Integrated single-cell RNA sequencing analysis reveals alterations of ageing human lung endothelium heterogeneity in idiopathic pulmonary fibrosis

Eamon C. Faulkner, Adam A. Moverley, Simon P. Hart, Leonid L. Nikitenko


Increasing age is the main risk factor for chronic lung diseases (CLD) including idiopathic pulmonary fibrosis (IPF). Halting or reversing progression of IPF remains an unmet clinical need due to limited knowledge of underlying mechanisms. In particular, the contribution of the endothelium to ageing in human lung under physiological conditions and in IPF remains insufficiently understood. In this study, we analysed heterogeneity of endothelium in physiologically ageing human lung and its alterations in IPF. We conducted a comprehensive in silico analysis of scRNAseq profiles of human lung tissues from older healthy donors and age-matched IPF patients (n=9 for each group) by integrating datasets from two independent cohorts. We generated a single-cell map of the ageing human lung and identified 17 subpopulations of ageing endothelium (12 for blood and 5 for lymphatic vessels, including 4 “de-differentiated”), with distinct transcriptional profiles, specific gene expression signatures and percentage contributions, revealing previously underappreciated extent of heterogeneity. In IPF lung, the balance of different endothelial sub-types was significantly altered both in terms of cell numbers and gene expression patterns, identifying disease-relevant subpopulations and transcriptional changes associated with specific signalling pathways and cellular processes. These findings reveal a previously unrecognised phenomenon of ageing human lung endothelium re-programming towards an “IPF endothelium” state, suggesting potential avenues for therapeutic management or biomarker discovery for diagnostics or prognostics of IPF. Our study creates a conceptual framework for appreciating the heterogeneity of ageing endothelium and its alterations in CLDs and diseases associated with fibrosis in other organs, including lymphoedema and cancer.


Schematic of methodology for samples integration and comprehensive data analysis.
UMAP representation of all cells and cell clusters from all 18 samples (pooled).
Heatmap of top ten differentially expressed genes by the cluster.
Stacked bar chart of percentage contribution of each cluster to total lung population, split by cohort and condition - donor or fibrosis (IPF).