial names were also searched to identify primary reports, supplementary materials, and registry records.

The search focused mainly on studies published from January 1, 2010 through July 10, 2026, while earlier landmark trials or mechanistic studies were included when necessary for context. Priority was given to COPD guidelines, randomized controlled trials of biologics, biomarker analyses, pooled analyses, systematic reviews, meta-analyses, and human airway or lung tissue studies. Studies were selected if they informed the interpretation of blood or airway eosinophils, target-specific biologic therapy, epithelial alarmins, mucus pathology, or biomarker-based patient enrichment in COPD. Evidence from asthma, animal models, or cell studies was used only as supportive mechanistic context when direct COPD evidence was limited. Because this was a focused narrative review rather than a systematic review or meta-analysis, no formal quantitative synthesis or risk-of-bias assessment was performed.

Blood EOS as a Patient Enrichment Entry Point, Not a Complete Pathway Readout

Blood EOS is readily obtainable and repeatable and is used in COPD to estimate exacerbation risk and the likelihood of benefit from ICS.^9,13 It is also commonly used as an enrollment or enrichment variable in biologic trials.^16,17 However, its clinical interpretation depends on the threshold applied, sampling time point, recent treatment exposure, and disease setting.⁹ The following considerations are therefore important when blood EOS is used to inform treatment decisions or trial eligibility.

Blood EOS is best interpreted as a continuum of the probability of ICS benefit for exacerbation prevention, rather than a dichotomous measure.⁹ GOLD generally regards <100 cells/μL as a reference threshold below which ICS benefit is unlikely and ≥300 cells/μL as a threshold above which benefit is more likely and an ICS-containing treatment strategy is more strongly supported. Values of 100–300 cells/μL require consideration of prior moderate-to-severe exacerbations, exacerbation-related hospitalization, pneumonia risk, current inhaled therapy, and ICS exposure.¹⁸ These thresholds are probabilistic aids to treatment decisions, not diagnostic boundaries.⁹ In COPD biologic trials, blood EOS ≥300 cells/μL has also commonly been used as an enrichment threshold, but a single qualifying value should not be equated with longitudinally stable pathway activity.^16,17

Sampling time points and treatment exposures further add to the complexity of blood EOS interpretation.⁹ A single blood EOS measurement is more akin to a screening signal; repeated elevations during stable periods provide stronger support for a relatively sustained treatment-relevant trait. The temporal variability of higher blood EOS counts also suggests that a single time point should not be directly equated with long-term pathway status.¹⁹ Long-term ICS exposure, ICS withdrawal, and recent systemic corticosteroid use may alter blood EOS levels. During acute exacerbations, blood EOS is also shaped by the trigger, systemic corticosteroid treatment, and event-related inflammatory fluctuation; acute values should not be extrapolated directly to stable-phase pathway status.⁹ This means that the clinical significance of blood EOS depends not only on the numerical value itself but also on the treatment context and disease stage in which that value is observed.

Human cohort and phenotyping studies indicate that blood EOS and local airway eosinophilic inflammation are not fully concordant. SPIROMICS analyses demonstrated divergent associations of blood and sputum EOS with COPD clinical features;¹⁴ the UCLA COPD Phenotyping Study further showed that blood EOS and sputum EOS had limited classification agreement for type 2 inflammation.¹⁵ This discordance likely reflects genuine biological differences between the peripheral circulation, the airway lumen, and the lung tissue microenvironment. Peripheral blood EOS primarily reflects bone marrow production and circulating pool levels, whereas airway or tissue EOS is additionally influenced by local chemotaxis, tissue retention, cellular activation, and inflammatory trigger context, none of which can be fully captured by routine peripheral blood counts.²⁰

The value of blood EOS lies in context-specific interpretation of thresholds, longitudinal stability, treatment exposure, and blood–airway/tissue discordance. The following section considers how airway and tissue pathways can refine interpretation of this signal for different biologic targets.^9,10 These concepts are summarized in Figure 1.

Figure 1.

Clinical interpretation of blood eosinophils in COPD. (A) shows how single versus repeated measurements, corticosteroid exposure, and exacerbation or infection affect blood eosinophil (EOS) interpretation. (B) illustrates imperfect concordance among peripheral blood, the airway lumen, and lung tissue: these compartments reflect circulating abundance, luminal recruitment, and local tissue activity, respectively, and are related but non-equivalent. (C) summarizes the use of blood EOS for exacerbation-risk stratification, estimation of inhaled corticosteroid benefit, and trial enrichment, while indicating where additional pathway and clinical-context evidence is required. Solid arrows indicate the proposed clinical workflow; dashed bidirectional arrows indicate temporal or cross-compartment discordance.

Abbreviations: COPD, chronic obstructive pulmonary disease; EOS, eosinophils; FeNO, fractional exhaled nitric oxide; ICS, inhaled corticosteroid; OCS, oral corticosteroid.

Target-Tier Pathways Beyond Blood EOS in COPD Type 2 Inflammation

Beyond peripheral blood cell counts, type 2-related pathology within the COPD lung should be interpreted across epithelial barrier disruption, local chemotactic signals, inflammatory-cell retention in tissue, mucus abnormalities, epithelial alarmin initiation, and mixed inflammatory microenvironments shaped by neutrophils, macrophages, oxidative stress, and infection.^21,22 These processes can be considered at distinct target tiers, including the IL-4/IL-13 epithelial–immune program, the IL-33/ST2 and TSLP epithelial alarmin pathways, and the mixed inflammatory background.^23,24 This framework helps explain why comparable blood EOS levels may be associated with different local airway inflammation, mucus burden, and epithelial injury; whether these differences modify responses to specific biologics requires prospective validation.

