Gene Expression Signatures of Synovial Fluid Multipotent Stromal Cells in Advanced Knee Osteoarthritis and Following Knee Joint Distraction

Abstract

Osteoarthritis (OA) is the most common musculoskeletal disorder. Although joint replacement remains the standard of care for knee OA patients, knee joint distraction (KJD), which works by temporarily off-loading the joint for 6-8 weeks, is becoming a novel joint-sparing alternative for younger OA sufferers. The biological mechanisms behind KJD structural improvements remain poorly understood but likely involve joint-resident regenerative cells including multipotent stromal cells (MSCs). In this study, we hypothesized that KJD leads to beneficial cartilage-anabolic and anti-catabolic changes in joint-resident MSCs and investigated gene expression profiles of synovial fluid (SF) MSCs following KJD as compared with baseline. To obtain further insights into the effects of local biomechanics on MSCs present in late OA joints, SF MSC gene expression was studied in a separate OA arthroplasty cohort and compared with subchondral bone (SB) MSCs from medial (more loaded) and lateral (less loaded) femoral condyles from the same joints. In OA arthroplasty cohort (n = 12 patients), SF MSCs expressed lower levels of ossification- and hypotrophy-related genes [bone sialoprotein (IBSP), parathyroid hormone 1 receptor (PTH1R), and runt-related transcription factor 2 (RUNX2)] than did SB MSCs. Interestingly, SF MSCs expressed 5- to 50-fold higher levels of transcripts for classical extracellular matrix turnover molecules matrix metalloproteinase 1 (MMP1), a disintegrin and metalloproteinase with thrombospondin motifs 5 (ADAMTS5), and tissue inhibitor of metalloproteinase-3 (TIMP3), all (p < 0.05) potentially indicating greater cartilage remodeling ability of OA SF MSCs, compared with SB MSCs. In KJD cohort (n = 9 patients), joint off-loading resulted in sustained, significant increase in SF MSC colonies' sizes and densities and a notable transcript upregulation of key cartilage core protein aggrecan (ACAN) (weeks 3 and 6), as well as reduction in pro-inflammatory C-C motif chemokine ligand 2 (CCL2) expression (weeks 3 and 6). Additionally, early KJD changes (week 3) were marked by significant increases in MSC chondrogenic commitment markers gremlin 1 (GREM1) and growth differentiation factor 5 (GDF5). In combination, our results reveal distinct transcriptomes on joint-resident MSCs from different biomechanical environments and show that 6-week joint off-loading leads to transcriptional changes in SF MSCs that may be beneficial for cartilage regeneration. Biomechanical factors should be certainly considered in the development of novel MSC-based therapies for OA.

Keywords: chondrocytes; knee joint distraction; multipotent stromal cells; osteoarthritis; subchondral bone; synovial fluid.

FIGURE 1

Phenotypical and functional analyses of multipotent stromal cells (MSCs). (A) Comparison of growth rates of medial and lateral subchondral bone (SB) MSCs (SB MSC M and SB MSC L, respectively) and synovial fluid (SF) MSCs measured as population doubling (PD) times (in days). Dots represent individual donors; error bars represent interquartile range (IQR). (B) Phenotypic profile of culture-expanded SF and medial and lateral SB MSCs indicating no differences in the expression of standard MSC markers. Horizontal bars show means and error bars represent standard error (n = 3 donors). (C) Example images of SF MSC adipogenesis assay (Adipog) stained with Oil Red O after 14 days of adipogenic induction (top panel, original magnification ×100), osteogenesis (Osteog) showing positive alizarin red staining on day 14 post osteogenic induction (medium panel, original magnification ×100), and toluidine blue staining for glycosaminoglycan (GAG) deposition (purple) of chondrogenic (Chondrog) pellet cultures on day 21 post induction (bottom panel, original magnification ×40). (D) Analysis of SF MSC motility showing no significant differences between MSCs. Top panel: graph representing the median percentage of wound uncovered by migrating cells after 24 h (relative to the corresponding 0-h area); data represent individual measurements from nine SB MSC donors and three SF MSC donors; error bars represent IQR. Bottom panels: example images showing uncovered area measured at time 0 and at 24 h (t24).

