Force-mediated recruitment and reprogramming of healthy endothelial cells in cerebral cavernous malformations

Cerebral cavernous malformations (CCMs) are vascular abnormalities in the brain characterized by clusters of leaky, tumor-like vessels that often lead to cerebral hemorrhage. Disruptions in signaling pathways involved in endothelial mechanotransduction and angiogenesis due to loss-of-function mutations in CCM genes are known to contribute to lesion formation. Recent studies have highlighted the critical role of wild-type cell recruitment from the surrounding endothelium by CCM mutants in lesion progression. However, the specific mechanisms underlying this recruitment process remain poorly understood. We report here that hyper-angiogenic CCM2-deficient endothelial cells enhance angiogenic invasion of neighboring wild-type cells. Mechanically hyperactive CCM2-deficient tips generate pulling forces on wild-type stalk cells and leave degraded paths in the matrix as cues to promote invasion. We wanted to further answer if mutant cells are capable of modifying the transcriptomic signature of neighbouring healthy endothelial changes in addition to the phenotypic modifications. We used single-cell RNA-sequencing (10x Genomics Chromium) to identifiy endothelial cell clusters. We enriched the endothelial cells and then used 10X Genomics Chromium technology to capture 1434 cells. The analysis was then performed using Seurat to identify each cell cluster. The results revealed that mutant cells were able to reprogram wild-type cells into stalk cells, with the activation of matrisome and DNA replication programs, eventually leading to cell proliferation. These findings unveil a novel mechanism by which proliferative wild-type cells contribute to CCM lesion growth and identify dysregulated cell mechanics as a “third hit” in CCM pathogenesis. Overall design: Wild type (WT) human umbilical vein endothelial cells (HUVECs) were labeled with Lifeact-RFP, while control (CT) or CCM2 siRNA silenced HUVECs were labeled with Lifeact-GFP. Red wild type HUVECs were combined in a 1:1 ratio with either siCT or siCCM2 HUVECs and allowed to invade overnight as mosaics into a PEG based hydrogel using our 3D-in vitro model of angiogenic sprouting. Hydrogels were then dissolved and the cells collected. WT cells were separated from the siRNA silenced HUVECs using Fluorescence activated cell sorting (FACs) according to the presence of lifeact-GFP or lifeact-RFP signal. The 3 samples collected for single cell RNA sequencing were WT cells invading in mosaics with siCT cells termed "WTinCT", WT cells invading in mosaics with siCCM2 cells termed "WTinCCM2", and siCCM2 cells from the mosaics termed "CCM2". The cells were collected from 4 independent experiments and in triplicates of hydrogels per trial. All collected and sorted cells were then pooled such that each sample (WTinCT, WTinCCM2, CCM2) consisted of the respective cell type from all trials and replicates.