Hydrogel-based Delivery of rhBMP-2 Improves Healing of Large Bone Defects Compared With Autograft

Abstract

Background: Autologous bone grafting remains the gold standard in the treatment of large bone defects but is limited by tissue availability and donor site morbidity. Recombinant human bone morphogenetic protein-2 (rhBMP-2), delivered with a collagen sponge, is clinically used to treat large bone defects and complications such as delayed healing or nonunion. For the same dose of rhBMP-2, we have shown that a hybrid nanofiber mesh-alginate (NMA-rhBMP-2) delivery system provides longer-term release and increases functional bone regeneration in critically sized rat femoral bone defects compared with a collagen sponge. However, no comparisons of healing efficiencies have been made thus far between this hybrid delivery system and the gold standard of using autograft.

Questions/purposes: We compared the efficacy of the NMA-rhBMP-2 hybrid delivery system to morselized autograft and hypothesized that the functional regeneration of large bone defects observed with sustained BMP delivery would be at least comparable to autograft treatment as measured by total bone volume and ex vivo mechanical properties.

Methods: Bilateral critically sized femoral bone defects in rats were treated with either live autograft or with the NMA-rhBMP-2 hybrid delivery system such that each animal received one treatment per leg. Healing was monitored by radiography and histology at 2, 4, 8, and 12 weeks. Defects were evaluated for bone formation by longitudinal micro-CT scans over 12 weeks (n = 14 per group). The bone volume, bone density, and the total new bone formed beyond 2 weeks within the defect were calculated from micro-CT reconstructions and values compared for the 2-, 4-, 8-, and 12-week scans within and across the two treatment groups. Two animals were used for bone labeling with subcutaneously injected dyes at 4, 8, and 12 weeks followed by histology at 12 weeks to identify incremental new bone formation. Functional recovery was measured by ex vivo biomechanical testing (n = 9 per group). Maximum torque and torsional stiffness calculated from torsion testing of the femurs at 12 weeks were compared between the two groups.

Results: The NMA-rhBMP-2 hybrid delivery system resulted in greater bone formation and improved biomechanical properties compared with autograft at 12 weeks. Comparing new bone volume within each group, the NMA-rhBMP-2-treated group had higher volume (p < 0.001) at 12 weeks (72.59 ± 18.34 mm(3)) compared with 8 weeks (54.90 ± 16.14) and 4 weeks (14.22 ± 9.59). The new bone volume was also higher at 8 weeks compared with 4 weeks (p < 0.001). The autograft group showed higher (p <0.05) new bone volume at 8 weeks (11.19 ± 8.59 mm(3)) and 12 weeks (14.64 ± 10.36) compared with 4 weeks (5.15 ± 4.90). Between groups, the NMA-rhBMP-2-treated group had higher (p < 0.001) new bone volume than the autograft group at both 8 and 12 weeks. Local mineralized matrix density in the NMA-rhBMP-2-treated group was lower than that of the autograft group at all time points (p < 0.001). Presence of nuclei within the lacunae of the autograft and early appositional bone formation seen in representative histology sections suggested that the bone grafts remained viable and were functionally engrafted within the defect. The bone label distribution from representative sections also revealed more diffuse mineralization in the defect in the NMA-rhBMP-2-treated group, whereas more localized distribution of new mineral was seen at the edges of the graft pieces in the autograft group. The NMA-rhBMP-2-treated group also revealed higher torsional stiffness (0.042 ± 0.019 versus 0.020 ± 0.022 N-m/°; p = 0.037) and higher maximum torque (0.270 ± 0.108 versus 0.125 ± 0.137 N-m; p = 0.024) compared with autograft.

Conclusions: The NMA-rhBMP-2 hybrid delivery system improved bone formation and restoration of biomechanical function of rat segmental bone defects compared with autograft treatment.