The IL-4/IL-13 Axis Signals an Epithelial–Chemokine–Mucus Pathological Tissue Response

IL-4/IL-13 signaling represents a tissue-response axis that links airway epithelial activation, local chemokine release, humoral immune background, and mucus-related pathology. Unlike the IL-5/EOS axis, which primarily regulates EOS production, survival, and effector-cell function, this axis is more closely related to the epithelial–immune microenvironment in the airway.^25,26

IL-4 and IL-13 can mediate downstream signaling through IL-4Rα-associated receptor complexes, and IL-4Rα blockade can simultaneously affect multiple type 2 responses mediated by both IL-4 and IL-13.^25,27 At the airway epithelial level, IL-13-related signaling can induce NOS2 expression, rendering FeNO a detectable readout of epithelial type 2 activation; concurrently, IL-13 may participate in mucus transcription programs, goblet cell metaplasia, and mucin expression such as MUC5AC.^28,29 These mechanistic data derive predominantly from asthmatic airway studies, cell experiments, and animal models; direct validation in COPD remains limited. However, the association between IL-13-related gene signatures and tissue EOS infiltration in COPD lung tissue provides preliminary human evidence for this pathway.²⁶

FeNO, CCL26/eotaxin-3, PARC/CCL18, and IgE should be interpreted as pathway readouts at distinct levels of IL-4/IL-13-related responses, rather than as isolated markers.³⁰ FeNO more closely reflects epithelial type 2 activation; CCL26/eotaxin-3 reflects the chemotactic environment for EOS migration; PARC/CCL18 reflects a myeloid cell-involved inflammatory and repair-like microenvironment; and IgE reflects IL-4/IL-13-associated humoral immunity or atopic background. After tissue entry, EOS may contribute to epithelial injury, mucus secretion, and maintenance of inflammation through degranulation and local effector-molecule release.²⁰ Together, these indices reflect complementary components of the local tissue response.

This dimension also intersects with mucus pathology in COPD. The SPIROMICS cohort demonstrated that airway MUC5AC and MUC5B concentrations are associated with COPD development and progression, with MUC5AC showing a more stable association;³¹ CT-detected mucus plugs are also associated with airflow limitation, hypoxemia, exacerbations, and mortality risk.^32,33 These findings suggest that the epithelial–mucus dimension may contribute to clinical heterogeneity and disease progression in COPD. However, mucus hypersecretion in COPD cannot be solely attributed to IL-4/IL-13 or type 2 inflammation, as smoking, oxidative stress, neutrophil/macrophage responses, and infection-related stimuli may similarly contribute.^21,22 Therefore, MUC5AC, MUC5B, and CT-detected mucus plugs are more appropriately regarded as candidate readouts of epithelial–mucus pathology burden, rather than specific indicators that independently define type 2 inflammation.

The IL-4/IL-13 dimension provides a mechanistic basis for interpreting dupilumab trial results, particularly why IL-4Rα blockade may influence tissue responses beyond the IL-5/EOS axis.^25,30

The IL-33/ST2 and TSLP Axes Reflect Epithelial Injury-Driven Context Dependency

IL-33/ST2 and TSLP are upstream alarmin pathways that link epithelial injury to innate and type 2 immune amplification. Their location at the airway interface gives both pathways a biological rationale in COPD, where cigarette smoke, oxidative stress, infection, and recurrent epithelial damage can alter local signaling.^23,34 IL-33 signals through ST2 and TSLP through TSLPR, placing both upstream of several downstream cell programs. Unlike blood EOS, these pathways describe injury sensing and immune initiation rather than the abundance of one circulating effector cell.

Epithelial IL-33 and TSLP can activate ILC2s, Th2 cells, dendritic cells, and downstream IL-5/IL-13 networks.³⁴ In COPD, however, the same ligand may not operate consistently across patient states. Current smoking has been associated with lower airway IL-33 expression,³⁵ while oxidative modification can change IL-33 and receptor-associated signaling.³⁶ Infection and barrier disruption may further shift pathway activity over time. Consequently, the presence of pathway components in tissue cannot establish that pathway output is sustained or dominant when treatment is given. These observations support context-dependent activity but do not identify a uniformly targetable population.

Human evidence supports pathway involvement. Distal COPD lung tissue shows increased ST2/IL1RL1 expression and compartment-specific IL-33/ST2 patterns involving mast cells and endothelial cells.³⁷ Aberrant TSLP and TSLPR expression and localization have also been reported in COPD airway samples.³⁸

Additional disease-state evidence comes from reported associations of circulating IL-33/ST2 measures with acute COPD exacerbations and lung-function impairment.³⁹ These observations, like tissue localization, establish biological involvement but do not by themselves establish therapeutic efficacy.^24,34,40 Direct comparative evidence of pathway dominance in asthma and COPD is limited, so therapeutic success in asthma should not be extrapolated to COPD. Future COPD studies should measure target engagement and pathway activity in the relevant disease state rather than infer validity from a universal blood EOS threshold.^23,41

Mixed Airway Inflammatory Phenotypes Complicate Interpretation of Type 2 Signals in COPD

