The inhibition of EZH2 ameliorates osteoarthritis development through the Wnt/β-catenin pathway

Abstract

The purpose of our study was to elucidate the role of the histone methyltransferase enhancer of zeste homologue 2 (EZH2) in the pathophysiology of osteoarthritis (OA) and to develop a strategy to modulate EZH2 activity for OA treatment. The expression of EZH2 in normal and OA human cartilage was compared by western blotting. The effect of EZH2 overexpression and inhibition on chondrocyte hypertrophy related gene expression was examined by real-time PCR, and histone methylation on the promoter of the Wnt inhibitor SFRP1 was analyzed using a chromatin immunoprecipitation (ChIP) PCR. Histological assessment of OA mice joint was carried out to assess the in vivo effects of EZH2 inhibitor EPZ005687. We found EZH2 level was significantly increased in the chondrocytes of OA patients compared to normal humans. Overexpression of EZH2 promoted Indian Hedgehog, MMP-13, ADAMTS-5 and COLX expression, while inhibition of EZH2 reversed this trend. Furthermore, the induction of EZH2 led to β-catenin signaling activation by increasing H3K27me3 on the promoter of SFRP1, while the inhibition of EZH2 silenced β-catenin signaling. Finally, intraarticular injection of EPZ005687 delayed OA development in mice. These results implicated EZH2 activity in OA development. Pharmacological inhibition of EZH2 may be an effective therapeutic approach for osteoarthritis.

Figure 1. EZH2 expression in osteoarthritis (OA) and normal articular chondrocytes.

(A) HE staining showed smooth surfaces in human articular cartilage and rough surfaces in OA human articular cartilage. Scale bars, 50 μm. (B) Real-time PCR analysis revealed higher expression of EZH2 mRNA in OA chondrocytes (n = 8) cultivated from OA cartilage sample compared to normal chondrocytes (n = 10). *p < 0.05. (C) A representative comparison of the EZH2 level of human normal and OA chondrocytes by western blot. (D) Quantitative analysis of western blot confirmed a significant upregulation of EZH2 expression in OA chondrocytes compared with normal chondrocytes. Values are means ± SD, *p < 0.05.

Figure 2. Modulation of EZH2 expression controlled hypertrophy and matrix degradation-related gene expression.

(A) EZH2 expression increased gradually in normal chondrocytes (P2) after being treated with IL-1β (10 ng/ml) for 0, 12 and 24 hours demonstrated by Western blotting. (B) Successful overexpression of EZH2 in normal chondrocytes was validated by Western blotting analysis. Meanwhile, knockdown of EZH2 by siEZH2 greatly inhibited the enhanced expression of EZH2 which was induced by IL-1β treatment in normal chondrocytes. (C–G) Effect of overexpression and knockdown of EZH2 in normal chondrocytes on Indian Hedgehog, COLX, MMP-13, ADAMTS-4 and ADAMTS-5 expression was demonstrated by Realtime RT-PCR. Compared with control group, the expression of these catabolic genes was significantly increased by Lenti-EZH2. Similarly, IL-1β treatment for 24 h also increased these catabolic genes but the effect was weaker than Lenti-EZH2. Finally, knockdown of EZH2 greatly reduced the IL-1β-induced increase of these catabolic genes. Values are means ± SD (n = 3). *p < 0.05, **p < 0.01.

Figure 3. Expression of EZH2 influenced SOX9 expression in normal chondrocytes.

(A) Normal chondrocytes (P2) were infected with Lenti-EZH2. A significant increase of EZH2 and a obvious decrease of SOX9 were demonstrated by Western-blotting. (B,C) A quantitative analysis of EZH2 and SOX9 protein expression confirmed the Western-blotting results. (D) The normal chondrocytes (P2) were treated with IL-1β (10 ng/ml) for 24 hours to induce chondrocyte pathological change followed by knockdown by siEZH2 or treatment by pharmacological inhibitor EPZ005687. The effect of each treatment was investigated by Western-blotting and quantitative analysis (E,F). IL-1β treatment notably enhanced EZH2 expression and inhibited SOX9 expression. Subsequently, knockdown of EZH2 by siEZH2 downregulated EZH2 expression and ameliorate the inhibition of SOX9. Meanwhile, the inhibition of SOX9 was also attenuated following treatment with the EZH2 inhibitor EPZ005687 (5.6 μmol/L). Values are means ± SD (n = 3). *p < 0.05, **p < 0.01.

