Synovium and infrapatellar fat pad share common mesenchymal progenitors and undergo coordinated changes in osteoarthritis

Abstract

Osteoarthritis (OA) affects multiple tissues in the knee joint, including the synovium and intra-articular adipose tissue (IAAT) that are attached to each other. However, whether these two tissues share the same progenitor cells and hence function as a single unit in joint homeostasis and diseases is largely unknown. Single-cell transcriptomic profiling of synovium and infrapatellar fat pad (IFP), the largest IAAT, from control and OA mice revealed five mesenchymal clusters and predicted mesenchymal progenitor cells (MPCs) as the common progenitors for other cells: synovial lining fibroblasts (SLFs), myofibroblasts (MFs), and preadipocytes 1 and 2. Histologic examination of joints in reporter mice having Dpp4-CreER and Prg4-CreER that label MPCs and SLFs, respectively, demonstrated that Dpp4+ MPCs reside in the synovial sublining layer and give rise to Prg4+ SLFs and Perilipin+ adipocytes during growth and OA progression. After OA injury, both MPCs and SLFs gave rise to MFs, which remained in the thickened synovium at later stages of OA. In culture, Dpp4+ MPCs possessed mesenchymal progenitor properties, such as proliferation and multilineage differentiation. In contrast, Prg4+ SLFs did not contribute to adipocytes in IFP and Prg4+ cells barely grew in vitro. Taken together, we demonstrate that the synovium and joint fat pad are one integrated functional tissue sharing common mesenchymal progenitors and undergoing coordinated changes during OA progression.

Keywords: joint fat pad; mesenchymal progenitors; osteoarthritis; single-cell RNA-sequencing; synovium.

Graphical Abstract

Figure 1

Synovium and IAATs are an integrated joint tissue. (A) H&E staining of a mouse knee joint at midsagittal region. Boxed area in the image at the top (a) is magnified at the bottom (b). Arrows point to synovium. (B) Representative fluorescence images of Col2/td knee joints with Perilipin staining. Images (a) and (d) are posterior and anterior regions, respectively. Images (b), (c), (e), and (f) are magnified images from boxed areas from (a) and (d). (C) Schematic cartoon of a mouse knee joint shows sectioning sites. (D) H&E staining of medial, middle, and lateral sections from control knees with no surgery and DMM knees at 1 wk, 1 mo, or 3 mo postsurgery. (E) IFP area was quantified at three locations from control and DMM knees. N = 3 mice/group. ^**, P < .01; ^***, P < .001 vs control. F, femur; T, tibia; M, meniscus; PCL, posterior cruciate ligament; PT, patellar tendon; IFP, infrapatellar fat pad; PFP, posterior fat pad; EJFP, extra-joint fat pad; A, anterior; P, posterior. Scale bars, 200 μm.

Figure 2

Single-cell transcriptomics analyses of normal and DMM-injured Syn/IFP tissue identify distinct cell types and cell–cell interactions. (A) UMAP plots of 29763 Syn/IFP cells collected from control, DMM 1-wk, and DMM 2-mo mouse knees are shown as integrated (merged) and split view. Nine cell clusters are visualized. VEC, vascular endothelial cells; LEC, lymphatic endothelial cells; Mφ, macrophages. (B) Dot plot shows marker gene expression for each cluster. (C) Circle plot shows the number of interactions among cell clusters predicted by CellChat. Line width is proportional to the number of interactions. Line color is the same as the color of sender cells. (D) Quantification of outgoing/incoming cell–cell interaction number.