Type 2 signals in COPD are embedded within heterogeneous airway inflammatory phenotypes rather than an isolated Th2/EOS pattern. In a single large Chinese cohort of 895 patients with stable COPD, defined using sputum EOS ≥3% and neutrophils ≥61%, 58.0% had a neutrophilic phenotype, 32.6% a mixed-granulocytic phenotype, 5.4% an eosinophilic phenotype, and 4.0% a paucigranulocytic phenotype.⁴² These proportions should not be interpreted as universal prevalence estimates because phenotype frequencies depend on cohort composition, thresholds, treatment exposure, and whether sampling occurs at stability or exacerbation. Separately, a multicohort longitudinal analysis of 510 patients and 1706 sputum samples identified neutrophilic, eosinophilic, mixed-granulocytic, and paucigranulocytic states across stable and exacerbation visits.²²

Peripheral and airway measures were frequently discordant. In SPIROMICS, 42% of participants with paired measurements were classified discordantly by blood and sputum EOS.¹⁴ The longitudinal multicohort analysis also documented within-patient transitions among neutrophilic, eosinophilic, mixed-granulocytic, and paucigranulocytic states.²² In another COPD cohort, PPM-positive stable-state sputum samples had lower sputum EOS percentages than PPM-negative samples (median, 0.5% vs 1.25%).⁴³ These observations establish temporal and spatial heterogeneity but do not show that microbial status changes biologic response.

Together, these population data establish COPD inflammatory heterogeneity. They do not demonstrate that neutrophilic or mixed-granulocytic state, microbiome pattern, or PPM status modifies response to chronic biologic therapy. Sputum-EOS-guided corticosteroid treatment in stable COPD and blood-EOS-guided systemic corticosteroid treatment during acute exacerbations address corticosteroid responsiveness, not biologic efficacy.^44,45 This distinction is essential when translating steroid evidence to long-term biologic selection. These variables should therefore be evaluated as candidate stratification or treatment-interaction factors in future studies, not used as validated current clinical selection criteria.

BOREAS and NOTUS excluded current or previous asthma, yet enrolled participants with elevated baseline FeNO.^17,46 Excluding recorded asthma improves internal validity for COPD-specific inference but may limit generalizability to real-world COPD populations in whom asthma histories or traits coexist. Thus, a type 2-associated signal can persist in rigorously defined COPD after asthma exclusion. Neither FeNO, blood EOS, nor bronchodilator response alone proves unrecognized asthma; the trials also did not establish the prevalence of significant bronchodilator responsiveness or its relationship to treatment effect. Asthma history, asthma-associated traits, and type 2 pathway activity within COPD should therefore be recorded and analyzed as three separate dimensions.

Biologic Trials Reveal Differences in Target–Population Matching

Biologic trials are efficacy studies, but in heterogeneous COPD they also provide evidence about the clinical relevance of pathway–population matching. Reproducible benefit in a prespecified enriched population supports the therapeutic relevance of the target pathway in that setting; conversely, negative primary endpoints or exploratory subgroup signals may reflect limited target relevance in the enrolled population or inadequate matching of patient background, biomarker strategy, and target tier.^23,47

Clinical Trial Results Should Be Interpreted as Target-Population Matching Evidence

Recent systematic reviews, meta-analyses, and network meta-analyses describe overall efficacy and safety trends across COPD biologic studies, but cross-study comparisons remain constrained by differences in targets, enrichment criteria, background inhaled therapy, endpoints, follow-up, and subgroup definitions.^47,48 Trials should therefore be interpreted according to the pathway tier they test and the adequacy of target–population–biomarker matching, rather than exacerbation reduction alone or primary-endpoint positivity. IL-5/IL-5R targeting assesses the EOS effector-cell compartment; IL-4Rα blockade acts on IL-4/IL-13-related epithelial, chemokine, and tissue responses; and IL-33/ST2 and TSLP target upstream epithelial injury sensing and innate immune initiation.^25,34 Differences across trial results should consequently be considered in relation to patient context, pathway activity, and enrichment metrics.^23,24

IL-5/IL-5R Targeted Studies Support Partial Treatability at the EOS Effector-Cell Level

IL-5/IL-5R-targeted trials evaluate whether intervention at the EOS effector-cell level can reduce exacerbation risk in selected patients with COPD.^16,49 These programs differed in EOS enrichment thresholds, background inhaled therapy, prespecified analysis populations, multiplicity procedures, sample size, placebo event rates, and clinical endpoints. The endpoints included exacerbation frequency and, in some programs, symptoms and health status. Their findings should therefore be interpreted first in relation to trial design and population selection, before any negative or uncertain result is used to infer a biological limitation of the IL-5/EOS axis.

The mepolizumab program supports treatability of the IL-5/EOS axis in selected eosinophilic COPD populations. METREX enrolled 836 patients and prospectively defined an eosinophilic-phenotype population of 462 patients using blood EOS ≥150 cells/μL at screening or ≥300 cells/μL during the previous year. In this prespecified population receiving ICS-based triple therapy, mepolizumab reduced the annualized moderate-to-severe exacerbation rate to 1.40 compared with 1.71 with placebo, whereas no significant benefit was observed in the overall population.¹⁶ METREO enrolled 674 patients using the same composite EOS criterion and background therapy. The 100-mg group showed a numerical reduction in exacerbations, with a rate ratio of 0.80, but did not meet the prespecified multiplicity-adjusted significance criterion.^16,50 MATINEE subsequently enrolled 804 patients receiving triple inhaled therapy with blood EOS ≥300 cells/μL and significantly reduced moderate-to-severe exacerbations (rate ratio, 0.79). MATINEE demonstrated exacerbation reduction in a population prospectively enriched at blood EOS ≥300 cells/μL, although symptoms and health status did not improve consistently.⁵¹

