Matrix-bound nanovesicles alleviate particulate-induced periprosthetic osteolysis

Abstract

Aseptic loosening of orthopedic implants is an inflammatory disease characterized by immune cell activation, chronic inflammation, and destruction of periprosthetic bone, and is one of the leading reasons for prosthetic failure, affecting 12% of total joint arthroplasty patients. Matrix-bound nanovesicles (MBVs) are a subclass of extracellular vesicle recently shown to mitigate inflammation in preclinical models of rheumatoid arthritis and influenza-mediated "cytokine storm." The molecular mechanism of these anti-inflammatory properties is only partially understood. The objective of the present study was to investigate the effects of MBV on RANKL-induced osteoclast formation in vitro and particulate-induced osteolysis in vivo. Results showed that MBV attenuated osteoclast differentiation and activity by suppressing the NF-κB signaling pathway and downstream NFATc1, DC-STAMP, c-Src, and cathepsin K expression. In vivo, local administration of MBV attenuated ultrahigh molecular weight polyethylene particle-induced osteolysis, bone reconstruction, and periosteal inflammation. The results suggest that MBV may be a therapeutic option for preventing periprosthetic loosening.

Fig. 1.. MBV characterization and assessment of cytocompatibility.

(A) TEM imaging of MBV isolated from porcine urinary bladder matrix (UBM). Scale bar, 100 nm. (B) MBV size as measured by nanoparticle tracking analysis (NTA). (C) RAW264.7 cells were treated with phosphate-buffered saline (PBS), 5% dimethyl sulfoxide (DMSO), and serial dilutions of MBV ranging from 1.25 × 10⁹ to 1 × 10¹⁰ particles/ml (MBV versus cells ratio 1.25 × 10⁵ to 1 × 10⁶ particles/ml per cell) for 24 hours and assayed for cytotoxicity by live/dead staining. Representative fluorescence images. Scale bars, 100 μm. (D) Percentage of live cells. PBS control, 5% DMSO, and various doses of MBV. Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05. (E) Cell proliferation and cytotoxicity were assessed by cell counting kit 8 (CCK-8) assay. Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05. OD₄₅₀, optical density at 450 nm.

Fig. 2.. MBVs mitigate RANKL-induced osteoclast formation and activity in vitro.

(A) Cellular uptake of MBV by RAW 264.7 cells. CFSE-labeled (green) MBV were added to the RAW 264.7 cells culture media for 1 hour; nuclei were labeled with 4′,6-diamidino-2-phenylindole (DAPI; blue), and cytoskeleton F-actin was labeled with 594-phalloidin (orange) before imaging. Representative IF images (40× and enlarged). Scale bars for 40× images, 50 μm; enlarged images, 10 μm. (B) MBVs reduce osteoclasts differentiation. RAW 264.7 cells were cultured with RANKL (30 ng/ml) and serial doses of MBV ranging from 0 to 5 × 10⁹ MBV/ml for 5 days. Osteoclasts were determined with tartrate-resistant acid phosphatase (TRAP) staining. Representative light microscope TRAP staining images (10× and enlarged). Scale bars, 200 μm. (C) Raw 264.7 cells were cultured with RANKL (30 ng/ml) with or without MBV for 5 days. Actin ring formation was determined with FITC-phalloidin, and nuclei were labeled with DAPI. Representative immunofluorescence images (1× and 40×). Scale bars for 1× images, 2000 μm; 40× images, 100 μm. (D) MBVs reduce osteoclasts activity. RAW 264.7 cells were cultured with RANKL (30 ng/ml) and serial doses of MBV ranging from 0 to 5 × 10⁹ MBV/ml on a calcium phosphate–coated plate for 7 days. Resorption areas were determined with a light microscope (5×). Scale bars, 500 μm. (E) Numbers of osteoclasts (TRAP positive, nuclei > 3). Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05. (F) Average cell area (n = 4). Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05. (G) At day 6, bone resorption assay culture media were tested for fluorescence intensity. Fluorescence intensity (n = 4). Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05. (H) Resorption pit area (n = 4). Values are shown as mean ± SD (n = 4). Significant differences were determined as *P < 0.05.

Fig. 3.. MBVs mitigate RANKL-induced osteoclast formation and functions by suppressing NF-κB pathway in vitro.

