Development of a human osteochondral construct on a microfluidic chip – to advance functional studies of osteoarthritis risk genes

Purpose: To advance functional follow up research of strong OA risk genes that have been frequently found to play a role in the osteochondral unit of the joint tissues, we developed a human in-vitro model system, mimicking the interacting joint tissues bone and cartilage. Methods: The microfluidic chip used in this study consists of a chondrogenic and an osteogenic channel, separated by an electrospun polycaprolactone (PCL) matrix (Fig. 1). Primary osteogenic cells were seeded in the PCL matrix. After culturing the osteogenic cells for seven days, primary chondrocytes were seeded in the chondrogenic compartment, on top of the PCL matrix (RAAK-study, N=8 chips, N=4 patients). After a total culture period of 28 days, the chips were sacrificed and matrix production and gene expression were analyzed by histology and RT-qPCR. As a reference, we included 3D cell pellet cultures of the same patients and we included RNAseq data of preserved OA articular cartilage and subchondral bone of N=15 additional patients. Results: As shown in cross sections of the osteochondral construct under the light microscope after 4 weeks of culturing (Fig. 2, left panel), a layer of cartilaginous matrix was deposited on top of the PCL electrospun matrix, while the osteogenic matrix appeared to be deposited between the PCL fibres. Subsequent Alcian Blue staining (Fig. 2, middle panel) confirmed presence of glycosaminoglycans (GAGs), one of the major components of the cartilage, whereas Alizarin red staining (Fig. 2, right panel) visualized mineralized tissue. Upon measuring gene expression levels, we found cartilage marker COL2A1 being significantly higher expressed in the chondrogenic compartment compared to the osteogenic compartment (fold difference (FD)=9.0, P=0.020; Fig. 3, upper left panel). A similar pattern appeared in COL2A1 expression found in individual chondrogenic and osteogenic in vitro cell pellet cultures that were generated as a reference (Fig. 3, upper right panel). Similarly, the bone marker SPP1 expression was significantly higher in the osteogenic compartment compared to the chondrogenic compartment (FD=8.4, P=0.034 Fig. 3, lower left panel) and again a similar pattern was found in SPP1 expression levels in individual chondrogenic and osteogenic in vitro cell pellet cultures that were generated as a reference (Fig. 3, lower right panel) Conclusions: We developed a microfluidic chip that could serve as a tailored human in vitro OA disease model of the osteochondral unit of joints. Given the multi-tissue function of currently identified OA risk genes such as e.g. IL11, MGP and TGFA, such a model enables translation towards underlying biological mechanisms, preclinical studies on target discovery and eventually drug testing.