GALATHEA and TERRANOVA evaluated benralizumab in primary-analysis populations of 1120 and 1545 patients, respectively, using blood EOS ≥220 cells/μL and guideline-based dual or triple inhaled therapy.⁴⁹ Neither trial met its primary exacerbation endpoint, and their large populations make inadequate total sample size an insufficient explanation. Placebo-group exacerbation rates were 1.24 and 1.17 per year, respectively, compared with 1.71 in the METREX eosinophilic population and 1.49 in METREO; MATINEE was positive despite a lower rate of 1.01.^16,49,51 Low placebo event rates may reduce event accrual and statistical power, but do not determine trial outcome. Their comparison illustrates why event rate is a power consideration rather than a causal explanation. A post hoc benralizumab analysis suggested a responder profile among patients with elevated blood EOS, at least three previous exacerbations, and triple therapy; this remains hypothesis-generating rather than a validated selection rule.⁵²

Across these programs, IL-5/EOS-directed treatment reduced exacerbation risk in some eosinophilic COPD populations, supporting partial treatability of the EOS effector-cell compartment.^16,49,51 The benefit was not reproduced consistently across trials or drugs, and blood EOS alone did not fully define treatment responsiveness. Negative or uncertain findings cannot be attributed only to limited EOS-axis coverage, nor explained by any single feature such as EOS threshold, sample size, background therapy, or placebo event rate. The programs also evaluated different molecules, endpoint structures, and statistical hierarchies, preventing one cross-program variable from being isolated retrospectively. Interpretation should integrate trial design, statistical structure, population enrichment, drug-specific biology, and residual disease heterogeneity; separate, non-head-to-head programs cannot determine the independent causal contribution of each factor.

IL-4/IL-13 Blockade: Replicated Phase 3 Evidence and Implications for the Epithelial–Chemokine Axis

IL-4Rα blockade has generated replicated phase 3 positive evidence in COPD type 2 inflammation research.^17,46 By blocking IL-4Rα, dupilumab inhibits signaling mediated by both IL-4 and IL-13 and acts at a tissue-response tier that includes epithelial activation, chemokine release, humoral immune background, and the local tissue microenvironment.^25,30

BOREAS enrolled COPD patients with blood EOS ≥300 cells/μL who remained at risk for exacerbations despite standard inhaled therapy and were excluded for a history of asthma, demonstrating that dupilumab reduced the moderate-to-severe exacerbation rate and improved forced expiratory volume in 1 second (FEV1), St George’s Respiratory Questionnaire (SGRQ), and Evaluating Respiratory Symptoms in COPD (E-RS-COPD) scores.¹⁷ NOTUS further confirmed exacerbation reduction and FEV1 improvement in a similarly blood EOS-enriched population, but the change in SGRQ score from baseline to 52 weeks did not show a significant between-group difference.⁴⁶ A pooled analysis of both studies supported the overall consistency of exacerbation reduction and lung function improvement.⁵³ Therefore, the strength of dupilumab evidence primarily lies in the reproducibility of exacerbation and lung function benefits; patient-reported outcomes, while generally supporting the direction of symptom and quality of life improvement, showed less consistency across individual phase 3 studies than exacerbations and FEV1.

BOREAS and NOTUS should not be interpreted simply as showing that patients with elevated blood EOS respond to biologics. In blood EOS-enriched COPD populations with persistent exacerbation risk despite standard inhaled therapy, IL-4Rα blockade was associated with reproducible exacerbation reduction and lung function improvement.^17,46,53 Compared with the mixed IL-5/IL-5R evidence, these findings are consistent with treatment-relevant biology extending beyond the EOS effector-cell compartment to an IL-4/IL-13-related epithelial–chemokine–tissue response program.^25,30

The BOREAS biomarker analysis further supports this pathway positioning. Dupilumab reduced FeNO, total IgE, eotaxin-3, and PARC, indicating effects across multiple type 2-related pharmacodynamic readouts.³⁰ Higher baseline FeNO was also associated with a greater reduction in moderate-to-severe exacerbations (nominal treatment-by-biomarker interaction P=0.043).³⁰ This post hoc signal supports FeNO as a candidate response-related biomarker alongside blood EOS but does not establish a validated selection threshold; the continuous treatment-by-FeNO interaction has not been independently replicated in NOTUS. These findings help characterize an IL-4/IL-13-related epithelial–chemokine–tissue response program without establishing mucus pathology, CT-detected mucus plugs, airway remodeling, or epithelial transcriptional alterations as causal mediators of clinical benefit.³⁰

Thus, the significance of IL-4Rα blockade lies in the reproducibility of clinical benefit in a defined blood EOS-enriched population and in the accompanying changes in type 2-related pharmacodynamic readouts.^17,46,53 More refined responder characteristics and the causal contribution of epithelial or mucus pathology require prospective mechanism-embedded studies.³⁰

Evolving Evidence Across IL-33/ST2 and TSLP Alarmin Pathways

Evidence for IL-33/ST2 in COPD now spans human genetics, lung-tissue localization, pharmacodynamic target engagement, and randomized clinical trials.^37,54 This evidence supports therapeutic relevance, but inconsistency across molecules, trials, and dosing regimens does not establish a uniform class effect. Smoking, oxidative stress, infection, and abnormal epithelial repair may alter pathway activity,^35,36 limiting direct extrapolation from asthma and requiring COPD trial findings to be judged on their own evidentiary strength. TSLP evidence is considered separately below.

