Deciphering pathogenic cellular module at single-cell resolution in checkpoint inhibitor-related pneumonitis

Checkpoint inhibitor pneumonitis (CIP) is the most common fatal immune-related adverse event; however, its pathophysiology remains largely unknown. Comprehensively dissecting the key cellular players and molecular pathways associated with CIP pathobiology is critical for precision diagnosis and develop novel therapy strategy of CIP. Herein, we performed a comprehensive single-cell transcriptome analysis to dissect the complexity of the immunological response in the bronchoalveolar lavage fluid (BALF) microenvironment. CIP was characterized by a dramatic accumulation of CXCL13+ T cells and hyperinflammatory CXCL9+ monocytes. T-cell receptor (TCR) analysis revealed that CXCL13+ T cells exhibited hyperexpanded- TCR clonotypes, and pseudotime analysis revealed a potential differentiation trajectory from naïve to cytotoxic effector status. Monocyte trajectories showed that LAMP3+ DCs derived from CXCL9+ monocytes possessed the potential to migrate from tumors to the BALF, whereas the differentiation trajectory to anti-inflammatory macrophages was blocked. Intercellular crosstalk analysis revealed the signaling pathways such as CXCL9/10/11-CXCR3, FASLG-FAS, and IFNGR1/2-IFNG were activated in CIP+ samples. We also proposed a novel immune signature with high diagnostic power to distinguish CIP+ from CIP- samples (AUC = 0.755). Our data highlighted key cellular players, signatures, and interactions involved in CIP pathogenesis.