(358b) 3-Dimensional Agent-Based Simulation Framework to Investigate the Effect of Scaffold Pore Structure On Angiogenesis

Modeling and simulation frameworks that are built upon sound biological knowledge and are utilizing novel computational strategies have become an inevitable tool for understanding the underlying mechanisms of biomedical applications, often leading to improved theoretical knowledge and enhanced experimental strategies. In this work, vascularization of biomaterial structures used as tissue engineering scaffolds has been investigated using an agent-based model (ABM) developed.

Tissue engineering scaffolds are used as a physical support structure and insoluble regulator of cellular activities in a wide range of biomedical applications, including production of functional implants for regenerative medicine. Microstructure of biomedical scaffolds is characterized by parameters such as porosity, average pore size, interconnectivity, and pore shape as well as mechanical properties such as Young’s modulus. Development of a healthy blood vessel network inside these scaffolds is essential for transferring oxygen and nutrients to the tissue cells growing inside scaffold structure, and has been a major limiting step in tissue regeneration, largely restricting clinical application of these scaffolds.

A multi-layer multi-agent computational framework is developed for modeling the process of sprouting angiogenesis (formation of new blood vessels from pre-existing blood vessels) within polymeric porous scaffolds. The framework is implemented in Java, using Repast (Recursive Porous Agent Simulation Toolkit) which is an open-source agent-based modeling and simulation platform. Software agents are developed as independent computational entities to represent endothelial cells (ECs), which are the cells lining the inside of blood vessel capillaries. These agents interact together and with their micro-environment, leading to formation of new blood vessel capillaries and invasion into deeper parts of scaffolds.

An embedded rule base governs the behavior of individual EC agents. ECs are capable of sensing their micro-environment to perceive the location of other neighboring agents and the geometry of surrounding scaffold, as well as concentration and gradients of the soluble and insoluble factors, and then performing various actions such as elongation, proliferation, and sprouting or branching. Each EC agent is identified by two spherical nodes located in space and attached to each other with a connection in the network layer of the model. The leading node is the active node which performs the actions of an EC. The other node represents the trailing node and is fixed in space. The length of each EC is dictated as the distance between its leading and trailing nodes. Using this abstraction method, elongation of an EC is simply exhibited by movement of its leading node. In each capillary branch, only the leading EC, referred to as tip cell, is active and capable of elongation and proliferation. Other ECs in each branch are referred to as stalk cells, and are only activated randomly during simulation to perform sprouting, if other conditions in their environment allows them to do so.

2D version of the developed model has previously been published and includes simulation case studies to investigate the effect of scaffold pore size on angiogenesis. Undertaken simulation results indicate that pores of larger size (160-270 µm) support rapid and extensive angiogenesis throughout the scaffold. 2D model predictions are compared to experimental in vivo results of vascularization in porous poly-ethylene glycol (PEG) hydrogels in order to validate the results. There is a good agreement between experimental and computational results, with model accuracy increasing with the pore size.

The simulation framework is then extended to 3D and is improved by addition of volume considerations for EC agents, enabling the framework to be well-suited for understanding the effect of scaffold geometry and steric hindrances on bioactivity of the 3D construct. To achieve this goal, 3D scaffold models with homogeneous and heterogeneous spherical pores are developed to investigate the impact of scaffold mean pore size and interconnectivity on rate of vessel growth and the developed blood vessel density. Different pore sizes can be used as model input and for each pore size, various pore interconnectivities are studied. Simulation results for homogeneous pore structures support the positive effect of increasing pore size and interconnectivity on angiogenesis. Simulation results also indicate that at early times (less than one week) different cases yield similar results, showing that the positive effect of larger pore sizes become considerable at longer times, when deeper invasion of blood vessels inside scaffold becomes important.

Controlling blood vessel assembly by modulating EC behavior with changes in the characteristics of the extracellular environment is critical to development of replacement tissues. The model predicts behavior of blood vessel networks growing inside a matrix towards regions with higher concentration of growth factors, as expected from a developing blood vessel network and comparable to experimental findings. The computational and experimental results can be used to guide the design of optimized porous scaffolds that regulate vascularization.