Single Cell Transcriptomics Based Deconvolution Of Radiation Induced Lung Fibrosis (RILF) Model

Abstract

RILF is dose-limiting for curative treatment of thoracic malignancies. Deconvolution of cellular subtypes involved in RILF and their crosstalk might unravel novel insights into the pathophysiological processes governing this relevant late radiation side effect. C57BL6 mice were irradiated with a single dose of 15Gy whole thoracic irradiation (RT). Presence of radiation induced fibrosis was confirmed 24 weeks post RT by CT scans. Fibrotic lungs (n:3) and aged-matched non-irradiated reference lungs (n:2) were isolated and disaggregated. Single cell suspensions were analyzed by next-generation sequencing (RNAseq, 3’, 10X genomics). The Seurat package was used for data analysis: the shared nearest neighbor algorithm for cell clustering, t-test and ROC analyses for differential gene expression. Unsupervised clustering identified 17 cell type defining subpopulations. Overall TGFb1 expression was increased after RT (8% to 20%) and could be attributed to lipofibroblast (Chil3, Lpl) and endothelial cell (EC, CD31/Pecam1) subpopulations. Lipofibroblasts were found to coexpress PD-L1 and TGFb1, and split into TGFb1high and TGFb1interim populations following radiation. ECs showed increase of TGFb1, the cell cycle regulators Cdkn2a (p16), Ccnd1 and mesenchymal markers Serpine1 (PAI-1) and Col4a2. A differential EC-cluster (Sparc+) was detected following irradiation. Immune cell infiltration was increased after RT (11% to 36%): the APC-cell (H2-Aa, H2-Eb1) subgroup showed upregulation of inflammatory macrophages associated genes Cd83 and Ccl4. In the B/T-cell subgroup expression of Ccl17 (Th2 chemokine ligand) was increased in RT lungs. Together, RT induced a pro-fibrotic M2/Th2 state with increased TGFb1 expression and transcriptome changes hinting towards endothelial-to-mesenchymal transformation. Single cell RNAseq of RILF model reveals novel insights into the cellular heterogeneity and subpopulation specific responses to RT. Confirmation of these data in context of human disease will navigate towards development of new rationales for treatment of RILF.