Itepekimab targets IL-33. In its phase 2a COPD study, the accompanying genetic analysis strengthened the pathway rationale, while the trial did not meet the exacerbation primary endpoint in the overall population (rate ratio, 0.81; P=0.13); the prespecified former-smoker subgroup showed nominal exacerbation and lung-function signals.⁵⁴ Thus, disease relevance did not translate into uniform efficacy. AERIFY-1 and AERIFY-2 were described in a peer-reviewed phase 3 trial-design publication in former smokers with moderate-to-severe COPD.⁵⁵ Full peer-reviewed phase 3 efficacy reports were unavailable as of July 10, 2026.

Astegolimab blocks ST2. COPD-ST2OP did not significantly reduce exacerbations in its phase 2a intention-to-treat population.⁵⁶ Later, ALIENTO phase 2b and ARNASA phase 3 enrolled current or former smokers with frequent exacerbations irrespective of blood EOS or chronic bronchitis.⁵⁷ Although individual trial-dose findings were not fully concordant, the prespecified pooled analysis of 2682 participants demonstrated statistically significant reductions in moderate-to-severe exacerbations with dosing every 2 weeks (rate ratio, 0.85) and every 4 weeks (rate ratio, 0.88).⁵⁸ These findings provide direct clinical evidence that ST2 blockade can reduce exacerbation risk in COPD, while incomplete individual-trial replication argues against a uniform response. Because oxidized IL-33 can signal through an ST2-independent RAGE/EGFR complex, ST2 blockade may not cover all IL-33 biology, a biologically plausible but clinically unproven distinction.³⁶

Tozorakimab is an IL-33-directed antibody. In a Phase 1 study, it engaged endogenous IL-33 in serum and nasal airway samples and reduced circulating IL-5 and IL-13 in participants with COPD, providing proof of mechanism.⁵⁹ Target engagement and downstream biomarker modulation did not establish clinical efficacy: FRONTIER-4 did not meet its phase 2a primary endpoint for change in pre-bronchodilator FEV1, although exploratory post-bronchodilator FEV1, COPDCompEx, and frequent-exacerbator signals were observed.⁶⁰ Preclinical studies indicate that tozorakimab can inhibit reduced IL-33/ST2 signaling and limit subsequent oxidized IL-33/RAGE–EGFR signaling.⁶¹ ClinicalTrials.gov records accessed July 10, 2026 identify OBERON and TITANIA as completed and MIRANDA as active, not recruiting.^62–64 These records support trial identity, design, and status only; full peer-reviewed phase 3 efficacy reports were unavailable.

Evidence for TSLP targeting is less mature. In COURSE, tezepelumab did not significantly reduce moderate-to-severe exacerbations overall (rate ratio, 0.83; 90% CI, 0.64–1.06), although prespecified higher-EOS subgroups showed numerical trends.⁶⁵ Taken together, IL-33/ST2 has progressed beyond biological plausibility to direct clinical evidence of therapeutic relevance in COPD: human genetic and lung-tissue studies support disease involvement, pharmacodynamic data demonstrate target engagement, and astegolimab trials show that ST2 blockade can reduce exacerbation risk.^{37,54,57–59} However, inconsistent benefit across molecules, individual trials, and dosing regimens argues against a uniform class effect or universal pathway dominance. Three non-exclusive explanations remain: limited pathway dominance in COPD; mismatch between pathway-active states and trial entry criteria; and different pharmacologic coverage of IL-33 ligand and ST2 receptor blockade. Negative or inconsistent findings cannot be attributed mainly to inadequate enrichment, but neither do they negate the target. IL-33/ST2 is therefore therapeutically relevant but intervention- and context-dependent; TSLP remains exploratory.

Trial Differences Suggest Target-Directed Enrichment

Key design features and primary findings of COPD biologic trials are summarized in Table 1. Table 2 presents a proposed target-tier framework linking candidate pathway readouts with patient-context variables for future enrichment, stratification, pharmacodynamic assessment, and treatment-interaction analyses. Together, the tables separate trial-level evidence from the proposed enrichment framework and are not intended to provide a simple cross-drug efficacy ranking.

Table 1.

Key Design Features, Evidence Sources, and Evidence Status of Biologic Trials and Programs in COPD