Figure 4. EZH2 induced pathologic changes in chondrocytes partially through regulation ofWnt/β-catenin signaling pathway.

(A) Treatment with IL-1β (10 ng/ml) for 24 hours induced EZH2 expression and increased β-catenin expression in normal chondrocytes (P2). Meanwhile, overexpression of EZH2 by Lenti-EZH2 induced much higher increase of β-catenin expression in normal chondrocytes than IL-1β treatment. (B) Immunofluorescence of EZH2 revealed that β-catenin was enriched in the nuclei of normal chondrocytes transfected with EZH2, while mainly localised in cytoplasm and on cell membrane in chondrocytes transfected with vector. (C) Quantitative analysis confirmed β-catenin was enriched in the nuclei of chondrocytes transfected with EZH2. (D,E) EZH2 overexpression resulted in Wnt downstream signalling Axin2 and LEF-1 activation. (F) EZH2 overexpression led to the elevation of global EZH2 expression in the chondrocytes. (G) Cytocol/nuclear separation was undertaken to assess the enrichment of β-catenin in the nuclei of EZH2-overexpressing chondrocytes. The upregulation level of β-catenin localized in nucleus was more obvious than that remained in cytosol after infection of Lenti-EZH2. Values are means ± SD (n = 3), *p < 0.05, **p < 0.01.

Figure 5. Wnt inhibitor SFRP1 was inhibited by EZH2 in OA chondrocytes.

(A,B) The mRNA level and protein expression of SFRP1 were significantly higher in the normal chondrocytes than in OA chondrocytes demonstrated by real-time-PCR and western blot, respectively. (C,D) SFRP1 expression level was decreased when normal chondrocytes were treated with IL-1β as determined by real-time-PCR and Western blotting. (E,F) The inhibited mRNA and protein expression of SFRP1 induced by Lenti-EZH2 infection in normal chondrocytes were partly relieved by EZH2 inhibitor EPZ005687, while greatly enhanced by siEZH2 infection. Values are means ± SD (n = 3), *p < 0.05, **p < 0.01.

Figure 6. The state of H3K27 trimethylation at the promoter of SFRP1 was determined by the binding affinity of EZH2.

(A) Western-blotting analysis revealed that the global state of H3K27 trimethylation was significantly elevated in chondrocytes infected with Lenti-EZH2 compared with control. (B) The promoter of SFRP1 was ChIP-ed with anti-EZH2 antibody or IgG control. OA chondrocytes showed a higher level of EZH2 occupation at the promoter of SFRP1. (C) The promoter of SFRP1 was ChIP-ed with anti-H3K27me3 antibody or IgG control. The repressive trimethylation of H3K27 at the SFRP1 promoter was higher in OA chondrocytes compared with normal chondrocytes. (D) Knockdown of EZH2 by siEZH2 or EPZ005687 reduced EZH2 binding to the promoters of SFRP1 in OA chondrocytes. (E) Inhibition of EZH2 by siEZH2 or EPZ005687 decreased H3K27me3 levels at the promoters of SFRP1 in OA chondrocytes. Values are means ± SD (n = 3), *p < 0.05, **p < 0.01.

Figure 7. Inhibition of EZH2 activated SFRP1 in OA chondrocytes and delayed OA development in vivo.

(A) EPZ005687, a strong inhibitor of EZH2, enhanced the activity of SFRP1 in human OA chondrocytes shown by Realtime PCR. (B) EPZ005687 revived the inhibited SFRP1 in OA chondrocytes shown by Westernblot. (C) EPZ005687 delayed OA development demonstrated by SO staining. Scale bars, 200 & 50 μm. (D) OARSI scoring of OA severity upon EPZ005687 treatment. The in vivo experiments clearly demonstrated the therapeutic effects of EPZ005687 on OA cartilage degeneration in mice. (E–J) The cartilage samples from sham, OA + EZP205687 and OA + NS group were examined for the expression of COLX, ADAMTS-5, MMP3, MMP13, SFRP1 and β-catenin by Realtime PCR. The results confirmed the protective effect of EPZ005687 on chondrocytes. Values are means ± SD (n = 3). *p < 0.05, **p < 0.01.