Computational Investigation of Endothelial Heterogeneity During Homeostasis and Inflammation

Abstract

Endothelial cells (ECs) are a specialized cell type lining all vertebrate blood vessels and form an interface between circulating blood and the neighboring tissue. Understanding the tissue-specific characteristics of endothelial cells could markedly improve our understanding of the organ-specific roles of blood vessel function and the development of vascular disease. Computational approaches statistically modeling the gene expression data provided by high throughput sequencing technologies have been developed to analyze the cellular transcriptome. In our work, we present novel computational methods, namely HeteroPath and Subnetwork Signaling Entropy Analysis (SSEA), to analyze gene expression data and ascertain organ-specific endothelial heterogeneity during homeostasis and inflammation. We hypothesized that characterizing cells from distinct tissues based on the heterogeneity of the molecular signaling would allow for the precise identification of clusters of genes which are uniquely upregulated or uniquely downregulated in each tissue. Using HeteroPath alongside traditional gene set enrichment analysis methods, we demonstrated endothelial transcriptomic heterogeneity. HeteroPath specifically identified organ-specific signaling pathways and provided a comprehensive characterization of EC heterogeneity in the healthy state. We next adopted the RiboTag mRNA isolation technique to directly isolate tissue-specific mRNAs undergoing translation without cell disassociation to understand the nature of the endothelial translatome in vivo. By performing RNA-Sequencing and computationally analyzing the endothelial translatome, we identified specific pathways, transporters, and cell-surface markers expressed in an organ-specific manner. In addition, we found that ECs adopt the characteristics of the tissue by expressing genes typically expressed in the surrounding tissue such as genes associated with synaptic function in the brain endothelium and cardiac contractile genes in the heart endothelium. Once we established the organ-specific endothelial signature during homeostasis, we studied whether this heterogeneity persisted in response to a biological stimulus that induced systemic inflammation. Using differential expression approaches and our novel framework, SSEA, we quantified the organ-specific endothelial gene expression dynamics and found that the progression and resolution of endothelial injury during vascular inflammation in each organ is mediated by distinct endothelial signaling mechanisms. Using these methods and tools, we characterized organ-specific endothelial heterogeneity during homeostasis and inflammation and provided insights regarding the underlying endothelial biology and potential therapeutic targets.