Background Immunosuppressed patients with ongoing inflammation experience more often difficulties in the process of fracture healing. Hereby amongst others immunomodulated activation of osteoclasts leads to augmented osteolysis. Most notably numerous cytokines and invading lymphocytes provide an inflammatory environment within the fracture gap, utmost during the first phase of fracture healing. Nowadays researchers, whilst investigating fracture healing, use small rodent models, facing the problem of translation towards the human system. Hence, there is a lack of valid in vitro models to examine the first phase of fracture healing. To test new therapeutic strategies, we develop a valid 3D-model using human cells to mimic the first phase of fracture healing in vitro which is characterized by the formation of a fracture hematoma, a hypoxic microenvironment as well as inflammation that initiate the healing cascade and the process of endochondral ossification.
Objectives To develop 3D bone-like models displaying endochondral ossification.
Methods As a first step, we focus on establishing 3D bone-like, matrix-free models consisting of human mesenchymal stromal cells (hMSC). MSC were isolated from bone-marrow samples of patients undergoing total hip replacement and characterized with regard to their typical surface markers as well as their differentiation potential towards osteogenic, adipogenic and chondrogenic lineage. 3D bone-like models were generated from micro-mass culture of hMSC (Research Center of Medical Technology and Biotechnology, Bad Langensalza). After initial maturation, 3D bone-like models were cultured under hypoxic conditions (37°C, 1% O2) in osteogenic medium for up to 3 months. In vitro μCT analyses were performed at day 0, after one and two months focusing on the total volume (TV) and bone volume (BV). Additionally osteogenic-relevant genes/factors (Runx2, SPP1, Dlx5, ALP, RANKL, SPI1) as well as exclusion markers (SOX9, PPARg2) were investigated after 3 months of cultivation using qRT-PCR. Finally, we implemented histological/immunohistochemical methods (van Kossa, ALP, Col1 staining).
Results The 3D bone-like models achieved diameters between 0.5 and 0.7 and a thickness of 0.3 cm. In vitro μCt analysis revealed a high amount of mineralized tissue at day 0 and substantial increase in the bone volume (BV/TV) after cultivation. At day 0 in vitro μ Ct analysis implied mineralization mostly at the margin but penetrated the tissue within further cultivation. These results are also supported by positive van Kossa staining. Additionally, qRT-PCR results yielded higher expressions of RUNX2, SPP1, SPI1, RANKL and DLx5. Furthermore, immunohistochemistry showed high ALP-activity and ColI-expression.
Conclusions Preliminary results of our study focusing at developing a 3D bone-like model displayed a promising trend towards modelling endochondral ossification in vitro by increased mineralization (in vitro μCt analysis and van Kossa staining), and upregulation of osteogenic-relevant RUNX2 and SPP1 expression as well as ALP-activity and ColI-expression. High expression of SPI1 and RANKL could refer to osteoclast-like activities, which will be in the focus of further investigations. Finally, the complete 3D model will leave us the opportunity for studying the first phase of fracture healing under in vitro conditions.
Disclosure of Interest None declared