Microfluidic Osteoarthritis-on-a-Chip: Modeling Human Joint Inflammation

Abstract Osteoarthritis (OA) is a multifactorial joint disease driven by complex interactions among chondrocytes, osteoblasts, fibroblasts, and immune cells across cartilage, bone, and synovial tissues. Conventional monoculture systems are unable to capture this crosstalk, limiting their physiological relevance. Building on our previously established joint-on-a-chip platform, this study evaluated multicellular communication and assessed whether a microfluidic co-culture provides a more realistic representation of joint inflammation compared with monoculture models. Two configurations were established: a healthy, low-inflammation model containing M0 macrophages and an OA-like, high-inflammation model with M1 macrophages. In healthy models, co-culture significantly increased MMP-1 (∼4-fold), MMP-3 (∼15-fold), TIMP-2 (∼5-fold), IL-6 (∼6-fold), and IL-8 (∼5-fold) relative to monoculture, indicating that endogenous signaling initiates basal matrix remodeling and inflammatory pathways. In disease models, M1-driven co-culture elevated MMP-10 (∼300-fold) and MMP-13 (∼60-fold), along with TIMP-2 (∼5-fold), compared with monoculture, reflecting amplified catabolic activation. Direct comparison of disease versus healthy co-culture revealed additional increases in MMP-10 (∼55-fold), MMP-13 (∼95-fold), MCP-1 (∼1.6-fold), MMP-1 (∼1.6-fold), MMP-3 (∼1.8-fold), TIMP-1 (∼1.4-fold), and TIMP-2 (∼1.5-fold), representing a macrophage-mediated shift from homeostasis to OA-like pathology. However, neither IL-1 nor TNFα, each a key inflammatory mediator of OA, differed measurably between healthy and disease models under either monoculture or co-culture conditions. Thus, the microfluidic joint inflammation-on-a-chip model presented here more faithfully recapitulates the pathogenic MMP profile of OA than monoculture systems, but it does not yet fully recapitulate the pathogenic inflammatory environment of OA.