Endothelial sensitivity to pro-fibrotic signals links systemic exposure to pulmonary fibrosis

Abstract

Pulmonary fibrosis (PF) is a life-threatening condition characterised by excessive extracellular matrix deposition and tissue scarring. While much of PF research has focused on alveolar epithelial cells and fibroblasts, endothelial cells have emerged as active contributors to the disease initiation, especially in the context of systemic exposure to pro-fibrotic substances. Here, we investigate early transcriptomic and secretory responses of human umbilical vein endothelial cells (HUVEC) to subtoxic doses of bleomycin, a known pro-fibrotic agent, and TGF-beta, a key cytokine in fibrosis. Bleomycin exposure induced a rapid and extensive shift in the endothelial transcriptional programme, including signatures of endothelial to mesenchymal transition, cellular senescence, and immune cell recruitment. These findings suggest endothelial cells as early initiators of pro-fibrotic signals, independent of contributions from other cell types. In contrast, TGF-beta effects were limited and transient, indicating its pro-fibrotic action may require another initial stimulus and interplay with other cells like fibroblasts. This study highlights the sensitivity of endothelial cells to pro-fibrotic exposure and provides a blueprint of early pro-fibrotic mechanisms that may operate on organs such as the lungs systemically via the endothelium, emphasising its pivotal role in PF pathogenesis.

Fig. 1. Overview of RNA-Seq data.

A Total raw counts obtained per sample for both datasets. Wilcoxon rank sum test showed no significant difference in raw counts between the two datasets (p value = 0.12). B Distribution of total normalised counts per gene summed up over all samples (n = 12) of unexposed cells in both datasets. C Distribution of total normalised counts per gene summed up over all samples (n = 36) of exposed cells in both datasets. D Set size and intersection size of dose-dependent (DD) and differentially expressed (DE) genes in both exposures. The union of DD and DE genes detected in the different experimental sub-groups (timepoints, concentrations) is plotted.

Fig. 2. Bleomycin is a molecular initiator of fibrotic conditions.

A AOP:173 (Substance interaction with the pulmonary resident cell membrane components leading to pulmonary fibrosis). KEs in blue were enriched by dysregulated genes under bleomycin at each of the 3 timepoints. KEs in grey were not enriched at any timepoint under bleomycin exposure. The square indicates the MIE, the triangle the AO. B The fraction of enriched KEs under Bleomycin and TGF-beta exposure that are of molecular-cellular and organ-systemic biological scale, respectively. KEs of molecular and cellular biological level were merged, presenting MIEs and early KEs of an AOP, while KEs of tissue, organ, and individual biological level were considered as of organ-systemic biological scale, presenting later KEs and AOs of an AOP. Difference between the two substances was tested by two-sided Fisher’s exact test (p value = 2e−11). C The systems of enriched KEs as recently annotated by Saarimäki et al. for bleomycin (top) and TGF-beta (bottom). D Experimental data in comparison with public RNA-Seq data sets of PF. Sixteen public data sets and in total 51 comparisons were recently curated by Inkala et al. (10.5281/zenodo.10692129). Fold changes of samples of PF versus healthy controls were compared with fold changes computed for HUVEC cells exposed to bleomycin and TGF-beta, respectively. The multi-dimensional scaling plot was generated from the Euclidean distance between datasets based on the ranks of 8127 common genes according to their fold changes. BAL bronchoalveolar lavage.

Fig. 3. Transcriptomic changes under bleomycin exposure.

A Functional analysis of up- and downregulated genes under bleomycin exposure. Overrepresentation test with dysregulated genes (absolute LFC ≥1, adjusted p value ≤0.01) against the gene sets of KEGG pathways. The dot size indicates the gene ratio, i.e., the ratio of identified genes to pathway genes. The colour code indicates the Benjamini–Hochberg-adjusted p value of the Fisher’s exact test. The top five enriched pathways are plotted, and already existing pathways are filled even if they are not among the top five pathways. B Expression of apoptosis marker genes for the different experimental conditions of bleomycin exposure. Plotted are the size-factor normalised RNA-Seq counts.

Fig. 4. Fibrosis-associated cellular processes in endothelial cells.

A Cross-section of an alveolus and the surrounding capillaries highlighting key endothelial cell processes associated with PF. Processes are annotated as follows: a green checkmark indicates that RNA-Seq data showed the process was affected by bleomycin and TGF-beta exposure, respectively; a red cross indicates the process was not affected; and a yellow circle indicates that the process was only mildly or transiently affected. B Marker genes of EndMT under bleomycin exposure. The colour annotation of columns indicates the experimental condition, the row annotations refer to the label of the time-dose modelling, and the results of the differential expression and dose-dependent analysis. The heatmap colour code represents row-wise z-scores based on vst-transformed counts. C Marker genes of EndMT under TGF-beta exposure. Annotations and colour codes are the same as described in (B).