Clinical relevance: Delivery systems that allow prolonged availability of BMP may provide an effective clinical alternative to autograft treatment for repair of segmental bone defects. Future studies in a large animal model comparing mixed cortical-trabecular autograft and the NMA-rhBMP-2 hybrid delivery system are the next step toward clinical translation of this approach.

Fig. 1A–C

(A) Rat femur with polysulfone stabilization plates showing the middiaphyseal defect, polysulfone fixation plate, and poly-caprolactone (PCL) nanofiber (NFM) mesh tube fitting around the proximal and distal stumps to enclose the defect space. (B) Three-dimensional reconstruction of the micro-CT image of an autograft treated sample showing the relative sizes of the cortical bone fragments. (C) A schematic showing the experiment design including interventions, outcome measures, and sample numbers.

Fig. 2A–B

Representative radiographs (A) and micro-CT reconstructions (B) show a defect bridging over time in the two treatment groups. The autograft showed radiodense shadows around bone fragments as early as 2 weeks suggesting early mineralization. The NMA-rhBMP-2 group only showed limited bone formation at the early time points (2, 4 weeks). The analyzed midsection of the defects is shown with an overlay of the bone mineral density map (color map scale: 527–1237 mg HA/cm³).

Fig. 3A–B

Comparison of BV between the two groups with time is shown. (A) Total BV increased significantly in the NMA-rhBMP-2 group with time. This increase was much less marked in the autograft group. The dotted line represents the mean BV of autograft group at 8 weeks. The NMA-rhBMP-2 BV was significantly lower than the autograft group earlier than 8 weeks and was significantly higher subsequently (p < 0.05). (B) Comparing only the new bone formed beyond 2 weeks (subtraction of 2-week values) revealed the significantly lower new bone formation in the autograft group compared with the NMA-rhBMP-2 group at 8 and 12 weeks (a, b: p < 0.001). Within each group, however, the new BV rose significantly over time.

Fig. 4

Comparison of bone mineral density (BMD) between groups with time. Bone density in the NMA-rhBMP-2 group was always lower than that of autograft (p < 0.001). The mineral density increased significantly with time within both groups (p < 0.05) but was less marked in the autografts (Autograft: c, d = higher than the preceding time points within the group, NMA-rhBMP-2: all increments with time, p < 0.001) (mean ± SD; n = 14).

Fig. 5A–J

Representative nondecalcified cryosections (5 μm) at 12 weeks were processed for routine H&E staining (A–B), vital bone labels (green = Week 4; red = Week 8) (C–D), mineral content (E–F), AP (osteoblastic activity) (G–H), and DAPI (nuclear) on the same sections. Images I and J are overlay images of the regions identified in A and B. The NMA-rhBMP-2 group showed diffuse mineral deposition in the defect at 4 weeks with evidence of continued activity (A, C, E, I [region of interest denoted by *]). Limited AP activity was also diffusely distributed in the defect and appeared at the mineralizing edge of the bone (G). The autograft group showed a more localized mineral deposition at the graft edges and focal areas within the grafts at 4 weeks followed by predominantly edge deposition at 8 weeks (B, D, F, J [region of interest denoted by ^{^}]). Very limited AP activity was noted on some autograft margins and the defect boundary (H). (Composite images show entire defect in both cases [A–H]; magnification × 5).