Figure 3

Subclustering fibroblasts predicts a common progenitor for synovium and IFP tissues. (A) UMAP plots of 17097 fibroblasts in control, DMM 1-wk, and DMM 2-mo datasets are shown as integrated (merged) and split view. Five cell clusters are visualized. MPC, mesenchymal progenitor cells; SLF, synovial lining fibroblasts; MFs, myofibroblasts; PreAD1 and 2, preadipocytes 1 and 2. (B) Dot plot shows marker gene expression for each cluster. (C) The percentage of cells in each cluster was calculated in control, DMM 1-wk, and DMM 2-mo datasets. (D) The percentage of proliferating cells in each cell cluster was calculated. (E) Monocle pseudotime trajectory of fibroblasts from the control dataset. The trajectories of cells from individual clusters are shown on the right. (F) Hierarchy analysis of fibroblast cell clusters from the integrated dataset. The top five genes ranked by avg_log2FC of each cluster are listed. (G) Pseudotemporal depiction of differentially expressed TFs during the differentiation from MPC toward SLF (left) and from MPC toward PreAD (right). The color bar indicates the gene expression level.

Figure 4

Dpp4+ MPCs are common progenitors for Syn/IFP tissue during postnatal joint development. (A) Schematic diagram shows the study protocol of lineage tracing in young mice. Mice received Tam at P14–16 daily, and their joints were harvested at P19 and 3 mo for histology analysis. (B) Representative fluorescence images of Dpp4ER/Td knee joints with Perilipin staining. Boxes in (a) and (d) are shown at a high magnification as (b, c) and (e, f), respectively. Arrows and arrow heads point to Td+Perilipin− and Td+Perilipin+ cells, respectively. (C) Representative fluorescence images of synovium from Dpp4ER/Td mice show that Td signals overlap with Dpp4 staining well. Arrows point to Td+Dpp4+ cells, and an arrow head points to Td−Dpp4+ cells. (D) Fluorescence images of synovium with Prg4 and Cd45 staining show that Td+ cells overlap with Prg4+ SLFs at 3 mo but not at P19 and that Td+ cells are Cd45−. Arrows and arrow heads point to Td+Prg4− and Td+Prg4+ cells, respectively. H, tissue harvest; F, femur; T, tibia; M, meniscus; Syn, synovium; IFP, infrapatellar fat pad. Scale bars: 200 μm.

Figure 5

Dpp4+ MPCs are common progenitors for Syn/IFP tissue during OA progression induced by DMM. (A) Schematic diagram shows the study protocol of lineage tracing in mice after DMM surgery. Mice received daily Tam injections for 5 d at 2 wk before DMM surgery. Their joints were harvested at 1 wk and 3 mo postsurgery for histology analysis.(B) Representative fluorescence images of Dpp4ER/Td knee joints with Perilipin staining. Boxes in (a), (d), (g), and (j) are shown at a high magnification as (b, c), (e, f), (h, i), and (k, l), respectively. Arrows and arrow heads point to Td+Perilipin− and Td+Perilipin+ cells, respectively. (C) Fluorescence images of synovium with Prg4 staining show that Td+ cells overlap with Prg4+ SLFs at 3 mo in DMM knees but not in control knees. Arrows point to Td+Prg4+ cells. (D) Representative flow cytometry images show the gating strategy for Td+ cells within Cd45− cells from Syn/IFP collected at 1 wk postsurgery. (E) The percentage of Td+ cells within Cd45− cells was quantified. ^***, P < .001 vs control. (F) Representative fluorescence images of EdU incorporation in the synovium at 1 wk postsurgery. Arrows and arrow heads point to EdU+Td− and EdU+Td+ cells, respectively. (G) Quantification of the percentage of EdU+ cells in Td+ cells. ^***, P < .001 vs control. (H) Fluorescence images of synovium (syn) and IFP of Dpp4ER/Td knee joints at 1 wk postsurgery with αSMA and Perilipin staining. (I) Fluorescence images of synovium (syn) and IFP of Dpp4ER/Td Scx-GFP knee joints at 1 wk postsurgery with Perilipin staining. (J) Fluorescence images of synovium (syn) and IFP of Dpp4ER/Td knee joints at 1 wk postsurgery with Cd45 and Perilipin staining. H, tissue harvest; F, femur; T, tibia; M, meniscus; Syn, synovium; IFP, infrapatellar fat pad. Scale bars: 200 μm.

