Front Physiol. 2026 Oct 13;17:1951891. doi: 10.3389/fphys.2026.1951891. eCollection 2026.
ABSTRACT
BACKGROUND: Interstitial lung disease (ILD) comprises a heterogeneous group of disorders characterized by variable clinical phenotypes and dysregulated immune responses. The marked diversity in disease behavior and therapeutic response suggests the existence of distinct molecular endotypes. In this study, we applied an unsupervised clustering analysis to bronchoalveolar lavage (BAL) cytokines to identify clinically meaningful ILD subpopulations.
METHODS: BAL cytokine data from ILD patients were analyzed using K-means clustering, identifying four primary clusters: (C1) TGF-β/IL-17A dominant, (C2) IL-6 dominant, (C3) IL-35/IL-10 dominant, and (C4) lacking a dominant cytokine signature. Immune cell composition, clinical characteristics, and disease progression were compared across clusters.
RESULTS: Clusters C1 and C2 were associated with greater fibrotic burden. C1, characterized by elevated TGF-β and IL-17A levels, showed a higher prevalence of idiopathic pulmonary fibrosis (IPF) and increased BAL macrophages. C2, with high IL-6, included a higher proportion of smoke-related ILD (SR-ILD) and was associated with increased cardiovascular comorbidities, hospitalizations, and BAL neutrophils. In contrast, C3 and C4, with higher IL-35 and IL-10, were associated with lower fibrosis severity and a greater proportion of connective tissue disease-associated ILD (CTD-ILD). C3 and C4 showed more favorable clinical outcomes and disease stability over one year; notably, only C3 maintained stability over two years and exhibited increased BAL CD4+ T cells.
CONCLUSION: Clustering of BAL cytokine profiles revealed distinct immunological endotypes in ILD that correlated with fibrosis severity, clinical phenotype, and disease trajectory. This approach holds promise for improving patient stratification and prediction of disease progression in ILD.
PMID:42947868 | PMC:PMC13710069 | DOI:10.3389/fphys.2026.1951891