Association of Synovial Innate Immune Exhaustion With Worse Pain in Knee Osteoarthritis

Abstract

Objective: Uncontrolled pain remains a major clinical challenge in the management of knee osteoarthritis (OA), the most common disabling joint disease. Worse pain is associated with synovial innate immune cell infiltration (synovitis), but the role of innate immune-regulatory cells in pain is unknown. Our objective was to identify synovial innate immune cell subsets and pathophysiologic mechanisms associated with worse pain in patients with knee OA.

Methods: Synovial tissue biopsies from 122 patients with mild-to-severe knee OA pain (Knee Injury and OA Outcome Score [KOOS]) were analyzed to identify associations between synovial histopathology and worse pain. We then used spatial transcriptomics and proteomics of synovial tissue microenvironments (n = 32), followed by single-cell RNA sequencing (n = 8), to identify synovial cell composition and cell-cell communication networks in patients with more severe OA pain.

Results: Histopathological signs of synovial microvascular dysfunction and perivascular edema were associated with worse KOOS pain (-10.76; 95% confidence interval [CI] -18.90 to -2.61). Patients with worse pain had fewer immune-regulatory macrophages, expanded fibroblast subsets, and enrichment in neurovascular remodeling pathways. Synovial macrophages from patients with worse pain expressed markers of immune exhaustion and decreased phagocytic function (-19.42%; 95% CI -35.96 to -2.89) and their conditioned media increased neuronal cell stress in dorsal root ganglia.

Conclusion: Although synovitis increases during OA, our findings suggest that exhaustion, dysfunction, and loss of immune-regulatory macrophages is associated with worse pain and may be an important therapeutic target.

Figure 1

Spatial profiling of synovial lining, subintima, and microvessel microenvironments. (A–C) Representative images of the synovial microenvironment regions of interest. Immunofluorescence images from the spatial profiling analysis with morphology markers for CD68+ macrophages (yellow), CD45+ cells (magenta), SMA+ cells (green), and nuclear stain (blue). Representative synovial microenvironment regions of interest for (A) synovial lining, (B) subintima, and (C) microvessel are outlined with white, dashed lines. Scale bar = 100 μm. Volcano plots displaying the top differentially expressed (increased vs decreased expression) (D–F) genes and (G–I) proteins for each synovial tissue microenvironment: synovial lining, subintima, and microvessel in patients with more pain (relative to less pain). Genes with increased expression in patients with more pain are represented by red dots, and genes with decreased expression in more pain are represented by blue dots. The y‐axis represents the log10 P value with cut‐off set at 1.3 (P < 0.05), and the x‐axis represents the log2 fold change with cut‐off set at 0.5. Differential gene expression (n = 16 per group) was derived by fitting linear mixed models while adjusting for patient ID and BMI and with a Benjamini‐Hochberg correction to control for FDR. BMI, body mass index; CD45, hematopoietic cell marker; CD68, macrophage marker; FDR, false discovery rate; ID, identifier; SMA, smooth muscle actin vessel marker.

Figure 2

Single‐cell transcriptomics of OA synovium from patients with more versus less pain. (A) UMAP plot of synovial cell clusters annotated by cell types. (B) Expression level of the selected marker(s) used for identification of each cluster. (C) Proportion of main cell types in patients reporting “less pain” (n = 4) and patients reporting “more pain” (n = 4). (D–K) Cell type–specific analyses for myeloid cells and fibroblasts. (D) Subclustering of myeloid cells plotted on UMAP and annotated by subtype: T‐MΦ, IFNS‐MΦ, IR‐MΦ, and S100A8+ macrophages (S100A8^hi). (E) Myeloid subclustering UMAPs plotted by pain group. (F) Subclustering of fibroblast cells plotted on UMAP and annotated by subtype: lining fibroblasts, LRRC15+ lining, senescent lining, sublining progenitor, fibroblasts, perivascular sublining, HLA‐DRA^hi fibroblasts, and intermediate fibroblasts. (G) Fibroblast subclutsering UMAPs plotted separated by pain group. (H) Volcano plot of differentially expressed genes from myeloid pseudobulk analysis. (I) Bubble plot of Reactome gene sets enriched by myeloid cells from more pain relative to less pain. (J) Volcano plot of differentially expressed genes from fibroblast pseudobulk analysis. (K) Bubble plot of Reactome gene sets enriched by fibroblasts from more pain relative to less pain. DC, dendritic cell; DEG, differentially expressed gene; EMT, epithelial–mesenchymal transition; IFNS‐MΦ, interferon‐stimulated macrophage; IL, interleukin; IR‐MΦ, immune‐regulatory macrophage; LRCC15+, leucine‐rich repeat containing 15; MHC, major histocompatibility complex; NGF, nerve growth factor; NTRK, neurotrophic tyrosine receptor kinase; OA, osteoarthritis; PDGF, platelet‐derived growth factor; ROBO, roundabout homolog; SLIT, slit guidance ligand; RNS, reactive nitrogen species; ROS, reactive oxygen species; T‐MΦ, transitional macrophage; TGF, transforming growth factor; TLR, Toll‐like receptor; TNFR, tumor necrosis factor receptor; TRKA, tropomyosin‐related kinase receptor A; UMAP, Uniform Manifold Approximation and Projection for Dimension Reduction; VEGFR, vascular endothelial growth factor receptor. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43089/abstract.

Figure 3

CD68+ macrophage phagocytic index among patients with more versus less pain. (A) Representative immunofluorescence images of CD68+ macrophages (red) and phagocytosing beads (green) from patients with more or less pain. (B) Dot plot displaying the individual measures of and mean ± 95% CI phagocytic index for more pain (n = 8) and less pain (n = 8) groups. Phagocytic index is defined as the number of CD68+ macrophages with one or more phagocytosed beads out of the total number of CD68+ macrophages. An unpaired t‐test was used to compare phagocytic index between groups. DAPI nuclear stain. Scale bar = 10 μm. *P < 0.05. CD68, macrophage marker; CI, confidence interval; FDR, false discovery rate; FITC, fluorescein isothiocyanate latex bead marker. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43089/abstract.

Figure 4

CC3 and Iba1 quantification in rat DRGs after stimulation with media conditioned by synovial cells from patients with more versus less pain. (A–D) Representative immunofluorescence images for each stimulation group (n = 8 per group). Groups are as follows: (A) unconditioned, (B) LPS stimulated, (C) less pain, and (D) more pain. Dot plots displaying the individual measures and mean ± 95% CI of (E) CC3+ cell bodies (green) and (F) Iba1+ cells (magenta) per stimulation group. Unpaired t‐tests were used to compare the percentage of positive cells between groups. White arrow and arrowhead indicate CC3+ cell bodies and Iba1+ cells, respectively. DAPI nuclear stain. Scale bar = 20 μm. *P < 0.05. CC3, cleaved caspase 3; CI, confidence interval; DRG, dorsal root ganglia; Iba1, ionized calcium‐binding adapter molecule 1 activated microglia marker; LPS, lipopolysaccharide. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/doi/10.1002/art.43089/abstract.