Target/Drug	Trial/Phase (n)	Key Population and Background	Primary Finding	Evidence Status
IL-5/Mepolizumab	METREX, phase 3 (n=836; eosinophilic n=462); METREO, phase 3 (n=674); MATINEE, phase 3 (n=804)16,51	METREX/METREO: EOS ≥150 at screening or ≥300 cells/μL in the previous year; triple therapy.16 MATINEE: EOS ≥300 cells/μL; triple therapy51	METREX was positive in the prespecified eosinophilic population; METREO did not meet the multiplicity-adjusted criterion; MATINEE confirmed a modest exacerbation benefit16,51	Peer-reviewed phase 3; enriched-population benefit reproduced
IL-5Rα/Benralizumab	GALATHEA, phase 3 (primary n=1120); TERRANOVA, phase 3 (primary n=1545)49,52	EOS ≥220 cells/μL; frequent exacerbations; dual or triple therapy49	Both primary endpoints were negative;49 a pooled post hoc analysis identified a possible responder profile52	Negative phase 3; post hoc responder signal
IL-4Rα/Dupilumab	BOREAS, phase 3 (n=939); NOTUS, phase 3 (n=935)17,46	EOS ≥300 cells/μL; chronic bronchitis symptoms; exacerbations despite optimized inhaled therapy; asthma history excluded17,46	Both trials reduced exacerbations and improved prebronchodilator FEV117,46	Replicated peer-reviewed phase 3 evidence
IL-33/Itepekimab	Phase 2a (n=343);54 AERIFY-1 and AERIFY-2, phase 3 trial-design program55	Phase 2a: moderate-to-severe COPD.54 AERIFY design: former smokers with moderate-to-severe COPD receiving dual or triple inhaled therapy55	Phase 2a was negative overall with an exploratory former-smoker signal;54 phase 3 efficacy was not assessed because a full peer-reviewed efficacy report was unavailable55	Peer-reviewed phase 2a efficacy report; peer-reviewed phase 3 design publication only
ST2/Astegolimab	COPD-ST2OP, phase 2a (n=81);56 ALIENTO, phase 2b (randomized ITT n=1302; initiated treatment n=1301); ARNASA, phase 3 (randomized ITT n=1380; initiated treatment n=1375); prespecified pooled ITT analysis (n=2682)57,58	Frequent exacerbations; current or former smokers; optimized inhaled therapy; no EOS requirement56,57	COPD-ST2OP was negative;56 ALIENTO/ARNASA dose-regimen results were incompletely concordant;57 the prespecified pooled analysis showed modest benefit58	Peer-reviewed phase 2b/3; prespecified pooled benefit
IL-33/Tozorakimab	FRONTIER-4, phase 2a (n=135);60 OBERON, TITANIA, and MIRANDA, registered phase 3 programs62–64	FRONTIER-4: COPD with chronic bronchitis receiving dual or triple inhaled therapy.60 Phase 3 programs: symptomatic COPD with an exacerbation history on optimized maintenance inhaled therapy62–64	FRONTIER-4 did not meet its primary endpoint;60 no phase 3 efficacy inference was made from registry status62–64	Peer-reviewed phase 2a efficacy report; registry-supported phase 3 status only
TSLP/Tezepelumab	COURSE, phase 2a (n=333)65	≥2 moderate or severe exacerbations; triple therapy; no EOS requirement65	The primary endpoint was not met; higher-EOS subgroups showed numerical trends65	Peer-reviewed phase 2a exploratory evidence

Notes: Evidence status distinguishes complete peer-reviewed efficacy reports from peer-reviewed trial-design publications and ClinicalTrials.gov registry records. ClinicalTrials.gov records are used only to support trial names, design features, and registry status as of the access date and are not used to infer clinical efficacy. Sample sizes and efficacy estimates are reported only when supported by complete peer-reviewed efficacy publications.

Abbreviations: COPD, chronic obstructive pulmonary disease; EOS, eosinophils; FEV₁, forced expiratory volume in 1 second; IL, interleukin; IL-4Rα, interleukin-4 receptor alpha; IL-5Rα, interleukin-5 receptor alpha; ST2, suppression of tumorigenicity 2; TSLP, thymic stromal lymphopoietin.

Table 2.

Proposed Target-Tier Framework for Biomarker Readouts and Candidate Variables in COPD Biologic Studies

Target Tier	Candidate Readouts	Candidate Patient/Context Variables	Evidence Boundary
IL-5/IL-5R effector-cell axis	Repeated blood EOS; sputum EOS; pharmacodynamic EOS depletion16,19,49,51,52	Exacerbation burden; corticosteroid exposure; background inhaled therapy	Single blood EOS values incompletely represent airway inflammation; responder profiles remain post hoc14–16,19,49,51,52
IL-4/IL-13 tissue-response axis	FeNO; IgE; eotaxin-3/CCL26; PARC/CCL18; epithelial type 2 signatures17,30,46,53,66,67	Repeated EOS/FeNO; chronic bronchitis symptoms or mucus burden	BOREAS FeNO interaction was post hoc; no selection threshold has been established17,30,46,53,66,67
IL-33/ST2 alarmin axis	Research IL-33/ST2 pathway assays; epithelial-injury and mucus-related readouts35,36,54,56,60	Smoking; chronic bronchitis; infection-triggered exacerbations; epithelial injury	Pharmacologic coverage and evidence maturity differ across agents35,36,54,56,60
TSLP alarmin axis	Research TSLP-related assays; EOS; FeNO38,65	Exacerbation burden; smoking and infection context	Evidence remains phase 2a and exploratory38,65
Cross-target contextual modifiers	Sputum inflammatory phenotype; microbiome/PPM status19,22,43,68,69	Asthma history and associated traits; corticosteroid exposure; infection and sampling state	Direct modification of biologic treatment effects remains untested19,22,43,68,69

Notes: Cross-target contextual modifiers are not therapeutic targets. The listed readouts and context variables are proposed for prospective enrichment, stratification, pharmacodynamic assessment, or treatment-interaction analyses and should not be interpreted as validated clinical prescribing criteria.

Abbreviations: CCL, C-C motif chemokine ligand; COPD, chronic obstructive pulmonary disease; EOS, eosinophils; FeNO, fractional exhaled nitric oxide; IgE, immunoglobulin E; IL, interleukin; IL-5R, interleukin-5 receptor; PARC, pulmonary and activation-regulated chemokine; PPM, potentially pathogenic microorganism; ST2, suppression of tumorigenicity 2; TSLP, thymic stromal lymphopoietin.

