Computational model of the in vivo development of a tissue engineered vein from an implanted polymeric construct

Abstract

Advances in vascular tissue engineering have been tremendous over the past 15 years, yet there remains a need to optimize current constructs to achieve vessels having true growth potential. Toward this end, it has been suggested that computational models may help hasten this process by enabling time-efficient parametric studies that can reduce the experimental search space. In this paper, we present a first generation computational model for describing the in vivo development of a tissue engineered vein from an implanted polymeric scaffold. The model was motivated by our recent data on the evolution of mechanical properties and microstructural composition over 24 weeks in a mouse inferior vena cava interposition graft. It is shown that these data can be captured well by including both an early inflammatory-mediated and a subsequent mechano-mediated production of extracellular matrix. There remains a pressing need, however, for more data to inform the development of next generation models, particularly the precise transition from the inflammatory to the mechanobiological dominated production of matrix having functional capability.

Keywords: Constrained mixture theory; Inflammation; Interposition graft; Mechanosensing; Mouse model; Tissue engineering.

Fig. 1

Mass density production is shown for each constituent (collagen families 1,2,3,4, and circumferential smooth muscle) throughout the G&R simulation (168 days). (Panel A) Inflammation-mediated production peaked at 10 days (reflecting the time of highest measured macrophage infiltration) with a 45-fold increase relative to basal production for all constituents, then decreased following loss of the polymeric scaffold (by 8 weeks). Increases in inflammation-mediated production were isotropic (i.e., consistent fold increases relative to basal production for all families of collagen fibers). (Panel B) Stress-mediated production increased following the loss of structural integrity of the polymeric scaffold around 4 weeks. Production increases were anisotropic and remained elevated relative to basal for the remainder of the simulation. (Panel C) Total increases in production relative to basal rates for all constituents. Note the change of scale from previous panels to account for additive effects of all contributing constituents. These results suggest a steady state, but not optimal (relative to native vein), long-term turnover.

Fig. 2

Time-course of the survival fractions, which range over [0,1], for sample constituents produced at some time tau during the simulation for the inflammation- and stress-mediated mechanisms of production. For the inflammation-mediated collagen (dashed line), β_infl = 2, which resulted in a shorter half-life (smaller survival fraction) compared to the stress-mediated collagen (solid line).

Fig. 3

Circumferential tension-stretch relationships for the native SCID/bg inferior vena cava (Panel A), simulated polymeric scaffold at time 1 day (Panel B), and both predicted (solid curves) and experimental data (circles) for TEVGS at 2 (Panel C), 6 (Panel D), 12 (Panel E), and 24 weeks (Panel F). Note that the mechanical behavior (predicted and experimental) evolved from a stiff response dominated by polymer and inflammation-mediated collagen (Panel C) to a compliant more vascular-type response due primarily to collagen produced via the stress-mediated phase (Panel F). Note the similarity in the overall predicted behaviors and the experimental results.

Fig. 4

Pressure-normalized diameter relationships are shown for the native SCID/bg inferior vena cava (Panel A), simulated polymeric scaffold at 1 day (Panel B), and both predicted (solid waves) and experimental data (circles) for TEVGs at 2 (Panel C), 6 (Panel D), 12 (Panel E), and 24 weeks (Panel F). Note the transition from a stiff polymer-dominated response (Panel C) to a more compliant, vascular type response (Panel F) as well as the agreement between the predictions and experimental data.

Fig. 5

Circumferential tension-stretch relationships at 2 (Panel A), 6 (Panel B), 12 (Panel C), and 24 weeks (Panel D) for indicated values of β_infl for collagen produced via the inflammation-mediated response. Decreasing the half-life of collagen (increasing β_infl) resulted in an increase in TEVG compliance over time. A value of β_infl = 2 yielded the best predictions with respect to the experimental data.