Asthma-COPD overlap. Clinical relevance of genomic signatures of type 2 inflammation in chronic obstructive pulmonary disease

Abstract

Rationale: Chronic obstructive pulmonary disease (COPD) is a heterogeneous disease and likely includes a subgroup that is biologically comparable to asthma. Studying asthma-associated gene expression changes in COPD could add insight into COPD pathogenesis and reveal biomarkers that predict a favorable response to corticosteroids.

Objectives: To determine whether asthma-associated gene signatures are increased in COPD and associated with asthma-related features.

Methods: We compared disease-associated airway epithelial gene expression alterations in an asthma cohort (n = 105) and two COPD cohorts (n = 237, 171). The T helper type 2 (Th2) signature (T2S) score, a gene expression metric induced in Th2-high asthma, was evaluated in these COPD cohorts. The T2S score was correlated with asthma-related features and response to corticosteroids in COPD in a randomized, placebo-controlled trial, the Groningen and Leiden Universities study of Corticosteroids in Obstructive Lung Disease (GLUCOLD; n = 89).

Measurements and main results: The 200 genes most differentially expressed in asthma versus healthy control subjects were enriched among genes associated with more severe airflow obstruction in these COPD cohorts (P < 0.001), suggesting significant gene expression overlap. A higher T2S score was associated with decreased lung function (P < 0.001), but not asthma history, in both COPD cohorts. Higher T2S scores correlated with increased airway wall eosinophil counts (P = 0.003), blood eosinophil percentage (P = 0.03), bronchodilator reversibility (P = 0.01), and improvement in hyperinflation after corticosteroid treatment (P = 0.019) in GLUCOLD.

Conclusions: These data identify airway gene expression alterations that can co-occur in asthma and COPD. The association of the T2S score with increased severity and "asthma-like" features (including a favorable corticosteroid response) in COPD suggests that Th2 inflammation is important in a COPD subset that cannot be identified by clinical history of asthma.

Keywords: biological markers; chronic obstructive pulmonary disease; gene expression profiling.

Figure 1.

Study design. Gene expression overlap in asthma and chronic obstructive pulmonary disease (COPD) was identified by comparing microarray profiles from an asthma cohort with profiles from three COPD cohorts: (1) the bronchial airway epithelium COPD (BAEC) cohort; (2) the small airway epithelium COPD (SAEC) cohort; and (3) the Groningen and Leiden Universities study of Corticosteroids in Obstructive Lung Disease (GLUCOLD) cohort. GSEA = gene set enrichment analysis; ICS = inhaled corticosteroid; LABA = long-acting β agonist; T2S = Th2 signature; TGM = three gene mean.

Figure 2.

Genes most altered in the airway epithelium in asthma evaluated in chronic obstructive pulmonary disease (COPD) using gene set enrichment analysis (GSEA). Line plots of the 100 genes most up-regulated in asthma compared with control (fold change > 1.25; false discovery rate q-value < 0.01) show enrichment among (A) genes associated with low FEV₁ % predicted (FEV₁%pred) in the bronchial airway epithelium COPD (BAEC) cohort (P < 0.001), and (B) genes up-regulated in subjects with COPD compared with smokers without COPD in the small airway epithelium COPD (SAEC) cohort (P < 0.001). The 100 genes most down-regulated in asthma were similarly associated with (A) high FEV₁%pred in the BAEC cohort (P < 0.001) and (B) COPD in the SAEC cohort (P < 0.001). Red and blue bars represent all genes ranked by (A) t statistic for correlation of gene expression with FEV₁%pred in the BAEC cohort or (B) fold change for association with COPD in the SAEC cohort. Vertical black lines represent the position of genes in the gene set along the ranked gene list. The length of the lines represents the magnitude of the GSEA running enrichment score.

Figure 3.

Relationship between T helper (Th) type 2 signature (T2S) score and measurements of lung function decline. The 100-gene T2S score (A) increases with decreasing FEV₁ % predicted (FEV₁%pred) in the bronchial airway epithelium chronic obstructive pulmonary disease (COPD) cohort (fold change = −0.7 for every 10% increase in FEV₁%pred; P < 0.001), and (B) is higher in subjects with COPD (3.51 ± 2.89 [mean ± SD]) than in smokers without airway obstruction (0.16 ± 2.39) and in nonsmokers without airway obstruction (−2.19 ± 4.70) in the small airway epithelium COPD cohort (P < 0.001 for all comparisons).

Figure 4.

Heatmap of the 100-gene Th2-high gene signature supervised by T helper (Th) type 2 signature (T2S) score. Shown for the (A) bronchial airway epithelium chronic obstructive pulmonary disease (COPD) dataset and (B) small airway epithelium COPD dataset. Genes are in rows and samples are in columns arranged from low to high T2S score. Red indicates high relative expression; blue indicates low relative expression; yellow and purple bar above each heatmap indicates COPD status. The green and red bar in A indicates asthma status. Unsupervised clustering gives a similar result, with approximately 20% as Th2 high; however, the genes first cluster by smoking status, then by the T2S metric.

Figure 5.

T helper (Th) type 2 signature (T2S) score in the Groningen and Leiden Universities study of Corticosteroids in Obstructive Lung Disease (GLUCOLD) dataset. (A) The T2S score derived from endobronchial biopsy data was associated with increasing eosinophil counts in subjects with chronic obstructive pulmonary disease in the GLUCOLD cohort at baseline (steroid naive; P = 0.003). In the longitudinal analyses, increased baseline T2S score was associated with a greater decrease in residual volume (RV)/total lung capacity (TLC) % predicted in (C) the treatment group (inhaled corticosteroid [ICS] ± long-acting β-agonist [LABA]) compared with (B) placebo at 30 months (P = 0.019 for the interaction between treatment and T2S score).