KL-6 assisted subtyping of ARDS: from subtype-specific metabolomics to LPCAT1 as a pathogenic target.

BACKGROUND: Acute lung injury (ALI) and the resulting acute respiratory distress syndrome (ARDS) exhibit marked clinical and biological heterogeneity, hindering subtype identification, biomarker discovery, and targeted therapy development. We used untargeted metabolomics to identify serum and bronchoalveolar lavage fluid (BALF) biomarkers across ARDS subtypes and potential therapeutic targets. We introduce a KL-6–assisted refinement of the Berlin Definition to separate patients with similar PaO₂/FiO₂ ratios but differing degrees of alveolar epithelial damage. METHODS: We enrolled 166 participants (137 ARDS, 29 controls). Using the Berlin Definition with KL-6–assisted classification, patients were stratified into common and severe subtypes. Serum metabolomes were compared between ARDS and controls and between subtypes; paired serum–BALF profiles characterized systemic vs. pulmonary features. Overlapping metabolites from four comparisons informed target prediction (SwissTargetPrediction, GeneCards), followed by in vitro and in vivo validation in LPS models. RESULTS: Serum profiles and pathway enrichment differed between ARDS and controls and between subtypes. The severe subtype was enriched in fatty-acid oxidation, dicarboxylic acids, and phospholipid remodeling. Three serum metabolites discriminated severe from other less severe ARDS subtypes (AUC > 0.8). This 3-metabolite panel achieved AUC 0.917 in 10-fold cross-validation. Paired analyses showed predominant systemic alterations with subtype-specific BALF signatures; five BALF metabolites were specific to the severe subtype and correlated with KL-6, oxygenation, inflammation, and CT scores. Five shared LysoPCs highlighted LPCAT1 as the top target. LPS upregulated LPCAT1, and this effect was partially reversed by TAK-242, a TLR4 inhibitor. Knockdown of LPCAT1 reduced inflammatory cytokine and MUC1 (KL-6) release in cell culture supernatants. CONCLUSION: KL-6–assisted subtyping reveals distinct systemic and pulmonary metabolic signatures. Three serum metabolites provide evidence that justifies KL-6 as a marker of epithelial-injury–informed ARDS classification, and our results suggest that LPCAT1 is a promising therapeutic target for ARDS.