LATE-BREAKING ABSTRACT: Sputum transcriptome analysis yields eosinophilic and non-eosinophilic inflammatory mechanisms in UBIOPRED asthma cohort

Rationale: Granulocytic inflammation in sputum may be linked to diverse mechanisms. Clustering of the sputum transcriptome may help define granulocyte-specific disease drivers. Methods: In the UBIOPRED cohort, s putum eosinophils (EOS) ≥1.5% was used to define an EOS or ≤1.5% non-EOS phenotype. Differentially-expressed genes (DEGs) were obtained that differentiated between EOS, non-EOS and healthy volunteers (HV). Gene signatures were refined using nearest shrunken centroids algorithm (Robert et al. PNAS 2002, 99; 6567). Results: Clustering of 478 DEGs yielded clusters defined by 3 disease drivers, namely EOS- (n=31) and neutrophil- (NEU, n=31) enriched as well as a non-EOS non-NEU enriched (n=56) cluster (Fig 1). In the EOS-enriched cluster, signature refinement revealed 28 DEGs encoding multiple cytokine receptors and proteases including IL1RL1, TSLPR, CCR3, VSTM1, IL3RA, PRSS33 and CLC . For the NEU-enriched cluster, 27 DEGs mainly related to TNF and interferon regulation were found: TNFIP3, TNFSF10, CASP4, GBP1P1, CLEC4D, IFIT2, IFIH1, IFIT3 and IFI16 . Conclusion: Clustering on sputum cell genes reveal diverse mechanisms associated with sputa from asthma patients, with TNF- and IFN-associated genes in neutrophil-rich sputa.