Agent Based Model of Endothelial Cell and Pericyte Interactions During Angiogenesis in the Germinal Matrix

The germinal matrix is a highly vascularized region in the developing brain, which is susceptible to intraventricular hemorrhage (IVH) in premature infants1. Blood vessels of the germinal matrix have been shown to have less pericyte coverage than in surrounding regions of the brain1. Pericytes are known to provide important cues to endothelial cells that stabilize blood vessels, promote sprouting, and regulate vessel permeability. Thus, we hypothesize that a lack of pericyte coverage on blood vessels contributes to the risk for IVH in premature infants. To test this hypothesis, we developed an agent‐based model of endothelial cell and pericyte interactions in the germinal matrix to elucidate their role in IVH. Methods Here we extend our recently published model, implemented in NetLogo, of angiogenesis in the developing retina, which explores the role of pericytes in vascular pattering and the selection of sprouting tip cells through the Notch1‐DLL4 pathway2 (Fig. 1A). We implemented the new agent‐based model in Repast, a widely used Java based modeling software. Then, modifications were made to the model to more accurately represent the anatomy of the germinal matrix, cell phenotype and density, and source of VEGF. Outputs measured from the Repast model, including Notch, DLL4, VEGF receptor, and phosphorylated VEGF receptor levels, were then verified by implementing identical initial conditions with the NetLogo model. Results New functionalities in the Repast model allow endothelial cells and pericytes to be seeded along the bottom of the simulation, or as a cluster in the middle, with the VEGF gradient increasing from bottom‐to‐top or radially outward, respectively (Fig. 1B and 1C). Endothelial cells migrate along the VEGF gradient and are identified as tip cells based on Notch1‐DLL4 signaling. Pericytes migrate to maintain contact with endothelial cells. Model verification was done by comparing to simulations with the same initial conditions in the NetLogo model. Important parameters shared by both models follow similar trends over the time course of the simulation (Fig. 2A–D). Discussion We have implemented a model that simulates angiogenesis in the germinal matrix using Repast and verified the outputs using a previously published model of angiogenesis in the retina. This model will potentially serve as a predictive tool for elucidating how current clinical interventions, such as antenatal steroids, affect brain vascular development, and for identifying novel therapeutic interventions. Further development of the model will aim to incorporate PDGF‐B/PDGFR‐B signaling between pericytes and sprouting endothelial cells. Additionally, images acquired from the germinal matrix in the developing mouse brain stained with CD31 and NG2 to identify endothelial cells and pericytes, respectively, will be used to directly seed an initial vascular network in the model. Support or Funding Information UVA VTC Neuroscience Seed Fund This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .