A sputum 6-gene signature predicts future exacerbations of poorly controlled asthma

Background: Improved diagnostic tools for predicting future exacerbation frequency in asthmatic patients are required. A sputum gene expression signature of 6 biomarkers (6-gene signature [6GS], including Charcot-Leyden crystal galectin [CLC]; carboxypeptidase 3 [CPA3]; deoxyribonuclease 1-like 3 [DNASE1L3]; alkaline phosphatase, liver/bone/kidney [ALPL]; CXCR2; and IL1B) predicts inflammatory and treatment response phenotypes in patients with stable asthma. Recently, we demonstrated that azithromycin (AZM) add-on treatment in patients with uncontrolled moderate-to-severe asthma significantly reduced asthma exacerbations (AMAZES clinical trial). Objectives: We sought to test whether the 6GS predicts future exacerbation and inflammatory phenotypes in a subpopulation of AMAZES and to test the effect of AZM therapy on 6GS expression and prognostic capacity. Methods: One hundred forty-two patients (73 placebo-treated and 69 AZM-treated patients) had sputum stored for quantitative PCR of 6GS markers at baseline and after 48 weeks of treatment. Logistic regression and receiver operating characteristic and area under the curve (AUC) determination were performed on baseline measures, and in an exploratory analysis the predictive value of the 6GS was compared with conventional biomarkers for exacerbation and inflammatory phenotypes. Results: The 6GS significantly predicted all future exacerbation phenotypes tested. Calculated AUCs for the 6GS were significantly greater than AUCs for peripheral blood eosinophil counts, sputum neutrophil counts, and combined sputum eosinophil and neutrophil counts. 6GS AUCs were also numerically but not significantly greater than those for fractional exhaled nitric oxide values and sputum eosinophil counts. AZM treatment altered neither 6GS expression nor the predictive capacity of the 6GS for future exacerbation phenotypes. The 6GS was a significant predictor of airway inflammatory phenotype in this population. Conclusion: We demonstrate that a sputum gene signature can predict future exacerbation phenotypes of asthma, with the greatest biomarker performance in identifying those who would experience frequent severe exacerbations. AZM therapy did not modify 6GS expression or biomarker performance, suggesting the therapeutic action of AZM is independent of 6GS-related inflammatory pathways.