Piezo1 activation accelerates osteoarthritis progression and the targeted therapy effect of artemisinin

Abstract

Introduction: Osteoarthritis (OA) is a devastating whole-joint disease affecting a large population worldwide with no cure; its mechanism remains poorly defined. Abnormal mechanical stress is the main pathological factor of OA.

Objectives: To investigate the effects of Piezo1 activation on OA development and progression and to explore Piezo1-targeting OA treatment.

Methods: The expression levels of Piezo1 were determined in human OA cartilage and experimental OA mice. Mice with genetic Piezo1 deletion in chondrocytes or intra-articular injection of the Piezo1 activator Yoda1 were utilized to determine the effects on DMM-induced OA progression. Effects of artemisinin (ART), a potent antimalarial drug, on Piezo1 activation, chondrocyte metabolism and OA lesions were determined.

Results: Piezo1 expression was elevated in articular chondrocytes in human OA and DMM-induced mouse OA cartilage. Piezo1 deletion in chondrocytes largely attenuates DMM-induced OA-like phenotypes. In contrast, intra-articular injection of Yoda1 aggravates the knee joint OA lesions in mice. PIEZO1 activation increases, while PIEZO1 siRNA knockdown decreases, expression of RUNX2 and catabolic enzymes MMP13 and ADAMTS5 in primary human articular chondrocytes in a PI3K-AKT dependent manner. We have provided strong evidence supporting that ART is a novel and potent inhibitor of Piezo1 activation in primary OA-HACs and all cell lines examined, including human endothelial HUVEC cells, ATDC5 chondrocyte-like cells and MLO-Y4 osteocytes-like cells. Results from in vitro experiments confirmed that ART decreases the Yoda1-induced increases in the levels of OA-related genes and p-PI3K and p-AKT proteins in OA-HACs and alleviates DMM-induced OA lesions in mice.

Conclusions: We establish a critical role of Piezo1 in promoting OA development and progression and define ART as a potential OA treatment.

Keywords: Artemisinin; Chondrocyte; Osteoarthritis; Piezo1.

