A novel 3D spheroid model of rheumatoid arthritis synovial tissue incorporating fibroblasts, endothelial cells, and macrophages

Abstract

Objective: Rheumatoid Arthritis (RA) is a progressive and systemic autoimmune disorder associated with chronic and destructive joint inflammation. The hallmarks of joint synovial inflammation are cellular proliferation, extensive neoangiogenesis and infiltration of immune cells, including macrophages. In vitro approaches simulating RA synovial tissue are crucial in preclinical and translational research to evaluate novel diagnostic and/or therapeutic markers. Two-dimensional (2D) settings present very limited in vivo physiological proximity as they cannot recapitulate cell-cell and cell-matrix interactions occurring in the three-dimensional (3D) tissue compartment. Here, we present the engineering of a spheroid-based model of RA synovial tissue which mimics 3D interactions between cells and pro-inflammatory mediators present in the inflamed synovium.

Methods: Spheroids were generated by culturing RA fibroblast-like-synoviocytes (RAFLS), human umbilical vein endothelial cells (ECs) and monocyte-derived macrophages in a collagen-based 3D scaffold. The spheroids were cultured in the presence or absence of vascular endothelial growth factor (VEGF) and fibroblast growth factor 2 (bFGF) or RA synovial fluid (SF). Spheroid expansion and cell migration were quantified for all conditions using confocal microscopy and digital image analysis.

Results: A novel approach using machine learning was developed to quantify spheroid outgrowth and used to reexamine the existing spheroid-based model of RA synovial angiogenesis consisting of ECs and RAFLS. A 2-fold increase in the spheroid outgrowth ratio was demonstrated upon VEGF/bFGF stimulation (p<0.05). The addition of macrophages within the spheroid structure (3.75x10⁴ RAFLS, 7.5x10⁴ ECs and 3.0x10⁴ macrophages) resulted in good incorporation of the new cell type. The addition of VEGF/bFGF significantly induced spheroid outgrowth (p<0.05) in the new system. SF stimulation enhanced containment of macrophages within the spheroids.

Conclusion: We present a novel spheroid based model consisting of RAFLS, ECs and macrophages that reflects the RA synovial tissue microenvironment. This model may be used to dissect the role of specific cell types in inflammatory responses in RA, to study specific signaling pathways involved in the disease pathogenesis and examine the effects of novel diagnostic (molecular imaging) and therapeutic compounds, including small molecule inhibitors and biologics.

Keywords: 3D model; endothelial cells; fibroblasts; macrophages; rheumatoid arthritis.

Figure 1

Schematic workflow of the spheroid formation and the sprouting assay in a collagen-based 3D scaffold. (1) Cells pooled together in 20% Methocell solution (2) Spheroid formation: distribution of 150 µl cell mix/well into a 96 U-well suspension plate and incubation at 37°C for 24h (3) Collection of the spheroids with a 10 ml pipet (4) Spheroids resuspended in a 1.5 mg/ml collagen solution (5) Sprouting assay: 20 µl of collagen solution containing the spheroids placed dropwise using the sandwich method in a 8-chamber slide and cultured in the suitable media containing stimuli at 37°C for 40h. Created with BioRender.com.

Figure 2

Application of a machine learning quantitative analysis (QuPath) in the pre-established spheroid-based model of RA synovial angiogenesis to measure spheroid outgrowth and morphological changes. Spheroids were either left unstimulated (UNSTIM.), stimulated with VEGF/bFGF (10 ng/ml) or SF (20%). Representative confocal Z-stack projection pictures (10X) of the 3D model containing 3.75x10⁴ RAFLS (magenta) and 7.5x10⁴ EC (cyan), including the specific fluorescence signal for each cell type and the output segmentation of the outgrowth area (pink) versus the core area (blue) using trained pixel classifiers in QuPath (A). Ratio of the spheroid outgrowth area to core area (n=5) (B). Percentage of the total integrated density of RAFLS (C) and ECs (D) calculated in the outgrowth area (n=5). Integrated density is the product of mean fluorescence intensity and area. Statistical significance was determined by RM one-way ANOVA (*=p<0.05).

Figure 3

Incorporation of macrophages results in maintenance of the spheroid structure containing the RAFLS and ECs. Spheroids were all cultured in basal medium. Representative confocal Z-stack projection pictures (10X) of the 3D model containing 3.75x10⁴ RAFLS (magenta), 7.5x10⁴ EC (cyan) and macrophages (yellow) (A). Number of macrophages migrating out from the core detected by QuPath software (n=4-6) (B). Percentage of the total integrated density of macrophages calculated in the core area (n=4-6). Integrated density is the product of mean fluorescence intensity and area (C). Ratio of the spheroid outgrowth area to core area (n=4-6) (D). Statistical significance was determined by RM one-way ANOVA for the analysis of outgrowth ratio and integrated density of RAFLS/EC, and by ratio paired t-test for the integrated density of macrophages (*=p<0.05, **=p<0.01).

Figure 4

VEGF/bFGF significantly induces spheroid outgrowth in the new 3D model containing macrophages, whereas RASF enhances macrophage containment and compaction. Spheroids were left either unstimulated (UNSTIM.), stimulated with VEGF/bFGF (10 ng/ml) or SF (20%). Representative confocal Z-stack projection pictures of the 3D model containing 3.75x10⁴ RAFLS (magenta), 7.5x10⁴ EC (cyan) and 3x10⁴ macrophages (yellow) (A). Representative pictures of H&E staining in paraffin-embedded spheroid sections for the stimulated condition (5 µm) (B). Ratio of the spheroid outgrowth area to core area (n=6) (C). Percentage of the total integrated density of RAFLS (D) and ECs (E) calculated in the outgrowth area (n=6). Percentage of the total integrated density of macrophages calculated in the core area (n=6) (F). Concentrations of TNF and IL-6 proteins in the spheroid supernatants as detected by ELISA (n=5) (G, H). Relative mRNA expression of MMP3 expressed in fold-change value compared to the unstimulated condition (n=3). The relative expression was normalized to both GAPDH and RPLP0 reference genes using ΔΔCT method (average of the two fold-change values) (I). Statistical significance was determined by RM one-way ANOVA for the analysis of outgrowth ratio and integrated density of RAFLS/EC and by ratio paired t-test for the integrated density of macrophages. A Friedman test was applied for TNF and IL-6 protein concentrations (*=p<0.05).