Cross-Target Comparisons Support Pathway-Tier Heterogeneity

Cross-target comparison should focus on the pathway tier each biologic perturbs and the strength of evidence within the population studied.^41,47 IL-5/IL-5R targeting examines the EOS effector-cell compartment. Replicated phase 3 results with IL-4Rα blockade support an IL-4/IL-13-related epithelial–chemokine–tissue response program in blood EOS-enriched patients at continued exacerbation risk.^16,17,46,49 IL-33/ST2 has direct but program-specific clinical evidence of therapeutic relevance, whereas TSLP remains exploratory; interpretation must consider pathway dominance, patient context, pharmacologic coverage, and evidence maturity.^{54,57–60,65} Reproducible phase 3 benefit supports therapeutic relevance in prespecified populations, whereas program-specific or mixed findings define the limits of generalization and motivate target-specific enrichment hypotheses.^{16,23,41,54,58}

These comparisons, summarized in Figure 2, support a practical framework based on target tier, tissue pathway, and patient context. The next section translates this framework into target-directed composite biomarker strategies for patient selection and trial design.^10,24,41,47

Figure 2.

Target-specific pathway heterogeneity of type 2 inflammation in COPD. Pathway triggers and patient-context modifiers are shown above the principal type 2-related tiers. (A) depicts the upstream IL-33/ST2 and TSLP alarmin pathways; (B), the IL-4/IL-13 tissue-response branch through IL-4Rα; (C), the IL-5/EOS effector branch; and (D), representative clinical outcomes and pathway readouts. Solid arrows indicate conceptual relationships and convergence, not an obligatory linear cascade. The right-hand column summarizes the maturity and consistency of clinical evidence across target programs. Colors indicate evidence status rather than comparative efficacy: green, replicated benefit; blue, exploratory evidence; Orange, program-specific or mixed evidence; and red, negative primary endpoints with post hoc evidence only.

Abbreviations: COPD, chronic obstructive pulmonary disease; EOS, eosinophils; FeNO, fractional exhaled nitric oxide; IL, interleukin; IL-4Rα, interleukin-4 receptor alpha; IL-5R, interleukin-5 receptor; IL-5Rα, interleukin-5 receptor alpha; ILC2, group 2 innate lymphoid cell; ST2, suppression of tumorigenicity 2; Th2, T helper type 2; TSLP, thymic stromal lymphopoietin.

Target-Directed Composite Biomarker Enrichment Strategy

Cross-target trial evidence supports prospective testing of target-specific patient selection and mechanism-embedded validation. The aim is to determine whether target perturbation, pathway readouts, and clinical responses are coherent, rather than further refine a universal blood EOS threshold. All variables proposed below are investigational candidates for prospective study, not current biologic-selection or prescribing criteria.^{7,9,10,30,41,47}

Composite Biomarkers Should Be Prespecified by Target Hypothesis, Not Post Hoc Explained

Composite biomarker design should assign each variable a prespecified role. These roles include enrollment to enrich for a candidate pathway, stratification to balance relevant background factors, treatment-by-biomarker interaction testing, pharmacodynamic verification of pathway perturbation, and longitudinal sampling. This structure helps distinguish limited target relevance from inadequate enrichment or pathway-readout selection without relying on post hoc subgroup explanations.^7,41,52

For IL-5/IL-5R trials, repeated or longitudinally stable blood EOS and exacerbation history despite optimized inhaled therapy may serve as candidate enrollment or stratification variables, whereas sampling timing, recent systemic corticosteroid exposure, and acute infection should be documented and controlled as potential confounders.^9,16,51 For IL-4Rα trials, blood EOS plus FeNO, CCL26/eotaxin-3, PARC/CCL18, total IgE, epithelial type 2 signatures, and type 2 pharmacodynamic readouts may serve as candidate variables; their complementary contribution should be tested prospectively through treatment-by-biomarker interaction analyses.^12,30 For IL-33/ST2 or TSLP trials, smoking status, chronic bronchitis phenotype, infection- or virus-triggered exacerbations, mucus burden or CT-detected mucus plugs, epithelial injury markers, and the mixed inflammatory background may serve as candidate stratification or interaction variables. CT-detected mucus plugs have prognostic relevance in COPD, while their predictive value for biologic response remains untested.^32,35,60,65 These proposals align with contemporary review evidence emphasizing careful patient selection and broader phenotyping.⁷⁰

Longitudinal Inflammatory Assessment Helps Reduce Misclassification Between Blood EOS and Pathway Status

Longitudinal assessment should begin with stable-state blood EOS and document sampling timing, background inhaled therapy, recent inhaled or systemic corticosteroid exposure, acute infection, and exacerbation status. Repeated blood EOS and FeNO measurements during screening, before randomization, early treatment, and around exacerbations may help distinguish persistent treatment-relevant signals from event-related fluctuations and determine whether pharmacodynamic changes are directionally consistent with clinical outcomes. For upstream alarmin targets, exacerbation-related samples should be interpreted alongside smoking status, chronic bronchitis, infection context, epithelial injury, and mixed inflammatory features.^{9,19,22,30,34,35,39}

Mechanism-Embedded Trials Should Link Target Perturbation, Pathway Readouts, and Clinical Responses

