Transcriptome analysis of sputum cells reveals two distinct molecular phenotypes of "asthma and chronic obstructive pulmonary disease overlap" in the elderly

Abstract

Background: Little is known about the pathogenesis of asthma and chronic obstructive pulmonary disease (COPD) overlap (ACO). This study examined the molecular phenotypes of ACO in the elderly.

Methods: A genome-wide investigation of gene expression in sputum cells from the elderly with asthma, ACO, or COPD was performed using gene set variation analysis (GSVA) with predefined asthma- or COPD-specific gene signatures. We then performed a subsequent cluster analysis using enrichment scores (ESs) to identify molecular clusters in the elderly with ACO. Finally, a second GSVA was conducted with curated gene signatures to gain insight into the pathogenesis of ACO associated with the identified molecular clusters.

Results: Seventy elderly individuals were enrolled (17 with asthma, 41 with ACO, and 12 with COPD). Two distinct molecular clusters of ACO were identified. Clinically, ACO cluster 1 (N = 23) was characterized by male and smoker dominance, more obstructive lung function, and higher proportions of both neutrophil and eosinophil in induced sputum compared to ACO cluster 2 (N = 18). ACO cluster 1 had molecular features similar to both asthma and COPD, with mitochondria and peroxisome dysfunction as important mechanisms in the pathogenesis of these diseases. The molecular features of ACO cluster 2 differed from those of asthma and COPD, with enhanced innate immune reactions to microorganisms identified as being important in the pathogenesis of this form of ACO.

Conclusion: Recognition of the unique biological pathways associated with the two distinct molecular phenotypes of ACO will deepen our understanding of ACO in the elderly.

Keywords: Asthma; Asthma-chronic obstructive pulmonary disease overlap syndrome; Chronic obstructive pulmonary disease; Cluster analysis; Sputum; Transcriptome.

Fig. 1

A graphical summary of analysis. ACO asthma-COPD overlap, COPD chronic obstructive pulmonary disease, ES enrichment score

Fig. 2

Results of gene set variation analysis using asthma- and COPD- specific gene signatures. The GSVA enrichment scores were calculated across 70 whole-sputum gene expression profiles. ACO clusters 1 and 2 were identified from k-means clustering using the enrichments scores obtained from the GSVA with asthma- and COPD-specific gene signatures in the elderly with ACO. Dots represent the individual enrichment scores, and box and whisker plots show the median and interquartile range. Gene lists of asthma- and COPD-specific gene signatures are provided in the online supplement. GSVA gene set variation analysis, ACO asthma-COPD overlap, COPD chronic obstructive pulmonary disease, NS not significant

Fig. 3

Results of gene set variation analysis using gene signatures from the Reactome database. Dots represent the individual enrichment scores, and box and whisker plots show the median and interquartile range. Gene signatures that were significantly different in their enrichment between ACO cluster 1 and cluster 2 are shown