Background Our ultimate goal is to develop a valid in vitro 3D model to simulate the immune mediated pathogenesis of arthritis in order to present an alternative experimental setup for the traditional animal models. Therefore, we work to establish an in vitro simulation of a healthy joint, followed by the depiction of an inflamed arthritic joint to finally study the efficacy of drug treatment. The in vitro 3D joint model consists of different components including an (1) osteogenic and (2) chondrogenic part, (3) the joint space with synovial fluid and (4) the synovial membrane. The model is suggested to include all involved cell types and thus, to allow interactions between cells by cell contacts and signaling molecules. To our knowledge, there is currently no valid 3D model which is able to mimic an inflamed arthritic joint.

Objectives Here, we aim to mimic the (1) osteogenic and (2) chondrogenic part of the joint for our in vitro multi-component 3D joint model.

Methods We used β-tricalcium phosphate (TCP) particles as a mineralized 3D bone scaffold and human bone marrow derived mesenchymal stem cells (hMSCs), non- and pre-differentiated towards osteoblastic lineage. Both were cultured up to 21 days on β-TCP. Osteogenic differentiation was performed in the presence of osteogenic supplements under normoxic conditions (37 °C, 18% O2). Adhesion and proliferation of hMSCs on β-TCP were evaluated by immunofluorescence and histological analysis. To confirm cell attachment and biocompatibility of β-TCP particles cellular release of LDH was assessed. Osteogenic differentiation was analyzed on gene expression level using qRT-PCR. The chondrogenic model, a scaffold-free 3D cartilage construct (fzmb GmbH) was generated using hMSCs. Chondrogenic differentiation was performed under hypoxia (37 °C, 1% O2) with intermittent mechanical stimulation and analyzed by histology.

Results We developed an in vitro 3D trabecular bone model by seeding hMSCs on β-TCP scaffold after pre-incubation for 24 hours. The analysis of cell viability via LDH detection showed no toxic effects on the cells seeded as compared to the corresponding control. Furthermore, we assessed cell attachment and proliferation by measurement of LDH activity after scaffold crushing. As a result, samples showed higher LDH activity compared to the controls. Histological and immunofluorescence analysis based on DNA and actin staining demonstrated cell attachment until day 21. After 21 days, cells were located more inside the scaffold compared to day 1. qRT-PCR expression of bone-related genes such as RUNX2, SPP1 and COL1A1 confirmed the phenotypic change during osteogenic differentiation on the scaffold. Furthermore, the scaffold-free 3D chondrogenic structure was confirmed by HE staining representing the different zones. Cartilage phenotype was confirmed by the reduced expression of Col1a1, an abundant expression of Col2a1 and Aggrecan.

Conclusions The initial results from our in vitro 3D osteogenic and chondrogenic model confirm good cell vitality which indicates successful progression. To confirm the exchange of β-TCP through cellular matrix, we will now extend the assay co-cultivation time for up to 6 weeks. This 3D multi-component joint model should enable us to simulate arthritis and to study the efficacy of drug treatment in vitro.

Disclosure of Interest None declared