FIGURE 2

Gene expression analysis of culture-expanded medial and lateral subchondral bone (SB) multipotent stromal cells (MSCs) (SB MSC M and SB MSC L, respectively), synovial fluid (SF) MSCs, and medial and lateral chondrocytes (CH M and CH L, respectively). (A) Cluster analysis between CH, SB MSCs, and SF MSCs illustrating clear clustering of MSCs away from chondrocytes. Data were normalized to the housekeeping gene HPRT and log2 transformed. Data filtering was performed according to standard methods previously described (Churchman et al., 2012). Scores were assigned as: black = 1, red > 1, green < 1; gray = missing data (below detection). (B) Differentially expressed genes with a significant higher expression in chondrocytes than in MSCs. L.D. indicates transcripts which were rarely expressed (in < 50% samples), and B.D. indicates transcripts below detection. (C) Differentially expressed genes between chondrocytes and MSCs, which were statistically significantly higher expressed in MSCs than in chondrocytes. *p < 0.05; **p < 0.01; ***p < 0.0001. Horizontal bars show medians and error bars represent interquartile range (IQR). REU, relative expression units [relative to housekeeping hypoxanthine phosphoribosyltransferase (HPRT)].

FIGURE 3

Gene expression analysis of culture-expanded medial andlateral subchondral bone (SB) multipotent stromal cells (MSCs) (SB MSC M and SB MSC L, respectively) and synovial fluid (SF) MSCs pointing toward MMP and other pathway activation in SF MSCs. (A) Genes expressed significantly higher in SB MSCs than SF MSCs. (B) Differentially expressed genes with a significant higher expression in SF MSCs than in SB MSCs. *p < 0.05; **p < 0.01; ***p < 0.0001. Horizontal bars show medians, and error bars represent interquartile range (IQR). REU, relative expression units [relative to housekeeping hypoxanthine phosphoribosyltransferase (HPRT)]. L.D. indicates transcripts which were rarely expressed (detected in <50% samples).

FIGURE 4

Synovial fluid (SF) multipotent stromal cell (MSC) colony numbers and characteristics following knee joint distraction (KJD). (A) Colonies formed by MSCs at weeks 0, 3, and 6 from three patients. (B) Quantification of colony numbers (total numbers); each color represents individual patient. (C) Colony area frequency distribution showing a shift toward larger colonies following KJD. (D) Quantification of colonies integrated density using Image J (1,432, 833, and 744 total colonies at weeks 0, 3, and 6, respectively). ***p < 0.0001; horizontal lines represent medians.

FIGURE 5

Gene expression changes in synovial fluid (SF) multipotent stromal cells (MSCs) following knee joint distraction (KJD): sustained changes. (A) Reduction in chemokine transcripts CCL2 and CCL20. (B) Reduction in bone-anabolic transcripts TNFRSF11B and IBSP. (C) Reduction in FABP4 and Sox9 transcripts. (D) Increase in ACAN transcripts. *p < 0.05; **p < 0.01; ***p < 0.0001. Horizontal bars show medians, and error bars represent interquartile range (IQR). REU, relative expression units [relative to housekeeping hypoxanthine phosphoribosyltransferase (HPRT)].

FIGURE 6

Gene expression changes in synovial fluid (SF) multipotent stromal cells (MSCs) following knee joint distraction (KJD): temporary changes. (A) Increase at week 3 and reduction at week 6 in progenitor-marker transcripts GDF5 and GREM1. (B) Similar trend in anabolic growth factor transcripts TGFBR2 and TGFBR3. (C) Increase in insulin growth factor receptor (IGFR) and reduction in vascular endothelial growth factor C (VEGFC) transcripts at week 3. *p < 0.05; **p < 0.01; ***p < 0.0001. Horizontal bars show medians, and error bars represent interquartile range (IQR). REU, relative expression units [relative to housekeeping hypoxanthine phosphoribosyltransferase (HPRT)].