PRE-CONDITIONING WITH TGF-β1 DOES NOT IMPROVE CHONDROGENIC DIFFERENTIATION OF HUMAN MSCS IN A CLINICALLY RELEVANT MODEL UNDER MECHANICAL LOAD

Bone marrow-derived mesenchymal stromal cells (MSCs) are readily available with a large in vitro expansion potential, which makes them attractive candidates for cell-based cartilage therapies. In pellet culture, MSCs can adopt a chondrocyte-like phenotype when stimulated with exogenous TGF-β1. However, the quality of the chondrogenic differentiation varies considerably between cells of different donors, which is still one of the biggest obstacles for the clinical translation of MSC-based therapies. Pre-conditioning MSCs with TGF-β1 during the monolayer expansion phase has been shown to rescue the chondrogenic differentiation of poorly responsive cells in subsequent pellet culture 1 . As cartilage is a mechanosensitive tissue, which is constantly exposed to compression and shear, static pellet culture does not adequately represent the in vivo cellular environment. Here, we therefore tested the effect of TGF-β1 pre-conditioning on the chondrogenic differentiation of MSCs in a clinically more relevant model under mechanical load. Human MSCs ( N = 6) from bone marrow aspirates were expanded following standard methods. For the last two days of expansion, half of the cells were pre-conditioned with 10 ng/mL TGF-β1 (+TGF-β group) while the remainder were kept in regular medium (no TGF-β group). Cells were then seeded into polyurethane-fibrin scaffolds and divided into three groups: free swelling (FS) in chondrogenic medium (CH), FS in chondropermissive medium (CP), i.e. CH without TGF-β1, or multiaxial load in CP medium (CPLS) as previously described 2 . After 28 days, constructs were collected for histological evaluation and gene expression analysis. Conditioned medium was collected to quantify the release of sulphated glycosaminoglycans (sGAG) and TGF-β. In the CH group, pre-conditioning increased COL2A1 gene expression (fold change: 1.36; p<0.05) and slightly improved Saf-O staining compared to unconditioned CH constructs. In contrast, pre-conditioning led to a small, non-significant decrease in COL2A1 expression, a reduction in COMP expression (fold change: 0.63; p<0.05) and an increase in RUNX2 expression (fold change: 1.33; p<0.05) in CPLS constructs. Accordingly, histological sections showed reduced Saf-O staining in the CPLS group after pre-conditioning. Pre-conditioning had no effect on TGF-β and sGAG release. While TGF-β1 pre-conditioning seems beneficial in FS constructs, it does not improve chondrogenic differentiation in mechanically-induced chondrogenesis, which underlines the importance of testing new therapies in clinically relevant models.