Matrix metalloproteinase-10 promotes kidney fibrosis by transactivating β-catenin signaling

Abstract

Kidney fibrosis is characterized by excessive accumulation of extracellular matrix (ECM) and serves as a hallmark of chronic kidney disease (CKD). The turnover of ECM is controlled by a family of matrix metalloproteinases (MMPs), endopeptidases that play a crucial role in ECM remodeling and other cellular processes. In this study, we demonstrate that MMP-10 was upregulated in a variety of animal models of kidney fibrosis and human kidney biopsies from CKD patients. Bioinformatics analyses and experimental validation reveal that MMP-10 activated β-catenin in a Wnt-independent fashion. Knockdown of endogenous MMP-10 expression in vivo inhibited β-catenin activation and ameliorated kidney injury and fibrotic lesions, whereas over-expression of exogenous MMP-10 aggravated β-catenin activation and kidney fibrosis after injury. We found that MMP-10 cleaved and activated heparin-binding EGF-like growth factor (HB-EGF) via ectodomain shedding, leading to EGF receptor (EGFR) tyrosine phosphorylation and β-catenin transactivation via a cascade of events involving extracellular signal-regulated kinases and glycogen synthase kinase-3β. Consistently, treatment with erlotinib, a small-molecule EGFR inhibitor, effectively mitigated MMP-10-mediated kidney injury and fibrotic lesions in a dose-dependent fashion. Furthermore, β-catenin activation reciprocally upregulated the expression of MMP-10, thereby perpetuating kidney damage by forming a vicious cycle. Collectively, these results underscore that MMP-10 promotes kidney fibrosis through EGFR-mediated transactivating β-catenin in a Wnt-independent fashion. Our findings suggest that targeting MMP-10 could be a novel strategy for treatment of fibrotic CKD.

Fig. 1. Matrix metalloproteinase-10 (MMP-10) is upregulated in kidney tubular epithelium after various injuries.

A RNA sequencing revealed simultaneous induction of various MMPs in the kidney at 7 days after unilateral ureteral obstruction (UUO). Heat map depicts the mRNA expression profile of MMPs at 7 days after UUO. MMP-10 is highlighted by red color. B qRT-PCR showed renal MMP-10 mRNA expression at 7-day after UUO. Data are presented as the mean ± SEM. ***P < 0.001 versus the sham controls (n = 5). C, D Representative Western blotting (C) and quantitative data (D) showed the protein expression of MMP-10 in the kidney at different time points after UUO. Data are presented as the mean ± SEM. **P < 0.01 versus the sham controls (n = 6), ^†††P < 0.001 versus the 3-day UUO (n = 6). E, F The gelatin zymography image (E) and quantitative data (F) showed an increased MMP-10 enzymatic activity. The band near 56 kDa was indicated as MMP-10. Data are presented as the mean ± SEM. ***P < 0.001 versus the sham controls (n = 5). G, H Representative Western blotting (G) and quantitative data (H) showed the induction of MMP-10 in the kidney at different time points after unilateral ischemia-reperfusion injury (UIRI). Data are presented as the mean ± SEM. **P < 0.01 versus the sham controls (n = 6), ^†P < 0.05 versus the UIRI 3-day (n = 6), ^###P < 0.001 versus the UIRI 7-day (n = 6). I, J Zymographic analysis further showed an increased proteolytic activity of MMP-10 in UIRI. Data are presented as the mean ± SEM. ***P < 0.001 versus the sham controls (n = 5). K, L Representative Western blotting (K) and quantitative data (L) showed renal induction of MMP-10 in folic acid-induced nephropathy (FA) at 14 days after a single intraperitoneal injection of 250 mg/kg. Data are presented as the mean ± SEM. *P < 0.05 versus the sham controls (n = 5). M Representative micrographs of the immunofluorescence staining for MMP-10 protein in various CKD models as indicated. The arrow indicates positive staining in renal tubules. Scale bar, 50 µm. N Representative micrographs of the immunohistochemical staining for MMP-10 in human kidney biopsies from various CKD as indicated. Arrow indicates positive staining. Scale bar, 50 µm. O Quantitative real-time PCR analysis (qRT-PCR) demonstrated the induction of MMP-10 mRNA in human proximal tubule cells (HK-2) after treatment with TGF-β1 (2 ng/ml). Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 versus the controls (n = 4).

