Standard in vitro evaluations of engineered bone substitutes are not sufficient to predict in vivo preclinical model outcomes

Understanding the optimal conditions required for bone healing can have a substantial impact to target the problem of non–unions and large bone defects. The combination of bioactive factors, regenerative progenitor cells and biomaterials to form a tissue engineered (TE) complex is a promising solution but translation to the clinic has been slow. We hypothesized the typical material testing algorithm used is insufficient and leads to materials being mischaracterized as promising. In the first part of this study, human bone marrow – derived mesenchymal stromal cells (hBM-MSCs) were embedded in three commonly used biomaterials (hyaluronic acid methacrylate, gelatin methacrylate and fibrin) and combined with relevant bioactive osteogenesis factors (dexamethasone microparticles and polyphosphate nanoparticles) to form a TE construct that underwent in vitro osteogenic differentiation for 28 days. Gene expression of relevant transcription factors and osteogenic markers, and von Kossa staining were performed. In the second and third part of this study, the same combination of TE constructs were implanted subcutaneously (cell containing) in T cell-deficient athymic Crl:NIH-Foxn1rnu rats for 8 weeks or cell free in an immunocompetent New Zealand white rabbit calvarial model for 6 weeks, respectively. Osteogenic performance was investigated via MicroCT imaging and histology staining. The in vitro study showed enhanced upregulation of relevant genes and significant mineral deposition within the three biomaterials, generally considered as a positive result. Subcutaneous implantation indicates none to minor ectopic bone formation. No enhanced calvarial bone healing was detected in implanted biomaterials compared to the empty defect. The reasons for the poor correlation of in vitro and in vivo outcomes are unclear and needs further investigation. This study highlights the discrepancy between in vitro and in vivo outcomes, demonstrating that in vitro data should be interpreted with extreme caution. In vitro models with higher complexity are necessary to increase value for translational studies. Preclinical testing of newly developed biomaterials is a crucial element of the development cycle. Despite this, there is still significant discrepancy between in vitro and in vivo test results. Within this study we investigate multiple combinations of materials and osteogenic stimulants and demonstrate a poor correlation between the in vitro and in vivo data. We propose rationale for why this may be the case and suggest a modified testing algorithm.