Development of a Synovium-on-Chip Model with a Porous Membrane to Study Inflammatory Arthritis

Arthritis affects millions of people globally and commonly involves synovial inflammation. Unfortunately, for many arthritic diseases, such as Rheumatoid Arthritis (RA) and Osteoarthritis (OA) in particular, there is a limited number of disease-modifying treatments available. This is in part due to a lack of physiologically relevant preclinical models to assess these inflammatory forms of arthritis. Recently, we have developed a Synovium-on-Chip (SoC) platform that modeled the synovial lining, consisting (mainly) of synovial fibroblasts and macrophages, and the synovial vasculature in the subintima with endothelial cells (Figure 1A) [1,2]. However, this model had the limitation of possessing a ~20µm thick nonporous PDMS membrane acting as physical barrier separating the endothelial cells from the human synovial fibroblasts (hSFBs) and macrophages, limiting intercellular communication and the possibility to mimic key pathogenic events such as monocyte extravasation [1].
Here, we address these limitations and integrate a previously developed [3] porous 2µm thick PDMS membrane (pore diameter of 5µm and a 30µm pitch). First, SoC devices were fabricated at a wafer scale with up to 20 devices in a single process to enable higher throughput experiments (Figure 1B). Next, we co-cultured hSFBs from donors with THP-1-derived macrophages (top chamber) and Human Umbilical Vein Endothelial Cells (HUVECs, bottom chamber) for up to 10 days. All three cell types integrated well in the devices (Figure 1C) and confluent cell layers were observed on both sides of the membrane (Figure 1D-E). After verification of their integration and viability, the cells were challenged using 1 ng/mL TNF-α as a proof-of-concept for studying inflammatory arthritis (Figure 1F). After 4 days of stimulation (T10), RT-qPCR was performed on the hSFB and THP-1 cells (top chamber) for IL6, CCL2, MMP1, and TNFAIP6, which were all found to be upregulated by ~2-5 fold (Figure 1G). Surprisingly, microscopic analysis of the devices showed that HUVECs had started to delaminate from the walls of the channels while fibroblast-like cells were visible in the bottom channel (not shown). This prompted the hypothesis that the hSFBs could migrate through the porous membrane, which was tested in devices with nonporous membranes and in devices with only fibroblasts seeded in the top chamber. Interestingly, fibroblast migration was observed through the 5µm pores (Figure 2A), and 3D confocal microscopy revealed that the HUVECs formed a lumen-like structure (Figure 2B). Strikingly, hSFBs (labeled with green CellTracker) were found next to the HUVECs at the bottom of the chip and colocalized with the HUVECs in the lumen-like structure’s lining (Figure 2C). Future research will focus on how the interactions between HUVECs and hSFBs occur, leading to this lumen formation, assessing
stability in long-term culture through microscopy analysis. In short, we have successfully developed a novel SoC model with an integrated porous membrane and in future experiments, we will investigate different inflammatory stimuli, drug efficacy, monocyte extravasation, and the apparent self-organization of HUVECs and hSFBs.