Identifying Molecular Mechanisms of the Late-Phase Asthmatic Response by Integrating Cellular, Gene, and Metabolite Levels in Blood

RATIONALE: Individuals with allergic asthma respond differently, but reproducibly, to allergen inhalation challenge. Some individuals develop an isolated early response (early responders) (ERs), whereas others go on to develop a late response (dual responders) (DRs). It is not understood why late responses do not develop in all sensitized individuals. OBJECTIVES: The aim of this study was to identify blood biomarkers that can discriminate ERs and DRs using cellular frequencies and gene and metabolite expression from whole blood. METHODS: Thirty-two individuals participated in the allergen inhalation challenge as part of the AllerGen Clinical Investigator Collaborative. Fifteen participants were classified as ERs and 17 as DRs. Blood samples were collected before (pre) and 2 hours after (post) the allergen challenge. Cell counts were obtained using a hematolyzer, gene transcript relative levels using RNA sequencing, and metabolite concentrations using tandem mass spectrometry. An integrative ensemble algorithm that was based on canonical correlation analysis was used to classify ERs and DRs using all three data sets, adjusting for age and sex. The objective of this algorithm was to identify a correlated subset of molecules from each data set that best discriminated ERs from DRs. Gene set enrichment analysis was performed using Enrichr (Chen et al., BMC Bioinformatics 2013;128). MEASUREMENTS AND MAIN RESULTS: The pre-challenge multisignature classifier (error = 30%) outperformed the post-challenge multisignature classifier (error = 50%) in separating ERs from DRs. The cells selected in the prechallenge multisignature panel included eosinophils, lymphocytes, and neutrophils. The selected metabolites were enriched for glycerophospholipids. The subset of gene transcripts in the multisignature panel was enriched for the T-cell receptor and costimulatory signaling pathway (P = 3.4 × 10(-6)) (Wikipathways) and positive regulation of antigen receptor-mediated signaling pathway (P = 5.7 × 10(-4)) (GO Ontology). CONCLUSIONS: This study provides a systems perspective on the deregulated molecular processes between early and dual responses in whole blood. The integrative biomarker analysis suggests that a molecular signature that is predictive of the late-phase response can be identified. The variability in the onset of the late response may explain the poor predictive performance of the postchallenge multiomic biomarker signature. Replication of the prechallenge biomarker signature in additional independent samples is required to validate this panel.