Addition of heparin binding sites strongly increases the bone forming capabilities of BMP9 in vivo

Abstract

Bone Morphogenetic proteins (BMPs) like BMP2 and BMP7 have shown great potential in the treatment of severe bone defects. In recent in vitro studies, BMP9 revealed the highest osteogenic potential compared to other BMPs, possibly due to its unique signaling pathways that differs from other osteogenic BMPs. However, in vivo the bone forming capacity of BMP9-adsorbed scaffolds is not superior to BMP2 or BMP7. In silico analysis of the BMP9 protein sequence revealed that BMP9, in contrast to other osteogenic BMPs such as BMP2, completely lacks so-called heparin binding motifs that enable extracellular matrix (ECM) interactions which in general might be essential for the BMPs' osteogenic function. Therefore, we genetically engineered a new BMP9 variant by adding BMP2-derived heparin binding motifs to the N-terminal segment of BMP9's mature part. The resulting protein (BMP9 HB) showed higher heparin binding affinity than BMP2, similar osteogenic activity in vitro and comparable binding affinities to BMPR-II and ALK1 compared to BMP9. However, remarkable differences were observed when BMP9 HB was adsorbed to collagen scaffolds and implanted subcutaneously in the dorsum of rats, showing a consistent and significant increase in bone volume and density compared to BMP2 and BMP9. Even at 10-fold lower BMP9 HB doses bone tissue formation was observed. This innovative approach of significantly enhancing the osteogenic properties of BMP9 simply by addition of ECM binding motifs, could constitute a valuable replacement to the commonly used BMPs. The possibility to use lower protein doses demonstrates BMP9 HB's high translational potential.

Keywords: Bone morphogenetic protein 9 (BMP9); Bone regeneration; Heparin binding sites; Subcutaneous animal model.

Graphical abstract

Fig. 1

Biological activities of BMP9 HB compared to the BMP2 WT and BMP9 WT. (A) Coomassie Brilliant Blue staining of BMP9 WT and BMP9 HB separated by SDS-PAGE under reducing and non-reducing conditions; (B) Dose-dependent ALP activity in C2C12 cells induced by the indicated ligands. The calculated EC₅₀ values are: BMP9 WT: 0.87 nM ± 0.21 nM, BMP9 HB: 2.4 nM ± 0.85 nM and BMP2 WT 22 nM ± 9.3 nM). The background absorption (without ligand, OD_405nm: 0.045) was subtracted from all recorded values. (C) Selected surface plasmon resonance (SPR) sensorgrams depict the interactions of the indicated ligands with heparin at 200 nM ligand concentration. Selected SPR sensorgrams depict the interactions of the indicated ligands with (D) BMPR-IA_ECD or BMPR-II_ECD or (E) with ALK1_ECD at 200 nM ligand concentration and (F) with Noggin at 62.5 nM ligand concentration. Apparent K_D values (presented in text) were calculated as described in the material and methods section. (G) ALP activity in C2C12 cells was induced by the indicated ligands at concentration reflecting their individual EC₅₀ values with or without the presence of 50 nM Noggin.

Fig. 2

In vitro BMP9 WT and BMP9 HB release. (A) SEM of the collagen scaffold processed in 1 wt% 1,4-butanediol di-glycidyl ether bis-epoxy carbonate (BDDGE) buffer at room temperature. Scale bar 100 μm (B) The cumulative release of BMP9 WT (green) and BMP9 HB (black) in alpha MEM with 1% Fungizone/gentamicin detected by ELISA is demonstrated over 7 days. Data represent mean values ± Standard Error of Mean. Statistical significance was validated by a non-parametric test, p = 0.0002.

Fig. 3

Histological analyses of bone formed 3 weeks post implantation. Representative histological images (Masson's trichrome staining) of implants retrieved 3 weeks post implantation: A. and a. 1 μg BMP9 HB, B. 1 μg BMP9 WT, C. and c. 10 μg BMP9 HB, D. 10 μg BMP9 WT. Scale bar: 250 μm in A, B, C and D and 100 μm in a and c. White arrows indicate blood vessel, black arrows with dashed line show remnants of collagen scaffold, green arrows indicate newly formed bone which is also shown at a higher magnification in a and c.

Fig. 4

Longitudinal evaluation of bone volume (BV) and bone mineral density (BMD). (A, B, C) Comparison of the mineral volume obtained in the different conditions at 3 (A), 6 (B), and 8 (C) weeks‐periods by longitudinal imaging. (D, E, F) Comparison of bone mineral density at 3 weeks (D), 6 weeks (E) and 8 weeks (F). All the animals are included in the graphs (also values equal to 0). The table below the graph shows the number of subcutaneous mineralized tissues detected (values different from 0 were counted as detected bone) over a n = 6. Data represent mean values ± Standard Deviation. Statistical analyses were performed by one-way ANOVA by Dunn's multiple comparisons test.

Fig. 5

Evaluation of bone volume (BV) and histological examination of the implants retrieved 8 weeks post implantation. (A) The bone volume of the retrieved ossicles at 8 weeks is shown as bar diagram. Black symbols represent retrieved implants from the first batch of animals and red symbols represent the ones from the second batch (n = 6 for the conditions: 0.1 μg BMP9 HB, 1 μg BMP2 WT, 1 μg BMP9 WT, 10 μg BMP2 WT. n = 10 for the conditions: 1 μg BMP9 HB, 10 μg BMP9 WT and 10 μg BMP9 HB). Data represent mean values ± Standard Deviation. Statistical analyses were performed by one-way ANOVA by Dunn's multiple comparisons test, p < 0.01. (B–H) Representative Masson's trichrome-stained sections of implants at 8 weeks. (B_i and B_ii) 0.1 μg BMP9 HB, (C, D and E) 1 μg of BMP9 HB, BMP9 WT, BMP2 WT and (F, G and H) 10 μg BMP9 HB, BMP9 WT, BMP2 WT loaded collagen scaffold. Traces of the collagen scaffold (S), bone tissue (B) and fatty bone marrow (BM) are indicated. Scale bars in C, D, E and F, G, H: 2.5 mm. Scale bars in B_i 500 μm and in all the enlargements in C, D, E, F, G, H: 250 μm. Scale bar in B_ii 100 μm.