Fig. 6

Potential mechanism of smoking facilitating COPD development: Alveolar impairment due to the detrimental intercellular communications between macrophages and endothelial cells. A The sunburst diagrams illustrating the cell number proportion of active smoker (AS)/never-smoker (NS) and COPD/control in the macrophage three subtypes (A, B, and C). The inner ring represents the cell number proportion of macrophage subtype A, B and C in macrophage. The middle ring represents the cell number proportion of active smoker and never-smoker in macrophage subtype A, B and C. The outer ring represents the cell number proportion of COPD and control in macrophage subtype A from active smoker (AS)/never-smoker (NS), macrophage subtype B from active smoker (AS)/never-smoker (NS) and macrophage subtype C from active active smoker (AS)/never-smoker (NS). B Box plots showing the GSVA analysis enrichment score of inflammatory response, antigen processing and presentation via MHC class II, and mitochondrial gene expression in A, B, and C subtypes of macrophages. C Violin plots for the expression of genes involved in mitochondrial electron transport including MT-ATP8, MT-ND2, MT-ND3, MT-ND1, MT-ND5, and MT-ND4L in subtypes A, B, and C of macrophages. D Violin plots for the expression of 8 genes encoding enzymes implicated in sphingolipid metabolism (SPTLC1, SPTLC2, CERS2, CERS5, ASAH1, SGPP1, CERK, and SMPD4) in macrophages from active smokers (AS) and never-smokers (NS). E Bone marrow-derived macrophages (BMDMs) from C57BL/6 mice were treated with (0.5% CSE, gray bars; 1% CSE, black bars) or without (ctrl, white bars) cigarette smoke extract (CSE) for 24 h, when the mRNA levels of SPTLC1, SPTLC2, CERS2, CERS5, ASAH1, SGPP1, CERK, and SMPD4 were measured by qPCR in the BMDMs for each group (*P < 0.05 vs. ctrl, **P < 0.01 vs. ctrl). F Schematics of the correlation analysis we designed based on summarized expression profiles at the individual level between macrophages and other cell types. Cell-type-specific mean gene expression within each individual is first calculated. As a result, each individual has an expression matrix with rows corresponding to genes and columns corresponding to cell types. Next, Spearman correlation analysis between mean expression values of gene A of all individuals in sender cell type 1 and mean expression values of gene B of all individuals in recipient cell type 2 is performed to examine whether gene B in recipient cells changes coordinately with gene A dysregulated in sender cells. Finally, the correlation between the mean expression values at individual level of gene A in sender cells and the mean expression value at individual level of gene B in recipient cells across active smokers (AS) and never-smokers (NS) is visualized as scatterplot. G Heatmap showing the proportion of the number of genes with a correlation no less than 0.8 (Spearman rank correlation coefficient ≥ 0.8 or ≤ − 0.8) with each macrophage significantly up-regulated ligand gene expression and dysregulated sphingolipid metabolic enzyme gene expression to the number of expressed genes in each cell type. Each row is a cell type, and each column is a differential expressed ligand/enzyme gene in macrophages. White indicates low proportion and red indicates high proportion. Rows are ordered from top to bottom by the mean proportion of genes highly correlated with differential expressed ligands and enzymes in macrophages in each cell type. H Sankey plot demonstrating the enriched biological processes among the correlated genes in macrophages (left) and endothelial cells (right) to the smoking-associated dysregulated enzymes that participate in sphingolipid metabolism in macrophages (centre). I Scatter plots showing the cell-type-specific mean gene expression within each individual correlations between certain sphingolipid metabolic enzyme genes in macrophages and several representative highly correlated genes in macrophages or endothelial cells across active smokers (AS) and never-smokers (NS)