CFTR modulator therapy reshapes airway inflammation in cystic fibrosis: Insights from proteomics and machine learning.

BACKGROUND: CFTR modulators, including Elexacaftor/Tezacaftor/Ivacaftor (ETI), have markedly improved clinical outcomes for people with cystic fibrosis (pwCF), but the molecular impact on airway inflammation remains incompletely understood. This study aimed to characterise longitudinal changes in airway inflammation and sputum proteomes following ETI treatment. METHODS: Sputum from pwCF (n = 30) was collected before ETI initiation and after 3 and 9-12 months of treatment. Sputum from healthy controls (n = 7) were included for comparison. Proteomes were analysed using data-independent acquisition liquid chromatography tandem mass spectrometry (DIA LC-MS/MS), and cytokines using Mesoscale assays. Differential expression analysis and correlations between airway proteomes and inflammatory cytokines were performed. Machine learning (XGBoost with bootstrapping approach) was applied to identify proteins predictive of ETI response. RESULTS: ETI induced broad proteomic shifts, mainly related to decreased neutrophil degranulation and an increase in anti-proteases. Machine learning predicted proteins linked to RNA splicing, ER stress and lipid transport as contributors to treatment response. IL-1β, IL-8, and TNFα decreased with treatment, correlating with neutrophil-related proteins. In contrast, IL-6 levels increased and correlated with mucin O-glycosylation pathways. Despite these improvements, proteomic and cytokine profiles remained distinct from healthy controls. CONCLUSION: ETI therapy reduces neutrophilic inflammation and restores the protease/antiprotease balance but does not fully normalise airway biology. Machine learning provides novel insights into molecular determinants of ETI response, suggesting a role for RNA splicing, ER stress and lipid metabolism. This dataset provides a valuable resource for further exploration of CF airway biology under ETI therapy.