Bone lesions above a critical size become scarred rather than regenerated, leading to nonunion. We have attempted to obtain a greater degree of regeneration by using a resorbable scaffold with regeneration-competent cells to recreate an embryonic environment in injured adult tissues, and thus improve clinical outcome. We have used a combination of a coral scaffold with in vitro-expanded marrow stromal cells (MSC) to increase osteogenesis more than that obtained with the scaffold alone or the scaffold plus fresh bone marrow. The efficiency of the various combinations was assessed in a large segmental defect model in sheep. The tissue-engineered artificial bone underwent morphogenesis leading to complete recorticalization and the formation of a medullary canal with mature lamellar cortical bone in the most favorable cases. Clinical union never occurred when the defects were left empty or filled with the scaffold alone. In contrast, clinical union was obtained in three out of seven operated limbs when the defects were filled with the tissue-engineered bone.

Standardized particulate bone constructs, obtained by expanding autologous mesenchymal stem cells (MSCs) onto coral granules in vitro, were transplanted into long-bone, critical-size defects in sheep. Control experiments were also performed in which autologous bone grafts were implanted. Defect cavities were lined with a preformed vascularized membrane (induced by temporarily inserting a cement spacer for 6 weeks prior to bone construct implantation), which served as a mold keeping the engineered bone granules in place. Radiographic, histological, and computed tomographic tests performed 6 months later showed that the osteogenic abilities of the engineered construct and autograft were significantly greater than those of coral scaffold alone. No significant differences were found between the amount of newly formed bone in defects filled with coral/MSCs and those filled with autograft, yet radiological scores differed significantly between the two groups (21% and 100% healed cortices, respectively). The present study on a clinically relevant animal model provides the first evidence that standardized particulate bone constructs can be used to repair large bone defects and that their osteogenic ability approaches that of bone autograft, the bone repair benchmark. By proving feasibility, the present study makes possible the treatment of segmental bone losses with bone constructs engineered from granules, a process which is much simpler than preparing customized massive constructs using computer-assisted techniques. Important parameters, such as the rate of scaffold resorption and the number of MSCs to be seeded on the scaffolds, need to be optimized before reaching pertinent definitive conclusions. ß

Experiments have been performed to investigate the use of coral skeletons as bone graft substitutes. Skeletal fragments of different coral genera were implanted into cortical and spongy bone defects and used to bridge transcortical resections in the femur. The implant site was monitored for up to 18 months. Radiographically, both cortical and spongy bone defects were at least partially filled by new bone after 8 weeks while the implants underwent continuous resorption. Coral resorption and replacement by new tissue was also observed in the transcortical resections. The process of resorption was attributed to the enzymatic attack, especially carboanhydrase. This was confirmed by experiments in which dogs were implanted with coral in transcortical resections and treated daily with acetazolamide, a carboanhydrase inhibitor; the absorption appeared delayed and the resections failed to heal.

Previous studies showed that natural coral implanted into bone tissue was gradually resorbed and progressively replaced by newly formed bone. The objectives of this study were to compare the fate of two Madreporian corals, Porites and Acropora, after implantation during 1 and 2 months into sheep and pig long bones. These materials are identical in composition (CaCo3) but differ in volume (49 +/- 2%, 12 +/- 4%, respectively) and mean size (250 vs. 500 microns) of porosities. The non-decalcified histological slices were observed under light microscopy. Implant resorption and new bone formation were quantified through an automatic image analysis system. Quantitative results showed that the larger the porosity volume, the greater was the coral resorption as well as the new bone apposition. Large differences were found between the two animal species. Histological findings were identical to those previously reported: implants were resorbed and progressively replaced by newly formed bone. Coral was found to be an osteoconductive biomaterial which acted as a scaffold for a direct osteoblastic apposition and consequently could be an interesting alternative to bone auto-, allo-, or xenografts.