Lung IL-13 gene signatures are associated with raised tissue eosinophils in COPD

Abstract Background The role of eosinophils in COPD and their utility as biomarkers for cytokine targeting monoclonal therapies remains unclear. We investigated the distribution of eosinophils across different tissue compartments in COPD and analysed gene expression to understand the possible mechanistic drivers of eosinophilic inflammation in COPD. Methods Blood and BAL from ex-smoking volunteers with mild/moderate COPD (n = 31) and healthy ex-smoking controls (n = 20), and bronchial biopsy tissue in a subcohort (n = 19 and n = 8, respectively) was analysed. Differentially-expressed genes (DEGs) were characterised using RNASeq. Proteomic analysis of BAL was conducted using mass-spectrometry. Results COPD subjects had more eosinophils in blood and lung tissue compared to controls, with increased eosinophil protein CLC/Galectin-10 in BAL. However, peripheral blood eosinophil counts related poorly to numbers in lung tissue (rho = -0.09192, p = 0.3541) or proportions in BAL (rho = 0.01762, p = 0.4632). Tissue IL-5Rα expression was higher in frequent exacerbators and related to tissue eosinophils, but not peripheral blood eosinophils. Higher blood eosinophils were associated with DEGs that differed with compartment. Higher tissue eosinophil levels were associated with IL-13-induced DEGs including POSTN in bronchial brushes and CCL26 in bronchial biopsies. Gene-set enrichment analysis on data from brushings revealed significant enrichment of IL-4/IL-13, but not IL-5, pathways associated with eosinophil presence. Conclusion Eosinophilic lung inflammation is related to exacerbation frequency, but lung eosinophils are not predicted by blood eosinophil counts in COPD. Our data suggest IL-13-mediated pathways may be responsible for the presence of tissue eosinophils in COPD. Further work to establish more predictive biomarkers of lung eosinophil biology are required to unlock this axis to optimised treatment.