Background: Our ultimate goal is to study potential drug candidates in an experimental setting of arthritis. Therefore, we aim to develop a valid human in vitro 3D joint model mimicking features of joint inflammation by applying inflammatory conditions namely immune cells and pro-inflammatory cytokines. Our i n vitro 3D joint model consists of different components including an osteogenic and chondrogenic part, the joint space filled with synovial fluid, and the synovial membrane. Developed as an alternative experimental setup to animal experiments, our 3D joint model will enable us to study efficiently the effects of potential drug candidates in a human-based in vitro model.

Objectives: Here, we aimed to demonstrate the suitability of our human-based in vitro 3D osteochondral model by analyzing the influence of the main cytokines involved in the pathogenesis of RA as well as the impact of a specific therapeutic intervention.

Methods: Based on human bone marrow-derived mesenchymal stromal cells (hMSCs), we developed 3D bone and cartilage tissue components that were characterized in detail (e.g. cell vitality, morphology, structural integrity) using histological, biochemical and molecular biological methods as well as µCT and scanning electron microscope (SEM). In brief, to establish the osteogenic component, we populated β-tricalcium phosphate (TCP) – mimicking the mineral bony part – with hMSCs, while the scaffold-free cartilage component was generated by cellular self-assembly and intermittent mechanical stimulation (fzmb GmbH). Subsequently, we co-cultivated both tissue components for three weeks to generate an interconnected 3D osteochondral model. To test the suitability, we applied a cocktail of TNFα, IL-6 and MIF using concentrations reported from RA synovial fluid alone or in combination with specific therapeutic drugs and analyzed their impact by qPCR.

Results: We verified the osteogenic phenotype of our 3D bone tissue component by demonstrating an increase in mineralized bone volume and the induction of bone-related gene expression ( RUNX2 , SPP1 and COL1A1 ) as compared to the corresponding control. Secondly, we verified the chondrogenic phenotype of our cartilage tissue component by HE and Alcian Blue staining as well as by the reduced expression of COL1A1 and an abundant expression of COL2A1 . Interestingly, co-cultivation of both components for up to 3 weeks demonstrated colonization, connectivity and initial calcification implying a transitional bridging area. Cytokine stimulation with a cocktail of TNF, IL-6 and MIF leads to an upregulation of the metabolic marker LDHA and the angiogenic marker VEGF in both bone and cartilage. The inflammation markers IL8 and TNF are also upregulated in both components, while IL6 is downregulated in bone compared to the unstimulated control. In addition, a cytokine-induced upregulation of matrix-metalloproteases was observed especially in the cartilage component. All these cytokine-related effects could be antagonized with a cocktail of therapeutics (milatuzumab, adalimumab and tocilizumab).

Conclusion: The results of our study showed cytokine related effects of both tissue components, which can be therapeutically antagonized. By combining the components in a 96 well format, we aim to provide a mid-throughput system for preclinical drug testing.

Acknowledgments: This project is funded by the Federal Ministery of Education and Research (BMBF)

Disclosure of Interests: Alexandra Damerau: None declared, Moritz Pfeiffenberger: None declared, Annemarie Lang: None declared, Timo Gaber: None declared, Frank Buttgereit Grant/research support from: Amgen, BMS, Celgene, Generic Assays, GSK, Hexal, Horizon, Lilly, medac, Mundipharma, Novartis, Pfizer, Roche, and Sanofi.