Fig. 6A–P

Representative decalcified paraffin sections (5 μm) of NMA-rhBMP-2 samples at 2 to 12 weeks were processed for routine H&E staining (A–D), Safranin-O/Fast green (E–H), Mallory’s modified aniline blue stain (I–L), and Picrosirius red (M–P) staining to characterize bone regeneration. Dotted lines in H&E images indicate areas of higher magnification shown in Safranin-O or Mallory’s stain images, whereas Picrosirius stain images are from nonoverlapping areas. Bone appeared pink in H&E sections (A–D), whereas alginate and partly mineralized areas appeared pink-purple. Cartilage and alginate stained red with Safranin-O (E–H) and blue with Mallory’s aniline stain (I–L). Additionally, mature bone stained deep red, whereas newly formed bone stained orange-red with Mallory’s stain. Organized collagen structures, typical of mature bone, appeared brightly birefringent orange-red, yellow, or green with Picrosirius red stain under polarized light (M–P). Limited early bone formation was evident at 2 weeks (region of interest denoted by *) seen as pink areas (A), blue with Safranin-O (E), and orange-yellow with Mallory’s (I, region of interest denoted by “2”) and limited red staining areas with Picrosirius stain (M). Alginate appeared red in the Safranin-O-stained image and blue in Mallory’s (I, region of interest denoted by “1”). Progressively more bone deposition was evident with time appearing as better-defined pink regions in H&E-stained sections (A–D). Corresponding sections stained with Safranin-O (E–H) showed blue areas of bone and red areas of remaining alginate. Mallory’s stain with time (I–L) revealed an increase in deep red staining showing larger regions of mature bone (regions of interest denoted by ^{^} or “3”) with less of poorly mineralized tissue–orange stain (J region of interest denoted by “4”) and less residual alginate (1). Picrosirius red staining revealed increase in areas of birefringence with time suggestive of more fibrolamellar bone (regions of interest denoted by “5, 6” in N–P) (Scale: A–D: 300 μm, E–P: 150 μm).

Fig. 7A–P

Representative decalcified paraffin sections (5 μm) of the autograft group at 2 to 12 weeks were processed for routine H&E staining (A–D), Safranin-O (E–H), Mallory’s aniline blue stain (I–L), and Picrosirius red (M–P) staining to characterize bone regeneration as with the NMA-rhBMP-2 group. The dotted lines in H&E images indicate areas magnified for corresponding image areas in Safranin-O and Mallory’s aniline-stained sections (except K). Picrosirius red images are of nonoverlapping representative areas. The autograft group did not reveal clear evidence of appositional bone formation at 2 weeks by H&E (A, 1). However, small focal areas showing a mix of endochondral and fibrolamellar appositional bone formation were visible on autograft surfaces (A, *). Safranin-O staining did not reveal clear cartilage areas (E, *), but blue staining regions were evident with Mallory’s stain with areas of new immature bone deposition (I, *). Picrosirius stain confirmed this early mineralization near the bone stump with minimal activity in the autograft pieces (M, 1). At 4 weeks, cartilage (B, ^#) was clearly visible in a focal region between autograft fragments (B, 1) and more mature appositional bone at the autograft margins (B, 2). Red staining of corresponding areas of cartilage with Safranin-O (F, ^#) and blue cartilage (^#) and red adjacent bone in Mallory’s stain confirmed the presence of cartilage and more mature appositional bone, respectively (J). Picrosirius stain revealed organized fiber structures with stark birefringence (N, 2) at the autograft edges (N, 1). At 8 weeks, appositional mineralized connections (C, ^{^}) between autograft pieces were clearly visible (C, 1). This area did not stain for cartilage (G, ^{^}), but core areas of red staining suggestive of cartilage were visible in the autografts with Safranin-O stain (G). In other areas, varying amounts of new bone mainly on the surface of autograft-appositional bone were visible (K, ^{^}) without staining for cartilage tissue. Picrosirius staining confirmed this surface localization with brightly birefringent areas (O, 2) along autograft (O, 1) margins. At 12 weeks, the autograft pieces had progressed to bony union between autografts (^{^} in D, H, L) with Picrosirius staining showing typical organized structures in red-green birefringence (P) (Scale: A, B, C = 300 μm; D, H, L = 600 μm; others = 150 μm).

Fig. 8A–B

Evaluation of functional regeneration as indicated by biomechanical properties is demonstrated. (A) Torsional testing showed a higher torsional stiffness in the NMA-rhBMP-2 group compared with the autograft group (*p = 0.037). (B) Maximum torque at failure was also higher for the NMA-rhBMP-2 group compared with autograft (*p = 0.02). The dotted line on both graphs represents the mean value for intact bone from literature [28, 34] (mean ± SD; n = 9).