Developing an in vitro Osteochondral Micro-Physiological System for Modelling and Testing Therapeutics for Osteoarthritis

Introduction: Musculoskeletal diseases are the leading cause of disability, and often affect a combination of tissues including bone, muscle, and cartilage. Currently, animal models remain popular for musculoskeletal research, even though they may not accurately represent human biology. However, poorly planned and inconsistent animal research wastes up to $24 billion per year. In vitro models could be used to replace certain animal studies, but to be biologically relevant they must capture the three-dimensional nature of musculoskeletal tissues as well as their crosstalk. In this study, we developed a micro-physiological in vitro system that uses a biphasic bioreactor and multiple cell types and scaffolds to model native biology and crosstalk of the osteochondral unit of articular joints. We validated the efficacy of the model to emulate osteoarthritis (OA) by treating with inflammatory factors. We then probed whether the addition of the vascularized bone component to the model affects the response of the cartilage region to inflammation, and whether or not the inflammation of the cartilage region causes subsequent bone responses. We showed how our in vitro model can recapitulate cross talk across tissue types, enabling the study of OA development in vitro and the screening of candidate therapeutics.

Materials and

Methods: Creating the Osteochondral Construct: For the cartilage region (C), human bone-marrow derived mesenchymal stem cells (BM-hMSCs) were seeded in 5% methacrylated gelatin (GelMA), 2.5% methacrylated hyaluronic acid (MeHA), and photo-crosslinked. Cartilage constructs were pre-differentiated for 28 days in chondrogenic medium containing 10ng/mL TGF-β3. For the vascularized bone region (VB), BM-hMSCs are seeded onto a 3D printed 10 % GelMA, 50 % NanoHydroxyapatite, 0.2% genipin-crosslinked scaffold. After 14 days of osteogenic differentiation with 10mM β-glycerophosphate and 10nM 1,25-dihydroxy vitamin D3, human umbilical vein endothelial cells (HUVECS) and BM-hMCSs (4:1) were seeded into the pores of the 3D printed scaffold in a 5% GelMA, 0.0025% fibrinogen, 0.00004% thrombin suspension, crosslinked in situ, and cultured in 50:50 osteogenic and vasculogenic medium for an additional 14 days. The chondral and osseous components are combined in the biphasic bioreactor prior to experiments.

Simulating an Osteoarthritic Environment: The cartilage constructs are treated with inflammatory conditions for 7 days either alone, or in the bioreactor with the vascularized bone construct, using either: control (no additives) (CTRL), combination of interleukin 1β (IL-1β) (20ng/mL), interleukin 6 (IL-6) (100ng/mL), and tumor necrosis factor α (TNF-α) (1000ng/mL) (CMB), or pro-inflammatory macrophage conditioned medium (MCM).

Analysis of Differentiation, Inflammation, and Remodeling: Chondrogenesis is assessed by RT-qPCR (SOX9, ACAN, COL2), histology, and immunofluorescence (collagen II), and osteogenesis by histology, RT-qPCR (BSP2, RUNX2, ALP), and immunofluorescence (collagen I). Response to inflammatory challenge is assessed with RT-qPCR for MMP1, MMP13, and immunofluorescence for collagen I, and matrix metalloproteinase assay.

Results, Conclusions, and Discussions:

Results:
Isolated Cartilage Constructs: Chondrogenesis and inflammation of the cartilage constructs alone is shown in Figure 1. There is a noticeable decrease of chondrogenic genes in response to inflammatory conditions, and an upregulation of remodeling genes MMP1 and MMP13 in the inflammatory conditions, especially the MCM.

Structure of Osteochondral Construct: The combined osteochondral construct shows region specific expression for bone and cartilage markers, respectively (Figure 2). Masson’s trichrome staining shows strong collagen deposition in both compartments, with calcification visible in the osseous region. This is confirmed by collagen I and collagen II staining, the latter only in the cartilaginous region. Alcian blue staining shows GAG deposition in the cartilaginous region. This confirms the biphasic bioreactor supports osteochondral constructs maturation and distinct cartilaginous and vascularized osseous regions reminiscent of the osteochondral junction in articular joints.

Inflammation of the Osteochondral Construct: RT-qPCR shows similar trends in chondrogenic gene expression in the bioreactor compared to the cartilage only constructs. We additionally see expression of remodeling genes in the vascularized bone component when only cartilage is challenged, suggesting that the inflamed cartilage signals to the bone, driving bone inflammatory response (Figure 3).

Discussion:
Our preliminary data show that we have successfully developed a multi-tissue micro-physiological system simulating OA. We see that the cartilage region responds to inflammation similarly to OA in vivo, and that there is crosstalk between the cartilage and the vascularized bone region in our model. The macrophage conditioned media causes increased inflammation, reminiscent of the macrophage-driven synovial inflammation seen in vivo. Moving forward, we aim to use this system to test OA therapeutics like Rapamycin and TNF inhibitors.

Conclusion:
Using a micro-physiological system to test therapeutics for OA prior to animal studies can both minimize the number of animals required for preclinical testing and save on animal costs. These advanced in vitro systems will help advance knowledge of musculoskeletal disease development as well as allow rapid testing of potential therapeutics for treatment.

Acknowledgements (Optional): Support from the Children’s Hospital of Philadelphia Research Institute (RG), Ri.MED Foundation (RG), NIH T32-AR007132 (KWYS) and P30 AR069619, and the Fontaine Fellowship (KWYS). Thanks to the Penn Center for Musculoskeletal Disorders and CFD-Research.