Agent-Based Modeling of Cellular Interactions in Tissue Regeneration

Abstract

Tissue regeneration emerges from the coordinated interactions of heterogeneous cell populations -- stem cells, progenitors, immune cells, fibroblasts, endothelial cells, and differentiated parenchymal cells -- whose individual behaviours and intercellular communications collectively produce the tissue-level dynamics of repair. Agent-based modelling (ABM) provides a natural computational framework for capturing this emergent complexity: individual cells are represented as autonomous agents with state-dependent behaviours and local interaction rules, and tissue-level outcomes emerge from the simulated collective behaviour. While ABM has been applied to specific tissue regeneration contexts, a comprehensive ABM framework validated across multiple cell types, interaction mechanisms, and tissue regeneration scenarios with systematic comparison to experimental data has been absent. This paper proposes the Cellular Interaction Tissue Regeneration ABM (CITRA) framework, a validated agent- based modelling system for simulating multi-cellular regeneration dynamics, comprising five modelling components: a cell agent library (CAL) defining regeneration-relevant cell types with calibrated behaviour rules; an intercellular signalling simulator (ISS) modelling paracrine and juxtacrine communication between agents; an extracellular matrix dynamics model (EMDM) tracking ECM deposition, degradation, and mechanical properties; a spatial tissue geometry engine (STGE) managing agent positions and tissue boundary conditions; and a regeneration emergence quantifier (REQ) measuring tissue-level regeneration metrics from simulation output. CITRA is evaluated across five tissue regeneration scenarios: skeletal muscle satellite cell-mediated repair, hepatic zonation-directed regeneration, intestinal crypt regeneration after radiation injury, dermal wound healing, and cartilage chondrocyte repopulation. CITRA simulations reproduce key experimental observations across all five scenarios with mean quantitative agreement r = 0.912 (SD = 0.024) against experimental time-course data. ISS knockout simulations predict cell signalling dependencies subsequently confirmed by experimental validation in 84.6% (11/13) of tested predictions. STGE spatial pattern metrics agree with experimental histological patterns with SSIM = 0.864 (SD = 0.032). The study contributes the CITRA specification, a validated multi-scenario ABM benchmark, and 13 computationally predicted intercellular signalling dependencies as experimental hypotheses.