Bronchoalveolar lavage metabolome dynamics reflect underlying disease and chronic lung allograft dysfunction

BackgroundProgression of chronic lung disease often leads to the requirement for a lung transplant (LTX). Despite improvements in short-term survival after LTX, chronic lung allograft dysfunction (CLAD) remains a critical challenge for long-term survival. This study investigates the relationship between the metabolome of bronchoalveolar lavage fluid (BALF) from subjects post-LTX with underlying lung disease and CLAD severity. MethodsUntargeted LC-MS/MS metabolomics was performed on 960 BALF samples collected over 10 years from LTX recipients with alpha-1-antitrypsin disease (AATD, n=22), cystic fibrosis (CF, n=46), chronic obstructive pulmonary disease (COPD, n = 79) or pulmonary fibrosis (PF, n=47). Datasets were analyzed using machine learning and multivariate statistics for associations with underlying disease and final CLAD severity. ResultsBALF metabolomes varied by underlying disease state, with AATD LT recipients being particularly distinctive (PERMANOVA, p=0.001). We also found a significant association with the final CLAD severity score (PERMANOVA, p=0.001), especially those with underlying CF. Association with CLAD severity was driven by changes in phosphoethanolamine (PE) and phosphocholine lipids that increased and decreased, respectively, and metabolites from the bacterial pathogen Pseudomonas aeruginosa. P. aeruginosa siderophores, quorum-sensing quinolones, and phenazines were detected in BALF, and 4-hydroxy-2-heptylquinoline (HHQ) was predictive of the final CLAD stage in samples from CF patients (R=0.34; p[≤]0.01). Relationships between CLAD stage and P. aeruginosa metabolites were especially strong in those with CF, where 61% of subjects had at least one of these metabolites in their first BALF sample after transplant. ConclusionsBALF metabolomes after LTX are distinctive based on the underlying disease and reflect final CLAD stage. In those with more severe outcomes, there is a lipid transition from PC to predominantly PE phospholipids. The association of P. aeruginosa metabolites with CLAD stages in LTX recipients with CF indicates this bacterium and its metabolites may be drivers of allograft dysfunction. Key messagesDespite the high prevalence of CLAD among LTX recipients, its pathology is not well understood, and no single molecular indicator is known to predict disease onset. Our machine learning metabolomic-based approach allowed us to identify patterns associated with a shift in the lipid metabolism and bacterial metabolites predicting CLAD onset in CF. This study provides a better understanding about the progression of allograft dysfunction through the molecular transitions within the transplanted lung from the host and bacterial pathogens.