Fig. 2. Identification of β-catenin as the major effector mediating MMP-10 action.

A–C Western blotting showed that rhMMP-10 (100 ng/ml) induced fibronectin and α-smooth muscle actin (α-SMA) and repressed E-cadherin in HK-2 cells. Representative Western blotting (A) and quantitative data (B, C) are reported. Data are presented as the mean ± SEM. ***P < 0.001 versus the controls (n = 6). D STRING analysis revealed the protein-protein interaction (PPI) network involving MMP-10 (https://cn.string-db.org/, accessed on 8 August 2023). The interaction between MMP-10 and β-catenin was noticed. E Analysis of the Cancer Dependency Map (https://depmap.org, accessed on 10 August 2023) revealed the correlation between MMP-10 expression and alteration in β-catenin. Linear regression (left) and volcano plot (right) demonstrated a positive correlation between MMP-10 and β-catenin. F, H RhMMP-10 activated β-catenin signaling in vitro, leading to the induction of active β-catenin, PAI-1, and MMP-7 proteins. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus the controls (n = 6). I Representative micrographs showed the nuclear translocation of β-catenin after rhMMP-10 treatment for 12 h. White arrows indicate nuclear staining of β-catenin. Scale bar, 50 µm. J, K ICG-001 (10 μM) abolished MMP-10-mediated fibronectin and α-SMA expression. Data are presented as the mean ± SEM. ***P < 0.001 versus the controls (n = 6); ^†††P < 0.001 versus rhMMP-10 alone (n = 6). L Representative immunofluorescence staining showed that ICG-001 inhibited fibronectin expression and deposition induced by MMP-10. Scale bar, 50 µm. M, N ICG-001 inhibited MMP-10-mediated β-catenin activation and its downstream MMP-7 and PAI-1 induction. Representative Western blotting (M) and quantitative data (N) were presented. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus the controls (n = 6); ^††P < 0.01, ^†††P < 0.001 versus rhMMP-10 alone (n = 6). O qRT-PCR analysis showed that rhMMP-10 did not significantly affect the expression of Wnt ligands except Wnt10a. Data are presented as the mean ± SEM. *P < 0.05 versus controls (n = 6).

Fig. 3. Knockdown of endogenous MMP-10 ameliorates kidney injury and fibrosis in obstructive nephropathy.

A Experimental design. The red and green arrows show the timing of UUO surgery and plasmid injection, respectively. B, C Representative micrographs (B) and quantitative data (C) demonstrated MMP-10 protein expression in different groups as indicated. Scale bar, 50 µm. Data are presented as the mean ± SEM. **P < 0.01 versus sham controls (n = 5); ^†P < 0.05 versus UUO (n = 5). D, E Western blot analysis showed MMP-10 protein levels in various groups as indicated. Data are presented as the mean ± SEM. **P < 0.01 versus sham controls (n = 5), ^††P < 0.01 versus UUO (n = 5). F, G Western blot analysis showed the expression of several fibrosis-related proteins such as fibronectin, α-SMA, and vimentin. Representative Western blotting (F) and quantitative data (G) are presented. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^††P < 0.01, ^†††P < 0.001 versus UUO (n = 5). H, I Representative Western blotting (H) and quantitative data (I) showed KIM-1 and E-cadherin protein in various groups as indicated. Data are presented as the mean ± SEM. **P < 0.01 versus sham controls (n = 5), ^††P < 0.01 versus UUO (n = 5). J, K Representative micrographs (J) and quantitative data (K) showed immunohistochemical staining for α-SMA and KIM-1 and Sirius red staining for collagen deposition. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01 versus sham controls (n = 5); ^†P < 0.05, ^††P < 0.01 versus UUO (n = 5). Scale bar, 50 µm.

Fig. 4. MMP-10 transactivates β-catenin via HB-EGF/EGFR/ERK1/2/GSK-3β cascade.

