Hnrnpk protects against osteoarthritis through targeting WWC1 mRNA and inhibiting Hippo signaling pathway

Abstract

Osteoarthritis (OA) is an age-related or post-traumatic degenerative whole joint disease characterized by the rupture of articular cartilage homeostasis, the regulatory mechanisms of which remain elusive. This study identifies the essential role of heterogeneous nuclear ribonucleoprotein K (hnRNPK) in maintaining articular cartilage homeostasis. Hnrnpk expression is markedly downregulated in human and mice OA cartilage. The deletion of Hnrnpk effectively accelerates the development of post-traumatic and age-dependent OA in mice. Mechanistically, the KH1 and KH2 domain of Hnrnpk bind and degrade the mRNA of WWC1. Hnrnpk deletion increases WWC1 expression, which in turn leads to the activation of Hippo signaling and ultimately aggravates OA. In particular, intra-articular injection of LPA and adeno-associated virus serotype 5 expressing WWC1 RNA interference ameliorates cartilage degeneration induced by Hnrnpk deletion, and intra-articular injection of adeno-associated virus serotype 5 expressing Hnrnpk protects against OA. Collectively, this study reveals the critical roles of Hnrnpk in inhibiting OA development through WWC1-dependent downregulation of Hippo signaling in chondrocytes and defines a potential target for the prevention and treatment of OA.

Keywords: Hippo signaling; HnRNPK; WWC1; chondrocytes; osteoarthritis.

Graphical abstract

Figure 1

Hnrnpk is drastically reduced in OA cartilage of human and mice (A) Representative images of safranin O/fast green staining and immunohistochemistry of hnRNPK in OA and normal human cartilage (n = 6 biological replicates). Scale bars, 100 μm. (B) Immunoblotting of COL2A1, hnRNPK, and GAPDH in three normal and three OA human knee joint cartilage tissues. (C) Representative images of safranin O/fast green staining and immunohistochemistry of Hnrnpk in chondrocytes of sham-operated and DMM-operated mice at 8 and 14 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (D) Representative images of safranin O/fast green staining and immunohistochemistry of Hnrnpk in chondrocytes of 3-, 6-, 11-, and 25-month-old wild-type mice (n = 6 biological replicates). Scale bars, 100 μm. (E) Immunoblotting of COL2A1, P-P65, MMP13, Hnrnpk, and GAPDH in mice primary articular chondrocytes treated with IL-1β (10 ng/mL or 20 ng/mL) for 72 h. (F) qRT-PCR of Hnrnpk, ACAN, COL2A1, and MMP13 in mice primary articular chondrocytes treated with IL-1β (10 ng/mL or 20 ng/mL) for 72 h (n = 3 biological replicates). ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (D, F) or two-tailed unpaired Student t test (C). Data are shown as means ± SD.

Figure 2

Hnrnpk deficiency causes spontaneous OA and accelerates the development of post-traumatic OA in mice (A) Experimental scheme. A total of 12 male, 2-month-old CKO and WT mice were used in the experiment. After five daily injections of tamoxifen, all mice were harvested at 7 months of age (n = 6 biological replicates). (B) Hnrnpk protein and mRNA expression in response to 4-hydroxytamoxifen (4-OHT, 1 μM) or solvent treatment in primary articular chondrocytes from CKO mice, as determined by western blotting and qRT-PCR (n = 3 biological replicates). (C) Representative images of safranin O/fast green staining and immunohistochemistry of Hnrnpk, MMP13, ADAMTS4, and Aggrecan in chondrocytes of WT and CKO mice at 7 months of age. Scale bars, 100 μm. (D) Quantitative analysis of the OARSI score, Hnrnpk-positive, MMP13-positive, ADAMTS4-positive, and Aggrecan-positive chondrocytes in WT and CKO mice (n = 6 biological replicates). (E) Representative images of safranin O/fast green staining and immunohistochemistry of MMP13, ADAMTS4, and Aggrecan in chondrocytes of WT and CKO mice at 8 weeks after sham or DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (E) or two-tailed unpaired Student t test (B and D). Data are shown as means ± SD. NS, not significant.