Graphical abstract

Fig. 1

Piezo1 expression is upregulated in chondrocytes of human and mouse OA articular cartilage. (A) Representative X-ray and general views of knee joints of OA patients with total knee arthroplasty, with red boxes indicating respective healthy and OA areas. Scale bar: 1 cm. (B) Representative immunohistochemical images of safranin O/fast green (SO/FG) and PIEZO1 in OA cartilage with different degrees of injury. The images with higher magnification in different colors at the bottom are the corresponding areas in the dashed boxes of the upper images. Scale bar: 200 μm. (C) Quantitative analysis PIEZO1-positive chondrocytes as a proportion of the total chondrocytes in cartilage with different degrees of injury. N = 8 per group. (D) Representative images of SO/FG staining and immunofluorescence (IF) of PIEZO1, RUNX2, COL10A1, MMP13 and ADAMTS5 in normal and OA cartilages. Scale bar: 200 μm. (E) OARSI score of normal and OA cartilages. (F, G) Western blotting and Quantification of PIEZO1, COL2A1, ADAMTS5, MMP13 and RUNX2 in normal and OA primary articular chondrocytes. N = 3 biologically independent replicates per group. (H) Representative images of SO/FG staining and IF of Piezo1 in chondrocytes of controls and mice at 2, 4 and 8 weeks after DMM surgery, white dashed lines indicate the cartilage surfaces. Scale bar: 50 μm. (I) Quantification of PIEZO1-positive chondrocytes based on staining results in (H). N = 8 biologically independent replicates per group. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Piezo1 deletion in articular chondrocyte attenuates DMM-induced OA lesions in mice. (A) Schematic diagram illustrating the experimental design. (B,D) Representative images of IF staining of Piezo1 in articular cartilage of cKO and Control mice at 2 months after TAM induction. N = 8 biologically independent replicates per group. Scale bar: 100 μm. Quantitative data are shown in D. (C) Three-dimensional reconstruction from μCT scans of knee joints from control and cKO mice at 8 weeks after sham or DMM surgery. Scale bar, 1.0 mm. (E) The bone volume (BV) of calcified meniscus and synovial tissue was analyzed by μCT. N = 8 biologically independent replicates per group. (F) Representative images of SO/FG staining (first row) and hematoxylin-eosin(H/E) staining (second row) of knee joints, and IF staining of Adamts5, Col2a1 and Runx2 (rest rows) in chondrocytes of cKO and Controls at 8 weeks after DMM surgery. Black arrowheads show cartilage damage, green arrowheads show marked synovial hyperplasia, red arrowheads show marked hyperplastic osteophyte, white dashed lines indicate the cartilage surfaces. Scale bar: 50 μm. (G-L) OARSI score(G), osteophyte score (H), synovitis score (I) and quantification of Adamts5(J), Runx2(K) and Col2a1(L) positive chondrocytes based on staining results in (F). N = 8 biologically independent replicates per group. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Activation of Piezo1 induces metabolic disorder of chondrocytes and exacerbates OA development in mice. (A) Schematic diagram illustrating the experimental design. On the 3rd day after DMM, mice in the Yoda1 group were injected with Yoda1 dissolved in DMSO (150 μg/ml, 2 μl/time, once a week) into the knee joint cavity. Mice in the control group were injected with DMSO at the same time point in the knee joint cavity, and the samples were collected 8 weeks after DMM operation. (B) Three-dimensional reconstruction from μCT scans of knee joints from Veh and Yoda1 group mice at 8 weeks after sham or DMM surgery. Scale bar, 1.0 mm. (C) The BV of calcified meniscus and synovial tissue was analyzed by μCT. N = 8 biologically independent replicates per group. (D) Representative SO/FG staining images in articular cartilage of cKO and Control mice at 8 weeks after sham or DMM surgery. The images with higher magnification at the bottom are the corresponding areas in the dashed boxes of the upper images. Scale bar, 200 μm. (E) Representative images of IF staining of Col2a1, Runx2 and Adamts5 in chondrocytes of Yoda1 and Controls at 8 weeks after DMM surgery, white dashed lines indicate the cartilage surfaces. Scale bar, 50 μm. (F-K) OARSI score (F), osteophyte score (G), synovitis score (H) and quantitative data of expression of Runx2(I), Adamts5(J) and Col2a1(K) based on staining results in (D, E). N = 8 biologically independent replicates per group. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01.

Fig. 4

PIEZO1 activation promotes catabolism but inhibits anabolism in OA HACs in a PI3K-AKT-dependent manner. (A) KEGG pathway analysis of transcripts from experiments of si-NC- and si-PIEZO1-treated OA HACs. (B) KEGG pathway analysis of transcripts from experiments of control and Yoda1 (1 μM) treatment. (C) KEGG pathway analysis of transcripts from experiments of control and overexpression-PIEZO1 in OA HACs. (D) Western blot of PIEZO1, COL2A1, RUNX2 and MMP13 in OA HACs transfected with si-NC and si-PIEZO1 for 48 h. (E) Quantification of (D). (F) Western blot of PIEZO1, COL2A1, RUNX2 and MMP13 in OA HACs treated with Yoda1 with and without LY294002(PI3K inhibitor), SB239063(p38 MAPK inhibitor) for 24 h. (G) Quantification of (F). (H) Western blot of p-PI3K, PI3K, p-AKT and AKT in OA HACs treated with si-PIEZO1 for 48 h. (I) Quantification of (H). (J) Western blot of p-PI3K, PI3K, p-AKT and AKT in OA HACs treated with Yoda1 for 30 min. (K) Quantification of (J). All experiments were repeated at least three times independently with similar results. (L) Representative images of IF staining of p-AKT in chondrocytes of cKO, Yoda1 and Controls mice at 8 weeks after DMM surgery, white dashed lines indicate the cartilage surfaces. Scale bar, 50 μm. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01.

