Multi-scale Computational Model of Endothelial Cell-Pericyte Coupling in Idiopathic Pulmonary Fibrosis

INTRODUCTION: Microvascular stability is highly dependent on endothelial cell-pericyte coupling. In fibrotic diseases, such as idiopathic pulmonary fibrosis (IPF), extracellular matrix stiffening and elevated concentrations of profibrotic factors, such as TGF-β, can disrupt this cell-cell communication. Treatment of IPF with the antifibrotic drug Nintedanib, which targets PDGF-βR, FGF-R, and VEGF-R signaling, may counteract these effects to strengthen endothelial cell-pericyte coupling and promote microvessel stability1,2 . We have developed a multi-scale computational model that simulates intracellular signaling in endothelial cells and pericytes, heterotypic cell-cell communication, and the dynamic lung microenvironment to study how Nintedanib2 and other drugs affect microvascular network remodeling in IPF. We hypothesize that blocking PDGF-βR and FGF-R signaling with Nintedanib is sufficient to rescue TGF-β-induced endothelial cell-pericyte decoupling. MATERIALS AND METHODS: The multi-scale model is comprised of logic-based ordinary differential equations representing intracellular signaling networks in both endothelial cells and pericytes integrated into an agent-based model representing the lung environment wherein the simulated cells interact with one another, sensing and dynamically altering their microenvironment (Figure 1). The logic-based network signaling models were developed using the Netflux3 toolkit in MATLAB. The agent-based model of the spatiotemporal 2D lung environment was constructed in NetLogo4 . The multi-scale model was created by linking the logic-based network models with the agent-based model using the NL4Py5 package in Python. RESULTS AND DISCUSSION: Endothelial cell-pericyte decoupling was signified by decreased N-cadherin expression in pericytes, increased αSMA and Col1mRNA (indicating possible pericyte-to-myofibroblast transition), and increased physical separation and distances between endothelial cells and pericytes. The multi-scale model predicted that increasing TGF-β concentrations significantly elevated αSMA and Col1mRNA expression in simulated pericytes. N-cadherin levels did not change in response to TGF-β or Nintedanib treatment in the network model of the pericyte alone but were affected by the presence of endothelial cells in the agent-based model, highlighting the importance of endothelial cell-pericyte interactions in determining pericyte behaviors. CONCLUSIONS: Our multi-scale model demonstrates the importance of considering endothelial cell-pericyte communication in predicting cellular responses to pro-fibrotic environments. By representing intracellular biochemical signaling networks, heterotypic cell interactions, and the dynamic multi-cell microenvironment, we were able to demonstrate a potential mechanism through which Nintedanib treatment affects endothelial cell-pericyte coupling in IPF. REFERENCES: 1) Hanumegowda, C., et al.; Chest. 2012 2) Wollin, L., et al.; Eur Respir J. 2015 3) Kraeutler, MJ., et al.; BMC Syst Biol. 2010 4) Sklar, E., Artif. Life 2007 5) Gunaratne, C.; 2018. NL4Py. © FASEB.