Lung injury-induced activated endothelial cell states persist in aging-associated progressive fibrosis

Abstract

Progressive lung fibrosis is associated with poorly understood aging-related endothelial cell dysfunction. To gain insight into endothelial cell alterations in lung fibrosis we performed single cell RNA-sequencing of bleomycin-injured lungs from young and aged mice. Analysis reveals activated cell states enriched for hypoxia, glycolysis and YAP/TAZ activity in ACKR1+ venous and TrkB+ capillary endothelial cells. Endothelial cell activation is prevalent in lungs of aged mice and can also be detected in human fibrotic lungs. Longitudinal single cell RNA-sequencing combined with lineage tracing demonstrate that endothelial activation resolves in young mouse lungs but persists in aged ones, indicating a failure of the aged vasculature to return to quiescence. Genes associated with activated lung endothelial cells states in vivo can be induced in vitro by activating YAP/TAZ. YAP/TAZ also cooperate with BDNF, a TrkB ligand that is reduced in fibrotic lungs, to promote capillary morphogenesis. These findings offer insights into aging-related lung endothelial cell dysfunction that may contribute to defective lung injury repair and persistent fibrosis.

Fig. 1. Activation of mouse lung ECs following bleomycin challenge.

A UMAP embedded visualization of different EC subpopulations (n = 24,168 cells). B Heatmap showing differentially expressed activation marker genes in each EC subpopulation. C Heatmap showing differentially expressed hypoxia/glycolysis marker genes in each EC subpopulation. D, E Ingenuity pathway analysis shows canonical pathways and upstream regulators enriched in all activated ECs. P values were generated using Right-tailed Fisher’s Exact Test (log2 FC ≤ −0.5 or ≥0.5, P value ≤ 0.05). F Heatmap showing differentially expressed activation marker genes in each experimental group. Young Sham (YS, n = 5699 cells); Young Bleo (YB, n = 2208 cells); Aged Sham (AS, n = 6810 cells); Aged Bleo (AB, n = 9451 cells). G Heatmap showing differentially expressed hypoxia/glycolysis marker genes in each experimental group. H Hypoxia detection by hypoxyprobe in young and aged mouse lungs at 60 days post bleomycin administration. Immunofluorescence staining using antibodies against Pimonidazole adducts and PECAM-1 shows reduced number of alveolar ECs in fibrotic aged lungs exhibiting elevated hypoxia compared to young one in which hypoxia levels were undetectable. Values are summarized as mean ± SEM and analyzed using a two-tailed Student’s t-test. Young Bleo day 0 (n = 3), Young Bleo day 60 (n = 3), Aged Bleo day 0 (n = 3), Aged Bleo day 60 (n = 7). I Schematic showing the experimental strategy and t-SNE displaying different EC subpopulations in sham (Blue, n = 3), injured lungs after 14 days (B14D, Red, n = 2), and injured lungs at 35 days (B35D, Orange, n = 2;) post bleomycin-induced lung injury. J, K Violin plots showing increased expression of activation and hypoxia/glycolysis marker genes in injured aged lungs at 14 days followed by a return to baseline during the early resolution phase (day 35 post bleomycin challenge), sham (n = 1994 cells), B14D (n = 2164 cells), and B35D (n = 1807 cells). L t-SNE plots showing the expression of EGFP across different EC clusters (n = 5965 cells). M Schematic showing arterial and vein EC subtypes expressing capillary markers. N Violin plot showing the expression of EGFP across different EC subpopulations (n = 5965 cells). O Venn diagram shows nearly 60% transcriptional overlapping between EC states at day 14 post bleomycin (young lungs (Y)) and day 30 post bleomycin (aged lungs (A)). The top two thousand differentially expressed genes (P < 0.05, determined by BioTuring Browser 3) were included in this analysis. Each box plot displays the median value as the center line, the upper and lower box boundaries at the first and third quartiles (25th and 75th percentiles), and the whiskers depict the minimum and maximum values. Source data are provided as a Source Data file.

Fig. 2. Venous EC remodeling in young and aged mouse lungs following bleomycin challenge.