Figure 3

Hnrnpk deletion leads to activation of Hippo signaling (A) Volcano map of RNA-seq data, n = 3 for each group. (B) KEGG enrichment of pathways associated with significantly upregulated genes (p < 0.05) in the Hnrnpk knockout versus control groups. (C) GSEA of the enrichment of Hippo signaling pathway in RNA-seq. (D) Immunoblotting of indicated protein in WT mice primary articular chondrocytes transfected with Ad-GFP or Ad-Cre. (E) qRT-PCR of CYR61 and CTGF in WT mice primary articular chondrocytes transfected with Ad-GFP or Ad-Cre (n = 3 biological replicates). (F) Representative images of immunohistochemistry of CTGF and YAP in chondrocytes of 7-month-old WT and CKO mice (n = 6 biological replicates). Scale bars, 100 μm. (G) Representative images of immunohistochemistry of CTGF and YAP in chondrocytes of WT and CKO mice at 8 weeks after sham or DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. ∗∗p < 0.01, ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (G) or two-tailed unpaired Student t test (E and F). Data are shown as means ± SD.

Figure 4

Suppressing Hippo signaling reverses accelerated post-traumatic OA development induced by Hnrnpk ablation (A) Experimental scheme. A total number of 24 male, 8-week-old mice were used in the experiment. After five daily injections of tamoxifen, 10-week-old CKO and WT mice were subjected to DMM surgery. One week after DMM surgery, WT and CKO mice were subjected to intra-articular injection of LPA (25 μg/mice) or vehicle. All mice were harvested at 18 weeks of age (n = 6 biological replicates). (B) Representative images of safranin O/fast green staining and OARSI scores of WT and CKO mice received LPA or vehicle injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (C) Representative images of immunohistochemistry of MMP13, Aggrecan, YAP, and CTGF in chondrocytes of WT and CKO mice received LPA or vehicle injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (D) Micro-CT scans of knee joints from WT and CKO mice received LPA or vehicle injection at 8 weeks after DMM surgery (n = 6 biological replicates). Red arrows, medial tibial subchondral bone. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (B–D). NS, not significant.

Figure 5

Hnrnpk binds and degrades the mRNA of WWC1, a Hippo signaling upstream effector (A) Heatmap of genes with upregulated expression of Hippo signaling pathway in the Hnrnpk knockout versus control groups obtained by GSEA. Red arrows, WWC1 mRNA. (B) qRT-PCR and immunoblotting of WWC1 in WT mice primary articular chondrocytes transfected with Ad-GFP or Ad-Cre. (C) Representative images of immunohistochemistry of WWC1 in chondrocytes of WT and CKO mice at 8 weeks after sham or DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (D) Quality check of RIP. (E) RIP-PCR indicated significant binding between Hnrnpk and WWC1 mRNA. (F) RIP-qPCR indicated significant binding between Hnrnpk and WWC1 mRNA (n = 3 biological replicates). (G) Half-lives of WWC1 mRNA in chondrocytes with or without Hnrnpk (n = 3 biological replicates). (H) Schematic diagram of plasmid construction. The Hnrnpk mutant plasmid lacking the KH1, KH2, and KH3 RNA-binding domains was ΔHnrnpk-mutant all. The Hnrnpk overexpression plasmid was ΔHnrnpk-full length. The Hnrnpk mutant plasmids lacking KH1, KH2, or KH3 was ΔHnrnpk-KH1, ΔHnrnpk-KH2, or ΔHnrnpk-KH3. (I) Immunoblotting of Flag protein in primary articular chondrocytes without plasmid transfection or transfected with vector, ΔHnrnpk-full length, ΔHnrnpk-Mutant all, ΔHnrnpk-KH1, ΔHnrnpk-KH2, or ΔHnrnpk-KH3 plasmid. (J) qRT-PCR of WWC1 mRNA in WT primary articular chondrocytes transfected with vector, ΔHnrnpk-full length, ΔHnrnpk-mutant all, ΔHnrnpk-KH1, ΔHnrnpk-KH2, or ΔHnrnpk-KH3 plasmid after transfection with Ad-GFP or Ad-Cre. (K) Immunoblotting of WWC1 and GAPDH protein in WT primary articular chondrocytes transfected with vector, ΔHnrnpk-full length, ΔHnrnpk-Mutant all, ΔHnrnpk-KH1, ΔHnrnpk-KH2, or ΔHnrnpk-KH3 plasmid after transfection with Ad-GFP or Ad-Cre. ∗p < 0.05, ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (C, J) or two-tailed unpaired Student t test (B, F, and G). Data are shown as means ± SD. NS, not significant.

