Bone morphogenetic protein 2-induced cellular chemotaxis drives tissue patterning during critical-sized bone defect healing: an in silico study

Abstract

Critical-sized bone defects are critical healing conditions that, if left untreated, often lead to non-unions. To reduce the risk, critical-sized bone defects are often treated with recombinant human BMP-2. Although enhanced bone tissue formation is observed when BMP-2 is administered locally to the defect, spatial and temporal distribution of callus tissue often differs from that found during regular bone healing or in defects treated differently. How this altered tissue patterning due to BMP-2 treatment is linked to mechano-biological principles at the cellular scale remains largely unknown. In this study, the mechano-biological regulation of BMP-2-treated critical-sized bone defect healing was investigated using a multiphysics multiscale in silico approach. Finite element and agent-based modeling techniques were combined to simulate healing within a critical-sized bone defect (5 mm) in a rat femur. Computer model predictions were compared to in vivo microCT data outcome of bone tissue patterning at 2, 4, and 6 weeks postoperation. In vivo, BMP-2 treatment led to complete healing through periosteal bone bridging already after 2 weeks postoperation. Computer model simulations showed that the BMP-2 specific tissue patterning can be explained by the migration of mesenchymal stromal cells to regions with a specific concentration of BMP-2 (chemotaxis). This study shows how computational modeling can help us to further understand the mechanisms behind treatment effects on compromised healing conditions as well as to optimize future treatment strategies.

Keywords: Agent-based model; Bone defect healing; Bone morphogenetic protein 2; Finite element analysis; Mechanobiology.

Fig. 1

Overview of the different models used to simulate BMP-2-stimulated bone defect healing in multiple scales. Top: graphical representation of the different in silico 3D models. Bottom: flowchart of the relationship between the different models

Fig. 2

BMP-2 chemotactic effect on MSC and osteoblast migration (top), enhancement of MSC proliferation capacity (left bottom), and osteoblast bone tissue production (right bottom). Adapted from Ribeiro et al. (2015)

Fig. 3

Bone tissue patterning of critical-sized defect healing under untreated conditions (control: no external mechanical stimulation, no BMP-2; and only-load: external in vivo mechanical stimulation, no BMP-2) cases at 2, 4 and 6 weeks postoperation. Comparison between in vivo µCT images (Schwarz et al. 2013) (A) and in silico predictions under continuous (B) and limited (C) cellular recruitment conditions

Fig. 4

Bone tissue volume comparison between in silico (average) and in vivo (BV average ± SD) within the healing region at 2, 4, and 6 weeks postoperation for untreated case scenarios (control and only-load). Note: “ex-” prefix in legend identifies in vivo

Fig. 5

In silico predicted dynamics of BMP-2 concentration within the callus growth region (A) and its effects on the chemotactic attraction of MSCs (B), on enhancing MSC proliferation (C) and bone tissue production (D) at 3, 5, 7 and 10 days postoperation. The results refer to the simulation of a BMP-2 treatment instantly released. Note: The color scale is logarithmic for BMP-2 concentration plots

Fig. 6

Bone tissue patterning of critical-sized defect healing under the only-BMP-2 condition (no external mechanical stimulation, exogenous BMP-2) at 2, 4, and 6 weeks postoperation when the gradual release of the growth factor is not implemented in the in silico model (right). Under this condition, the model is not able to reproduce BMP-2-treated defect healing as observed in the pre-clinical study (left) (Schwarz et al. 2013)

Fig. 7

In silico predicted dynamics of BMP-2 concentration within the callus growth region (A) and its effects on the chemotactic attraction of MSCs (B), on enhancing MSC proliferation (C) and bone tissue production (D) at 1, 2, 4, and 6 weeks postoperation. The results refer to BMP-2 treatment when a collagen sponge where a gradual release was simulated. Note: The color scale is logarithmic for BMP-2 concentration plots

Fig. 8

Bone tissue patterning of critical-sized defect healing under treated conditions (only-BMP-2 and BMP-2 + load cases) at 2, 4, and 6 weeks postoperation. Comparison between µCT images (Schwarz et al. 2013) (A) and in silico predictions (B)

Fig. 9

Mineralized callus volume comparison between in silico (average) and in vivo (BV average ± SD) within the healing region at 2, 4, and 6 weeks postoperation for BMP-2 treated case scenarios (only BMP-2 and BMP-2 + load). Note: “ex-” prefix in legend identifies in vivo