Mechanism-embedded trials should prespecify a limited biomarker panel, the role of each readout, clinical endpoints, pharmacodynamic endpoints, and treatment-by-biomarker interaction analyses. They should test whether the target has been sufficiently perturbed, whether prespecified pathway readouts change accordingly, and whether clinical outcomes are directionally consistent with those changes.^7,41

For IL-4/IL-13 or IL-4Rα blockade, FeNO, CCL26/eotaxin-3, PARC/CCL18, total IgE, epithelial type 2 signatures, and airway eosinophilia-related transcriptional signals may be used at different interpretive levels. MUC5AC/MUC5B and imaging-based mucus measures may serve as exploratory structural or mucus-pathology endpoints, rather than as proof of causal mediation of clinical benefit.^{26,30,33,66,67}

For IL-33/ST2 or TSLP trials, the panel should place greater emphasis on epithelial injury or alarmin-related markers, exacerbation etiology, mixed inflammatory indicators, and repeated sampling across stable and exacerbation states. Prespecified interaction analyses should test whether these contextual variables modify treatment response, rather than use them solely to explain subgroup signals after a negative primary endpoint.^34,35,60,65

Future COPD biologic trials should document physician-diagnosed asthma, age at diagnosis or symptom onset, previous asthma treatment, symptom variability, atopy, and relevant upper-airway comorbidities separately from current type 2 biomarkers. Bronchodilator response should be reported continuously and using a prespecified categorical definition; blood EOS and FeNO should be repeated during stable disease.⁷¹ These variables should be recorded before randomization and analyzed independently. Protocols should prespecify exclusion, stratified and sensitivity analyses, and treatment-by-trait interaction tests. No single FeNO, blood EOS, or bronchodilator-response threshold is currently validated to distinguish unrecognized asthma from COPD-intrinsic type 2 inflammation.

Conclusion

Blood EOS remains a practical entry biomarker, but cross-trial evidence is best interpreted at the pathway tier actually perturbed. IL-5/EOS-directed therapy supports partial treatability of the eosinophil effector-cell compartment rather than broadly consistent efficacy, whereas IL-33/ST2 has progressed from biological support to direct, though intervention- and context-dependent, evidence of therapeutic relevance. Replicated IL-4Rα results show that reproducible benefit can emerge when the target, enriched population, and relevant tissue-response program are aligned.

Future enrichment should combine evidence of target engagement and pathway activity with stable or repeated biomarkers and prespecified patient-context variables. Protocols should assign these measures defined roles in enrollment, stratification, pharmacodynamic assessment, and treatment-interaction analysis, then test whether target perturbation changes pathway readouts and clinical outcomes coherently. This framework is intended for prospective validation and trial design, not as a current prescribing algorithm, and it should not convert a single biomarker into a universal definition of COPD type 2 inflammation.

Funding Statement

This work was supported by the National Key Research and Development Program of China (No. 2023YFC3502601) and the National Science and Technology Major Project of China (No. 2023ZD0506705). This work was also supported by the “Three 100s” Initiative Program of Henan Provincial Academy of Medical Sciences (HNCRD202435) and the Young Qihuang Scholars Award of The State TCM Academic Young Top-notch Talents Program (Document No. 9 [2026]). All of the above grants were awarded to Yang Xie. The funders had no role in the conception of the review, literature interpretation,article preparation, or the decision to submit the article for publication.

Declaration of Artificial Intelligence (AI) Use

The authors used OpenAI ChatGPT (GPT-5.5) for language polishing and structural editing assistance. The tool was not used to generate scientific conclusions, analyze data, or select references. All scientific content, citations, and final wording were reviewed and approved by the authors, who take full responsibility for the article.

Abbreviations

ACO, asthma-COPD overlap; ATS, American Thoracic Society; COPD, chronic obstructive pulmonary disease; E-RS-COPD, Evaluating Respiratory Symptoms in COPD; EOS, eosinophils; ERS, European Respiratory Society; FeNO, fractional exhaled nitric oxide; FEV1, forced expiratory volume in 1 second; GOLD, Global Initiative for Chronic Obstructive Lung Disease; ICS, inhaled corticosteroid; IgE, immunoglobulin E; IL, interleukin; ILC2, group 2 innate lymphoid cell; MUC5AC, mucin 5AC; MUC5B, mucin 5B; OCS, oral corticosteroid; PARC, pulmonary and activation-regulated chemokine; PPM, potentially pathogenic microorganism; RCT, randomized controlled trial; SCS, systemic corticosteroid; SGRQ, St George’s Respiratory Questionnaire; ST2, suppression of tumorigenicity 2; Th2, T helper type 2; TSLP, thymic stromal lymphopoietin.

Data Sharing Statement

Data sharing is not applicable to this article, as no new data were created or analyzed in this review.

Ethics Approval and Informed Consent

Ethics approval and informed consent were not required because this article is a review based on previously published literature and did not involve human participants, animals, identifiable personal data, or biological samples.

Author Contributions

Jinzhong Zhang: Conceptualization, Methodology, Investigation, Visualization, Writing – original draft, Writing – review & editing.

Peng Zhang: Investigation, Writing – review & editing.

Yiyin Zhang: Investigation, Writing – review & editing.

Yang Xie: Conceptualization, Supervision, Funding acquisition, Writing – review & editing. All authors made a significant contribution to the work reported; took part in drafting, revising, or critically reviewing the article; gave final approval of the version to be published; agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors report no conflicts of interest in this work.