Figure 6

Hnrnpk modulates Hippo signaling and OA development by interacting with the transcript of WWC1 (A) Experimental scheme. A total of 24 male, 8-week-old mice were used in the experiment. After five daily injections of tamoxifen, 10-week-old CKO and WT mice were subjected to DMM surgery. One week after DMM surgery, WT and CKO mice were subjected to intra-articular injection of AAV5-WWC1-RNAi (5 × 10⁹ particles in 10 μL) or AAV5-EGFP-RNAi (5 × 10⁹ particles in 10 μL). All mice were harvested at 18 weeks of age (n = 6 biological replicates). (B) GFP signals were strongly detected in articular cartilage 1 month after intra-articular injection of AAV5-EGFP-RNAi. Scale bar, 100 μm. (C) Representative images of safranin O/fast green staining and OARSI scores of WT and CKO mice received AAV5-WWC1-RNAi or AAV5-EGFP-RNAi injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (D) Representative images of immunohistochemistry of MMP13, ADAMTS4, and Aggrecan in chondrocytes of WT and CKO mice received AAV5-WWC1-RNAi or AAV5-EGFP-RNAi injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (E) Micro-CT scans of knee joints from WT and CKO mice received AAV5-WWC1-RNAi or AAV5-EGFP-RNAi injection at 8 weeks after DMM surgery (n = 6 biological replicates). Red arrows, medial tibial subchondral bone. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (C–E). Data are shown as means ± SD. NS, not significant.

Figure 7

Overexpression of Hnrnpk in cartilage protects against OA (A) Experimental scheme. A total of 24 male, 10-week-old WT mice were used in the experiment. We subjected 10-week-old WT mice to DMM or sham surgery. One week after DMM or sham surgery, WT mice were subjected to intra-articular injection of AAV5-Hnrnpk (5 × 10⁹ particles in 10 μL) or AAV5-EGFP-control (5 × 10⁹ particles in 10 μL). All mice were harvested at 18 weeks of age (n = 6 biological replicates). (B) GFP signals were strongly detected in articular cartilage one month after intra-articular injection of AAV5-EGFP-control. Scale bar, 100 μm. (C) Representative images of safranin O/fast green staining and OARSI scores of sham-operated and DMM-operated mice received AAV5-Hnrnpk or AAV5-EGFP-control injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (D) Representative images of immunohistochemistry of Hnrnpk, MMP13, ADAMTS4, and Aggrecan in chondrocytes of sham-operated and DMM-operated mice received AAV5-Hnrnpk or AAV5-EGFP-control injection at 8 weeks after DMM surgery (n = 6 biological replicates). Scale bars, 100 μm. (E) Micro-CT scans of knee joints from sham-operated and DMM-operated mice received AAV5-Hnrnpk or AAV5-EGFP-control injection at 8 weeks after DMM surgery (n = 6 biological replicates). ∗∗p < 0.01, ∗∗∗p < 0.001. p values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (C–E). Data are shown as means ± SD. NS, not significant.

Figure 8

Diagram of Hnrnpk protecting against OA through targeting WWC1 mRNA and inhibiting Hippo signaling pathway In articular chondrocytes, the KH1 and KH2 domains of Hnrnpk bound and degraded the mRNA of WWC1. Hnrnpk deletion increased WWC1 expression, which in turn led to the activation of Hippo signaling and ultimately aggravated OA.