RAW 264.7 cells were challenged with RANKL and various concentrations of MBV (1× 10¹⁰ to 2× 10¹⁰ particles/ml, MBV versus cells ratio 1× 10⁴ to 2 × 10⁴ particles/ml per cell), 1 day for osteoclast differentiation and activity–related gene RT-qPCR (A and B; RANKL, 50 ng/ml), 3 days for osteoclast differentiation and activity–related protein Western blot (C; RANKL, 50 ng/ml), and 5 days for TNF-α concentration in conditional media (D; RANKL, 30 ng/ml; MBV, 5× 10⁹ particles/ml). (A) Osteoclast differentiation regulators (NFATc1 and DC-STAMP) relative expression (2^−△△T) were measured by RT-qPCR (values = means ± SD; n = 9; significant differences, *P < 0.05). (B) Osteoclast activity–related genes (cathepsinK, c-Src, MMP-9, and β3-integrin) relative expressions (2^−△△T) were measured by RT-qPCR (values = means ± SD; n = 9; significant differences, *P < 0.05). (C) Osteoclast differentiation–related protein (NFATc1) and osteoclast activity proteins (c-Src and cathepsin K) relative expression was measured by Western blot (values = means ± SD; n = 3; significant differences, *P < 0.05). (D) TNF-α in conditional media from Raw 264.7 cells cocultured with RANKL with or without MBV for 5 days was determined by ELISA (values = means ± SD; n = 3; significant differences, *P < 0.05). (E) RAW 264.7 cells were challenged with RANKL (50 ng/ml) and MBV (2 × 10¹⁰ particles/ml) from 0 to 60 min. NF-κB pathway proteins were determined by Western blot. Representative Western blotting images of the effects of MBV on p-p65, p65, and IĸB-α. Quantification of the ratios of band intensity of p-p65/p65 and IĸBα relative to β-actin expression (values = means ± SD; n = 3; significant differences, *P < 0.05).

Fig. 4.. MBV biodistribution in mouse model of particulate-induced calvarial osteolysis.

(A) Experiment design. Mice skull and periosteum were exposed to UHMWPE particles. Near-infrared (NIR)–labeled MBV pericalvarial administrated post-UHMWPE implantation surgery. From 0 hour to 7 days postsurgery, fluorescence signal was examined by an in vivo imagine system (IVIS). (B) Representative IVIS images from 0 hour to 7 days. (C) Seven days postinjection, mice were euthanized, and organs and bones were taken for IVIS imaging. Representative organs and bone IVIS images. (D) Total radiant efficacy ([p/s]/[uW/cm²]) of the calvarial region from 0 hour to 7 days (values = means ± SD; n = 5). (E) Total radiant efficacy ([p/s]/[uW/cm²]) of skull, liver, kidneys, and femur-tibia (values = means ± SD; n = 5; significant differences, *P < 0.05).

Fig. 5.. Local administration of MBV alleviates osteolysis and bone remodeling in mouse model of UHMWPE particulate-induced osteolysis.

(A) Experimental design and treatment regimen. (B) Mice were euthanized to harvest skull specimens at day 28. Representative morphology images of mice skull. Scale bars, 10 mm. (C) Representative images of mice skull dorsal/ventral micro-CT imaging 3D reconstruction. Scale bars, 2 mm. (D) Representative images of coronal cross section of micro-CT scan of mouse skull. Scale bars, 2 mm. (E) Quantification of cortical porosity (Ct.Po) and bone volume versus tissue volume (BV/TV) of skull. (values = means ± SD; n = 5; significant differences, *P < 0.05).

Fig. 6.. MBV alleviates inflammatory and osteoclastic activities in UHMWPE particulate-induced osteolysis.

(A) Representative images of mice calvarial bone coronal paraffin sections (10× and 20×) hematoxylin and eosin (H&E) staining. Scale bars for 10× images, 500 μm; for 20× images, 200 μm. (B) Representative images of mice calvarial bone paraffin sections (20× and 40×) TRAP staining. Scale bars for 20× images, 200 μm; for 40× images, 100 μm. (C) Representative images of calvarial bone coronal paraffin sections (20× and 40×) TNF-α immunohistology [immunohistochemistry (IHC)] staining. Scale bars for 20× images, 100 μm; for 40× images, 50 μm. (D) Periosteal thickness of H&E staining images. Values are shown as mean ± SD (n = 5); significant differences, *P < 0.05. (E) TRAP⁺ cells numbers per 20× images. Values are shown as mean ± SD (n = 5); significant differences, *P < 0.05. (F) TNF-α cells percentage per 20× images. Values are shown as mean ± SD (n = 5); significant differences, *P < 0.05.

Fig. 7.. Schematic illustration of the mechanism by which MBVs reduce particulate-induced osteolysis.

MBVs were derived from a porcine urinary bladder extracellular matrix (UBM) bioscaffold. In vitro, MBVs reduce monocyte/macrophage differentiation to osteoclast by suppressing the NF-κB pathway and its downstream NFATc1, DC-STAMP, c-Src, and cathepsin K expression. In vivo, MBVs alleviate UHMWPE particle-induced osteolysis in a murine calvarial osteolysis model.