Predicting future exacerbations in uncontrolled asthma: A discovery/validation approach using large-scale clinical proteomics.

Abstract

Exacerbations contribute significantly to the burden of asthma. Some individuals are predisposed to recurring exacerbations; however, the underlying mechanisms are not well understood. Our aim was to generate a sputum protein signature associated with future exacerbations. A total of 22 baseline sputum samples from the control (placebo control) arm of the AMAZES study were analyzed by using an optimized high-throughput mass spectrometry method. With use of a log fold change of at least 1.5 and a P value of .05 as cutoffs, univariate analysis identified 533 differentially abundant sputum proteins in participants with and without future exacerbations over the ensuing 48 weeks. A multivariate signature of 260 proteins for predicting future exacerbations was developed by using sparse partial least squares data analysis, which was partially able to predict (with an area under the receiver operating characteristic curve of 0.95 and an error rate of 0.41) those who would likely experience an exacerbation. Next, from sputum samples collected from an additional 123 participants, the 20 most influential proteins were selected for validation and quantification. After validation, 9 proteins were found to be linked to future exacerbation risk. The final model was able to predict future exacerbations with an area under the receiver operating characteristic curve of 0.77 and an error rate of 0.40. Pathway analysis revealed major themes associated with exacerbations, including inflammation, recruitment, and proliferation of immune cells. This study has identified, for what we believe is the first time, a sputum proteomic signature and pathways associated with future exacerbations, which will facilitate the discovery of new biomarkers and novel therapeutic targets in uncontrolled persistent asthma.