A UMAP plots showing venous EC clusters (n = 1961 cells). C; cluster. B, C Composition and UMAP plots displaying venous EC clusters in young (YS, Blue, n = 302 cells) and aged (AS, Orange, n = 658 cells) uninjured lungs, and young (YB, Red, n = 199 cells) and aged (AB, Green, n = 802 cells) injured lungs. Venous ECs in Clusters 3 and 4 exclusively emerged in bleomycin-injured lungs and their number increased in fibrotic aged lungs compared to young ones. D Heatmap showing average expression of venous EC marker genes across different venous EC clusters (E) Violin plots showing the expression of distinctive marker genes in activated venous ECs (clusters 4). C1 (n = 873 cells), C2 (n = 231 cells), C3 (n = 778 cells), C4 (n = 79 cells). F, G Ingenuity pathway analysis shows canonical pathways and upstream regulators enriched in activated venous ECs (clusters 4). P values were generated using Right-tailed Fisher’s Exact Test (log2 FC ≤ −0.5 or ≥ 0.5, P value ≤ 0.05). H, I UMAP embedded visualizations of different venous EC clusters during different time point of bleomycin challenge, sham (n = 383 cells), bleomycin 14 days (B14D, n = 703 cells), bleomycin 35 days (B35D, n = 253 cells), Venous #1 (n = 404 cells), Venous #2 (n = 229 cells), Venous #3 (n = 271 cells), Venous #4 (n = 435 cells). A; Activated. J, K UMAP and violin plots showing the expression of EGFP across different venous EC clusters (n = 1339 cells). L UMAP and violin plots showing the expression of Selp and Ackr1 genes across different venous EC clusters (n = 1339 cells). M Immunofluorescence staining showing ACKR1 (venous EC marker) and CD31 (Pan-endothelial marker) expression in lung airways and alveoli at baseline. Br; Bronchiole, Av; Alveoli. N Immunofluorescence staining showing ACKR1 expression in Col1a1-GFP mouse lung at 14 days after bleomycin injury. O Immunofluorescence staining using antibodies against ACKR1 and α-SMA in young and aged lungs following bleomycin injury. ACKR1 positive venous EC accumulates in lung areas exhibiting extensive remodeling. Each box plot displays the median value as the center line, the upper and lower box boundaries at the first and third quartiles (25th and 75th percentiles), and the whiskers depict the minimum and maximum values. Source data are provided as a Source Data file.

Fig. 3. ACKR1+ venous ECs expand in human fibrotic lungs.

A Immunofluorescence staining of normal human lungs using antibodies against ACKR1, Collagen-I, and αSMA. ACKR1 is mainly expressed in peribronchial and alveolar venous ECs. Perivascular cells surrounding ACKR1 positive venous ECs strongly expressed Collagen-I and αSMA. Br Bronchiole, Av Alveoli. B Immunohistochemistry of human IPF lung sections showing normal-looking alveoli surrounded by highly vascularized (CD31 positive) fibrotic areas. C Masson trichrome staining of human IPF lung sections showing peribronchial fibrosis. D Immunofluorescence staining of human IPF lung sections showing ACKR1+ veins that surround a fibrotic small bronchus extend toward the alveolar parenchyma. E ACKR1 positive veins with αSMA+ mural cells are found in the fibrotic alveolar parenchyma.

Fig. 4. ACKR1+ venous ECs expand in lPF lung regions containing elevated number of pathogenic fibroblasts.

FACS analysis was performed on human normal and IPF lung tissues. A Flow cytometry plot analyses are shown across different regions of fibrotic lungs using antibodies for the epithelial cell marker EpCAM, the immune cell marker CD45, the endothelial cell marker CD31, and (B) the venous EC markers ACKR1 and P-Selectin. C Quantitation of flow cytometry data demonstrates that ACKR1 + P-Selectin+ venous ECs were enriched in fibrotic areas compared to normal tissues containing a high number of pathogenic fibroblasts (Thy1 + CTHRC1 + ). One section from one normal lung and two sections from three different areas of one IPF lung were analyzed. Box borders display the minimum and maximum values, with a central line at the mean. Source data are provided as a Source Data file.

Fig. 5. Lung injury promotes capillary EC transition to activated/immature states that persist in aged and IPF lungs.

A UMAP plot of capillary ECs (n = 20960). Each color indicates a distinct EC state at baseline and in response to injury. Quiescent (Q), activated (A). B, C Composition of gCap and aCap EC clusters across different groups. Young Sham (YS, n = 5206 cells); Young Bleo (YB, 1867 cells); Aged Sham (AS, n = 5972 cells); Aged Bleo (AB, n = 8015 cells). D Violin plots showing the de novo expression of activated EC marker genes across different gCap (Q, (n = 10256 cells), 1 (n = 4354 cells), 2 (n = 3070 cells), 3 (n = 1121 cells)) and aCap EC (Q (n = 1283 cells), 1 (n = 545 cells), 2 (n = 331 cells)) clusters. E Violin plots showing the expression of quiescent gCap EC marker genes across different clusters. Activated gCap EC in cluster 2 and 3 exhibit the strongest reduction of canonical gCap EC marker genes. F Violin plots showing the expression of quiescent aCap EC marker genes across different clusters. Activated aCap EC (cluster 1 and 2) exhibit the strongest reduction of canonical aCap EC marker genes. Each box plot displays the median value as the center line, the upper and lower box boundaries at the first and third quartiles (25th and 75th percentiles), and the whiskers depict the minimum and maximum values. G FACS analysis of normal (n = 5) and IPF (n = 7) lungs. Antibodies against TEK and EDNRB were used to discriminate gCap EC from aCap EC. Out of all CD31 positive cells gCap ECs were defined as (TEK^High EDNRB^Low) whereas aCap ECs strongly expressed EDNRB and were largely negative for TEK. IPF lungs exhibited an increased percentage of ECs expressing low levels of the gCap marker TEK compared to healthy lungs. Values are summarized as mean ± SEM and analyzed using a two-tailed Student’s t-test. Source data are provided as a Source Data file.

Fig. 6. TrkB is a marker of activated capillary ECs in response to bleomycin challenge.