Fig. 5

ART suppresses the PIEZO1-mediated calcium influx and reverses OA-related gene expression induced by PIEZO1 activation. (A) Schematic diagram of experimental design. (B) Representative Fura-2 ratio (340/380) traces obtained from single-cell Ca²⁺ imaging of OA HACs in response to the indicated conditions as shown in (A) (N = 82 cells). (C) Representative Fura-2 ratio (340/380) traces obtained from single-cell Ca²⁺ imaging of HUVECs in response to the indicated conditions as shown in (A) (N = 30 cells). (D) Representative Fura-2 ratio (340/380) traces obtained from single-cell Ca²⁺ imaging of ATDC5 cells in response to the indicated conditions as shown in (A) (N = 52 cells). (E) Representative Fura-2 ratio (340/380) traces obtained from single-cell Ca²⁺ imaging of MLO-Y4 cells in response to the indicated conditions as shown in (A) (N = 40 cells). (F) Single-cell Fura-2 Ca²⁺ imaging experiments showing the average response of OA HACs treated with Yoda1 and ART in turn. (N = 82 cells). The black trace represents the average response of 3-dish repeats. Error bars are shown in colors. Note: ART at the experimental concentration did not impact the cell viability. (G) CCK8 assay was used to detect the effects of different concentrations of ART on OA HACs proliferation. The black trace represents the average response of 3-dish repeats. Error bars are shown in colors. (H) KEGG pathway analysis of transcripts from experiments of Yoda1 and Yoda1 + ART. The three biologically independent samples in each group were combined into one. (I) Western blot of PIEZO1, COL2A1, RUNX2 and MMP13 in OA HACs treated with Yoda1 with and without ART for 24 h. (J) Western blot of p-PI3K, PI3K, p-AKT and AKT in OA HACs treated with Yoda1 with and without ART for 30 min. (K-N) Quantification of (I). (O, P) Quantification of (J). All experiments were repeated at least three times independently with similar results. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01.

Fig. 6

ART ameliorates OA lesions caused by Piezo1 activation in mice. (A) Schematic diagram illustrating the experimental design. On the 3rd day after DMM, mice in the Yoda1 group were injected with Yoda1 dissolved in DMSO (150 μg/ml, 2 μl/time, once a week) into the knee joint cavity, and the mice in the ART group were injected into the joint cavity with ART dissolved in DMSO (1 mM, 2 μl/time, once a week) 7 days after DMM operation. The control group mice were injected with DMSO at the same time point in the knee joint cavity, and the samples were collected 8 weeks after DMM operation. (B) Representative 3D reconstruction derived from μCT scans of knee joints from Yoda1-treated mice and controls with or without ART treatment at 8 weeks after DMM surgery. Scale bar, 1.0 mm. (C) Representative H/E staining images in articular cartilage of four group mice at 8 weeks after DMM surgery. Scale bar, 50 μm. (D) Representative SO/FG staining images in articular cartilage of four group mice at 8 weeks after DMM surgery. The enlarged images at the bottom are the corresponding areas in the dashed boxes of the upper images. Scale bar, 50 μm. (E) Representative IF staining images of Col2a1, Runx2, Adamts5 in chondrocytes of DMM joint sections, white dashed lines indicate the cartilage surfaces. Scale bar, 50 μm. (F) Results of Von Frey test. (G) The BV of calcified meniscus and synovial tissue was analyzed by μCT. N = 8 biologically independent replicates per group. (H-L) OARSI score(H), osteophyte score (I), synovitis score (J) and quantification of Runx2 (K) and Adamts5 (L) positive chondrocytes based on SO/FG and IF staining results in (D, E). N = 8 biologically independent replicates per group. Results were expressed as mean ± standard deviation (s.d.). **P < 0.01.