Effects of osteogenic ambulatory mechanical stimulation on early stages of BMP-2 mediated bone repair

Abstract

Purpose: Mechanical loading of bone defects through rehabilitation is a promising approach to stimulate repair and reduce nonunion risk; however, little is known about how therapeutic mechanical stimuli modulate early-stage repair before mineralized bone formation. The objective of this study was to investigate the early effects of osteogenic loading on cytokine expression and angiogenesis during the first 3 weeks of BMP-2 mediated segmental bone defect repair.Materials and Methods: A rat model of BMP-2 mediated bone defect repair was subjected to an osteogenic mechanical loading protocol using ambulatory rehabilitation and a compliant, load-sharing fixator with an integrated implantable strain sensor. The effect of fixator load-sharing on local tissue strain, angiogenesis, and cytokine expression was evaluated.Results: Using sensor readings for local measurements of boundary conditions, finite element simulations showed strain became amplified in remaining soft tissue regions between 1 and 3 weeks (Week 3: load-sharing: -1.89 ± 0.35% and load-shielded: -1.38 ± 0.35% vs. Week 1: load-sharing: -1.54 ± 0.17%; load-shielded: -0.76 ± 0.06%). Multivariate analysis of cytokine arrays revealed that load-sharing significantly altered expression profiles in the defect tissue at 2 weeks compared to load-shielded defects. Specifically, loading reduced VEGF (p = 0.052) and increased CXCL5 (LIX) levels. Subsequently, vascular volume in loaded defects was reduced relative to load-shielded defects but similar to intact bone at 3 weeks. Endochondral bone repair was also observed histologically in loaded defects at 3 weeks.Conclusions: Together, these results demonstrate that moderate ambulatory strains previously shown to stimulate bone regeneration significantly alter early angiogenic and cytokine signaling and may promote endochondral ossification.

Keywords: Bone repair; angiogenesis word; immune response; mechanobiology; rehabilitation.

Figure 1.

Representative image-based finite element model cross-sections demonstrate elevated strain magnitudes within UHMWPE-stabilized defects at 1 week. Regions of elevated strain localized to non-mineralized soft tissue regions at 3 weeks.

Figure 2.

(A) Mean 3rd principal strain of non-mineralized soft tissue within the defect increased between 1 and 3 weeks due to the increased proportion of much stiffer woven bone as healing progressed. n=5. *p<0.05 Wk 1 vs. 3, #p=0.077 UHMWPE vs. PSU via Two-way RM ANOVA. Soft tissue strain distribution throughout the defect within each of 5 experimental samples was spatially heterogeneous at (B) 1 and (C) 3 weeks. In all box plots, the mean is denoted by a dot and whiskers are defined using the Tukey method. (D) Mean 3rd principal strain magnitude within immature woven bone at 3 weeks averaged approximately 0.5% regardless of fixator stiffness. n=5. n/s via t-test. (E) Woven bone tissue strain was similarly heterogeneous throughout the defect in each experimental sample. (F) Peak femoral loads were variable between animals, averaging approximately four-fold body weight (BW) regardless of fixator stiffness or time point. n=5. n/s via Two-way RM ANOVA.

Figure 3.

Discriminant partial least squares regression (D-PLSR) analysis of the expression of 16 cytokines in the defect tissue at 2 weeks. (A) Latent variable 1 (LV1) defines a multivariate cytokine expression profile depicted along the x-axis that separates defects stabilized by load-shielding PSU fixators to the right and load-sharing UHMWPE fixators to the left. (B) Mean LV1 score was significantly upregulated by load-shielding PSU fixation. n=9-11. ***p<0.001 via t-test. (C) Values along LV1 describe individual cytokines with elevated expression for UHMWPE fixation (negative values) or PSU (positive values). Error bars were computed using leave-one-out cross validation (mean ± SD).

Figure 4.

Univariate comparisons of individual cytokine expression in defect tissue at 2 weeks. LIX (CXCL5) expression was significantly increased by increased load sharing with load-sharing UHMWPE fixation. Conversely, VEGF expression was elevated with by load shielding with load-shielding PSU fixation. Significant pairwise differences were not observed in the remaining cytokines. n=9-11. **p<0.01 via t-test.

Figure 5.

MicroCT angiography quantification of vascular volume at 3 weeks in and surrounding the bone defects. (A) Vascular volume within the central 5 mm of the defect is increased under load-shielding fixation, though load-sharing fixation remained similar to the intact femora. n=6-10. *p<0.05 via Kruskal-Wallis test with Dunn’s pairwise comparisons (B) Vascular volume throughout the defect and surrounding tissue was elevated in injured femora relative to naïve, irrespective of fixation stiffness. n=6-10. *p<0.05 via ANOVA with Tukey’s test. (C) In the surrounding tissue alone, vascular volume was similarly elevated in injured hindlimbs regardless of fixator. n=6-10. *p<0.05 via ANOVA with Tukey’s test.

Figure 6.

Bone defect histology at 2 and 3 weeks, including Hematoxylin & Eosin (left column for PSU and UHMWPE, except for bottom image), Safranin0O/Fast Green (right columns), and Picro Sirius red (bottom image in left columns), n=1 femur per group. Alginate hydrogel remnants are denoted as “Al”, bone is denoted as “B”, chondrocytes are denoted as “Ch”, and blood vessels are indicated with arrows extending from “V”. Dotted lines at sides of images denote intact diaphyseal bone stumps. Insets are registered to full defects by color-coded dotted rectangles. Black scale bar = 500 μm. White scale bar = 50 μm.