An Agent-Based Discrete Collagen Fiber Network Model of Dynamic Traction Force-Induced Remodeling
Authored by James W Reinhardt, Keith J Gooch
Date Published: 2018
DOI: 10.1115/1.4037947
Sponsors:
No sponsors listed
Platforms:
NetLogo
Model Documentation:
Other Narrative
Model Code URLs:
https://drive.google.com/file/d/0B0E2WdML7GEmNC1Fejh1emNONlk/view
Abstract
Microstructural properties of extracellular matrix (ECM) promote cell
and tissue homeostasis as well as contribute to the formation and
progression of disease. In order to understand how microstructural
properties influence the mechanical properties and traction
force-induced remodeling of ECM, we developed an agent-based model that
incorporates repetitively applied traction force within a discrete fiber
network. An important difference between our model and similar finite
element models is that by implementing more biologically realistic
dynamic traction, we can explore a greater range of matrix remodeling.
Here, we validated our model by reproducing qualitative trends observed
in three sets of experimental data reported by others: tensile and shear
testing of cell-free collagen gels, collagen remodeling around a single
isolated cell, and collagen remodeling between pairs of cells. In
response to tensile and shear strain, simulated acellular networks with
straight fibrils exhibited biphasic stress-strain curves indicative of
strain-stiffening. When fibril curvature was introduced, stress-strain
curves shifted to the right, delaying the onset of strain-stiffening.
Our data support the notion that strain-stiffening might occur as
individual fibrils successively align along the axis of strain and
become engaged in tension. In simulations with a single, contractile
cell, peak collagen displacement occurred closest to the cell and
decreased with increasing distance. In simulations with two cells,
compaction of collagen between cells appeared inversely related to the
initial distance between cells. These results for cell-populated
collagen networks match in vitro findings. A demonstrable benefit of
modeling is that it allows for further analysis not feasible with
experimentation. Within two-cell simulations, strain energy within the
collagen network measured from the final state was relatively uniform
around the outer surface of cells separated by 250 mu m, but became
increasingly nonuniform as the distance between cells decreased. For
cells separated by 75 and 100 mu m, strain energy peaked in the
direction toward the other cell in the region in which fibrils become
highly aligned and reached a minimum adjacent to this region, not on the
opposite side of the cell as might be expected. This pattern of strain
energy was partly attributable to the pattern of collagen compaction,
but was still present when mapping strain energy divided by collagen
density. Findings like these are of interest because fibril alignment,
density, and strain energy may each contribute to contact guidance
during tissue morphogenesis.
Tags
Agent-based modeling
remodeling
mechanics
transmission
elasticity
Organization
Tissue
Discrete fiber network simulation
Collagen
network
Mechanical properties
Fibril alignment
Matrix stiffness
Human-fibroblasts
Cell behavior
Gels
Tension