A transcriptomic atlas of chronic lung allograft dysfunction

Chronic lung allograft dysfunction (CLAD) is the leading cause of late mortality after lung transplantation. Bronchiolitis obliterans syndrome (BOS) and restrictive allograft syndrome (RAS) are the main underlying clinical entities. Their molecular and cellular signatures are unclear and, therefore, we aimed to identify molecular programmes associated with morphological disease severity in CLAD.

We performed high-resolution imaging-based gene expression profiling of 128 lung samples from explanted CLAD and donor lungs, using weighted gene co-expression network analysis, cellular and pathway enrichment, and hub gene identification. Findings were validated across four datasets, including a murine transplant model, human BOS lungs, transbronchial biopsies, and bronchoalveolar lavage fluid of lung transplant recipients.

Unsupervised clustering revealed two transcriptomic CLAD endotypes aligning with mild-fibrotic (BOS, mild RAS) and advanced-fibrotic disease (moderate/severe RAS). Samples from the same patient often diverged molecularly, underscoring intrapatient heterogeneity and limitations of current phenotypical classification. Five molecular programmes emerged: 1) epithelial stress and innate immunity in early-fibrotic CLAD; 2) progressive adaptive immunity and cytotoxicity in advanced CLAD; 3) transient extracellular matrix remodelling; 4) progressive endothelial loss/dysfunction; and 5) progressive loss of homeostasis, wherein multiple potential druggable targets were detected. Finally, we identified a 26 CLAD-hub-gene panel, which showed robust diagnostic performance to discriminate CLAD.

CLAD is a spatially heterogeneous, yet molecularly continuous disease process, wherein BOS and RAS represent variable stages of a shared immunopathological continuum. Our findings support the development of lung-specific molecular classifiers to guide diagnostics and reveal novel targets for personalised therapies in transplantation.