A t-SNE displaying different capillary EC subpopulations in sham (Blue, n = 1346 cells), injured lungs after 14 days (B14D, Red, n = 1067 cells), and injured lungs after 35 days (B35D, Orange, n = 1399 cells) post bleomycin-induced lung injury. B, C t-SNE and violin plots showing that the expression of Ntrk2 gene is enriched in capillary ECs at 14 days after injury and returned to baseline at day 35 post bleomycin challenge. Sham (n = 1346 cells), B14D (n = 1067 cells), B35D (n = 1399 cells). D, E t-SNE and violin plots showing the expression of EGFP gene following injury. gCap ECs (sham (n = 964 cells), B14D (n = 800 cells), B35D (n = 1002 cells)) aCap ECs (sham (n = 382 cells), B14D (n = 267 cells), B35D (n = 397 cells)). F Immunofluorescence staining using antibodies against GFP and CAR4 in young lungs of uninjured sham and 28 days after bleomycin injury. G Quantification of immunofluorescence staining at 28 days after bleomycin injury showing the retention of EGFP expression by gCap ECs at different time points and the absence of EGFP expression in aCap ECs (n = 3). Values are summarized as mean ± SEM, P values were generated using one-way ANOVA with Tukey’s post hoc test for comparison. H Immunofluorescence images showing the expression of TrkB in gCap ECs after bleomycin challenge. gCap ECs were lineage labeled in Aplnr-CreER(T)-mTmG mice 15 days prior to bleomycin administration (Day 0). Sham and bleomycin-injured lungs were harvested 28 days post bleomycin delivery followed by immunofluorescence analysis. An antibody against TrkB was used to detect injured gCap ECs (Red). gCap ECs co-expressing EGFP and TrkB (yellow) only emerged in injured lungs. Each box plot displays the median value as the center line, the upper and lower box boundaries at the first and third quartiles (25th and 75th percentiles), and the whiskers depict the minimum and maximum values. Source data are provided as a Source Data file.

Fig. 7. TrkB+ gCap ECs accumulated in fibrotic aged mouse lungs and human IPF lungs.

A Immunofluorescence images showing the localization of TrkB+ gCap ECs in lung areas of fibroblast aggregation (GFP+ fibroblasts) at 14 and 37 days after bleomycin challenge, n = 3. B Immunofluorescence staining using antibodies against TrKB and collagen-I in young and aged lungs following bleomycin injury (n = 4). TrkB+ gCap ECs are localized in lung areas exhibiting collagen deposition. Values are summarized as mean ± SEM and analyzed using a two-tailed Student’s t-test. Source data are provided as a Source Data file. C Images of normal (n = 3) and IPF (n = 3) human lungs stained by immunohistochemistry for TrKB and CD93.

Fig. 8. Activation of YAP/TrKB axis enhances lung capillary morphogenesis.

A, B qPCR analyses of human lung microvascular ECs (HLMVECs) treated with the LATS1/2 inhibitor TRULI and siRNAs targeting YAP and TAZ for 48 or 72 h, respectively. YAP activation in these cells partially recapitulates the gene expression signature observed in activated gCap ECs. Values are summarized as mean ± SEM and analyzed using a two-tailed Student’s t-test. Yap activation (n = 3 independent experiments), YAP/TAZ inhibition (n = 4 independent experiments). C Dot plot showing the expression of Bdnf in different populations of lung cells. Dot size indicates the proportion of expressing cells, colored by standardized expression levels. D t-SNE plot showing different alveolar epithelial cell clusters. ATII (Blue, n = 1470 cells), intermediate AE (intAE, Red, n = 176 cells), and ATI (Orange, n = 229 cells). E t-SNE plot showing the expression of Bdnf in intAE and ATI. F, G UMAP visualization of intermediated alveolar epithelial cell clusters and marker gene signatures. H Violin plots showing the expression of ATI cell marker gene, Pdpn, and Bdnf in intAE clusters, 1 (103 cells), 2 (73 cells). Each box plot displays the median value as the center line, the upper and lower box boundaries at the first and third quartiles (25th and 75th percentiles), and the whiskers depict the minimum and maximum values. I Cell composition in each intAE cluster. Most intermediate alveolar epithelial cells exhibiting ATII marker genes (less differentiated) were from bleomycin-treaded aged lungs. Young Sham (YS); Young Bleo (YB); Aged Sham (AS); Aged Bleo (AB). J Expression of BDNF ATI cells derived from human normal (n = 23) and IPF (n = 24) lung. Values are summarized as means ± SEM and analyzed using a two-tailed Student’s t-test (K) Schematic showing the in vitro 3D endothelial morphogenesis assay. Staining of HLMVECs with phalloidin shows that BDNF and TRULI treatment synergistically promote tube formation. L Quantitative analysis of vascular morphogenesis was indicated as the total branches length. Image analysis in the whole photographed area were performed by angiogenesis analyzer (ImageJ software). Values are summarized as means ± SEM, P values were generated using one-way ANOVA with Tukey’s post hoc test for comparison. n = 3 independent experiments. Source data are provided as a Source Data file.