Figure 6

Dpp4+ MPCs possess mesenchymal progenitor properties in vitro. (A). Representative bright field (BF) and fluorescence light (FL) images of Td− and Td+ CFU-F colonies formed by Syn/IFP cells from Dpp4ER/Td mice. Mice received Tam injections for 5 d at 2 wk before tissue harvest. Scale bar: 200 μm. (B) Crystal violet stain for CFU-F of Syn/IFP cells from Dpp4ER/Td control and DMM surgery mice. Mice received daily Tam injections for 5 d at 2 wk before DMM surgery. Their Syn/IFP cells were harvested at 1 wk postsurgery. (C) Quantification of Td− and Td+ CFU-F colonies per knee from control and DMM surgery groups. (D) The growth curves of Syn/IFP mesenchymal cells from control and DMM knees were measured in the culture. ^**, P < .01; ^***, P < .001 vs control. (E) CFU-F frequency of sorted Td− and Td+ cells from Syn/IFP tissue. ^***, P < .001 vs Td−. (F) The growth curves of sorted Td− and Td+ cells from Syn/IFP tissue were measured in the culture. ^***, P < .001 vs Td−. (G) Oil Red O staining of sorted Td− and Td+ cells that undergo adipogenic differentiation for 1 wk. (H) qRT-PCR analysis of adipogenic marker gene expression in Td− and Td + cells after 1 wk of adipogenic differentiation. ^**, P < .01; ^***, P < .001 vs Td−. (I) Alizarin red staining of sorted Td− and Td+ cells that undergo osteogenic differentiation for 2 wk. (J) qRT-PCR analysis of osteogenic marker gene expression in Td− and Td+ cells after 2 wk of osteogenic differentiation. The gene expression in cells cultured in growth medium for the same period was set as 1. ^*, P < .05; ^**, P < .01; ^***, P < .001 vs Td−.

Figure 7

Prg4+ SLFs are not common progenitors for Syn/IFP tissue. (A) Representative fluorescence images of Prg4ER/Td knee joints with Perilipin staining. Mice received daily Tam injections for 5 d at 2 wk before DMM surgery. Their joints were harvested at 1 wk and 3 mo postsurgery for histological analysis. Boxes in (a), (d), (g), and (j) are shown at a high magnification as (b, c), (e, f), (h, i), and (k, l), respectively. F, femur; T, tibia; M, meniscus. (B) Fluorescence images of synovium with Cd45 staining show that Td+ cells are Cd45−. (C) Fluorescence images of synovium (syn) and IFP of Prg4ER/Td knee joints at 1 wk postsurgery with αSMA staining. (D) Fluorescence images of synovium (syn) and IFP of Prg4ER/Td Scx-GFP knee joints at 1 wk postsurgery. (E) Representative flow cytometry images show the gating strategy for Td+ cells within Cd45− cells from Syn/IFP collected at 1 wk postsurgery. (F) The percentage of Td+ cells within Cd45− cells was quantified. ^**, P < .01 vs control. (G) Representative fluorescence images of EdU incorporation in the synovium at 1 wk postsurgery. Arrows and arrow heads point to EdU+Td− and EdU+Td+ cells, respectively. (H) Quantification of the percentage of EdU+ cells in Td+ cells. ^**, P < .01 vs control. (I) Quantification of Td− and Td+ CFU-F colonies per knee at 1 wk postsurgery. ^***, P < .001 vs Td−. (J) CFU-F frequency of sorted Td− and Td+ cells from Syn/IFP at 1 wk postsurgery was quantified. ^***, P < .001 vs Td−. (K) Bright field (BF) and fluorescence light (FL) images of CFU-F colonies formed by Td− and Td+ cells. (L) The growth curves of Td− and Td+ cells in the culture were measured. ^***, P < .001 vs Td−. F, femur; T, tibia; M, meniscus; Syn, synovium; IFP, infrapatellar fat pad. Scale bars: 200 μm.