Bone tissue engineering constructs (BTEC) combining natural resorbable osteoconductive scaffolds and mesenchymal stem cells (MSCs) have given promising results to repair critical size bone defect. Yet, results remain inconsistent. Adjonction of an osteoinductive factor to these BTEC, such as rh-BMP-2, to improve bone healing, seems to be a relevant strategy. However, currently supraphysiological dose of this protein are used and can lead to adverse effects such as inflammation, ectopic bone and/or bone cyst formation. Interestingly, in a preliminary study conducted in ectopic site in a murine model, a synergistic effect on bone formation was observed only when a low dose of rh-BMP-2 was associated with MSCs-seeded coral scaffolds but not with a high dose.

The objective of the study was then to evaluate a BTEC combining coral scaffold, MSCs and a low dose of rh-BMP-2 in a large animal model of clinical relevance.

Sixteen sheep were used for this study.

MSCs were isolated from an aspirate of bone marrow harvested from the iliac crest of each sheep receiving BTEC with MSCs, cultivated and seeded on Acroporacoral scaffolds one week before implantation.

Rh-BMP-2, used at two different doses (low dose: 68μg/defect and high dose: 680μg/defect), was diluted and absorbed on Acroporacoral scaffold one day before implantation.

Metatarsal segmental bone defects (25 mm) were made in the left metatarsal bone of the sheep, stabilized by plate fixation, and filled with Acroporacoral scaffolds loaded with either (i) MSCs and a low dose of rh-BMP-2 (Group 1;n=6), (ii) a low dose of rh-BMP-2 (Group 2;n=5), (iii) a high dose of rh-BMP-2 (Group 3;n=5). Standard radiographs were taken after each surgery and each month until sheep sacrifice, 4 months postoperatively. Bone healing and scaffold resorption were assessed by micro-computed-tomography (μCT) and histomorphometry. Results were compared to a historical control group in which coral scaffolds were loaded with MSCs.

Bone volumes (BV) evaluated by μCT and bone surfaces (BS) evaluated by histomorphometry did not differ between groups (BV: 1914±870, 1737±841, 1894±1028 and 1835±1342 mm³; BS: 25,41±14,25, 19,85±8,31, 25,54±16,98 and 26,08±22,52 %; groups 1, 2, 3 and control respectively); however, an higher bone union was observed in group 1 compared to the others (3, 1, 2 and 2 sheep with bone union in groups 1, 2, 3 and control respectively). No histological abnormalities were observed in any group. Coral resorption was almost complete in all specimens. No significant difference in coral volumes and coral surfaces was observed between groups. A trend towards a higher variability in coral resorption was noted in group 1 compared to the others.

There seems to be a benefit to associate low dose of rh-BMP-2 with MSCs-seeded coral scaffolds as this strategy allowed an increase of bone unions in our model. Yet, results remain inconsistent. Although, defective coupling between scaffold resorption and bone formation impaired bone healing in some animals, adjunction of rh-BMP-2 (even at low dose) to CSMs loaded construct is a promising strategy for bone tissue engineering.