A Representative micrographs showed the colocalization of MMP-10 (Red) and β-catenin (Green) in the obstructed kidney after UUO. Arrows indicate positive staining. Scale bar, 50 µm. B–D Knockdown of endogenous MMP-10 abolished the expression of β-catenin and its downstream genes in UUO model. Representative Western blotting results (B) and corresponding quantitative data (C, D) were shown. Data are presented as the mean ± SEM. ***P < 0.001 versus sham controls (n = 5); ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001 versus UUO injected with Ctrl-shRNA (n = 5). E, F Representative immunohistochemical staining for β-catenin further demonstrated that knockdown of MMP-10 inhibited β-catenin expression in renal tubules. Scale bar, 50 µm. Data are presented as the mean ± SEM. ***P < 0.001 versus sham controls (n = 5). ^†††P < 0.001 versus UUO injected with Ctrl-shRNA (n = 5). G Double immunofluorescence staining showed the colocalization between MMP-10 (Red) and β-catenin (Green) in patients with various CKD as indicated. Scale bar, 25 µm. H–L Western blotting showed that knockdown of MMP-10 inhibited the expression of cleaved HB-EGF, p-EGFR (Tyr845), p-ERK (Thr202/Tyr204), and p-GSK-3β (Ser9) in the obstructed kidney after UUO. Representative Western blotting (H) and quantitative data (I–L) were shown. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001 versus UUO injected with Ctrl-shRNA (n = 5). M, N Immunohistochemical staining for p-EGFR (Tyr845) demonstrated a decreased EGFR phosphorylation in renal tubules after MMP-10 knockdown. Scale bar, 50 µm. Data are presented as the mean ± SEM. ***P < 0.001 versus sham controls (n = 5), ^††P < 0.01 versus UUO injected with Ctrl-shRNA (n = 5). O Representative micrographs showed the colocalization of p-EGFR (Red) and β-catenin (Green) in UUO. Scale bar, 50 µm.

Fig. 5. Depletion of endogenous MMP-10 reduces kidney injury and β-catenin activation after ischemia-reperfusion injury.

A Experimental design. The timing of UIRI surgery, plasmid injection, unilateral nephrectomy, and sacrifice were indicated by red and black arrow or arrowhead as indicated, respectively. B, C Western blotting showed the protein level of MMP-10 in different groups as indicated. Representative Western blotting (B) and quantitative data (C) were shown. Data are presented as the mean ± SEM. **P < 0.01 versus sham controls (n = 5); ^††P < 0.01 versus UIRI injected with Ctrl-shRNA (n = 5). D Representative images showed the expression and localization of MMP-10 in the kidney after different treatments. Scale bar, 50 µm. E, F Graphic presentation showed serum creatinine (SCr) and blood urea nitrogen (BUN) levels in three groups as indicated. Data are presented as the mean ± SEM. ***P < 0.001 versus sham controls (n = 5). ^†††P < 0.001 versus UIRI injected with Ctrl-shRNA (n = 5). G, H Western blotting showed the protein levels of fibronectin, collagen I, α-SMA, and KIM-1 in different groups. Representative Western blotting (G) and quantitative data (H) were shown. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001 versus UIRI injected with Ctrl-shRNA (n = 5). I, J Knockdown of MMP-10 ameliorated kidney injury and fibrosis. Renal fibrotic lesions were evaluated using Sirius red staining and immunohistochemical staining for α-SMA. The tubular injury was assessed through immunohistochemical staining for KIM-1. Scale bar, 50 µm. Data are presented as the mean ± SEM. ***P < 0.001 versus sham controls (n = 5), ^†††P < 0.001 versus UIRI injected with Ctrl-shRNA (n = 5).

Fig. 6. MMP-10-triggered β-catenin transactivation is dependent on EGFR signaling.

A, B Blockade of EGFR signaling by small molecule inhibitor erlotinib abolished MMP-10-mediated β-catenin transactivation in vitro. Human proximal tubular epithelial cells (HK-2) were treated with rhMMP-10 in the absence or presence of erlotinib as indicated. Representative Western blotting (A) and quantitative data (B) showed the expression levels of p-EGFR (Tyr845), p-ERK1/2 (Thr202/Tyr204), p-GSK-3β (Ser9), and active β-catenin after various treatments. Data are presented as the mean ± SEM. ***P < 0.001 versus controls (n = 6); ^†††P < 0.001 versus the group with rhMMP-10 treatment alone (n = 6). C Representative micrographs showed fibronectin expression and deposition in HK-2 cells after various treatments. Scale bar, 75 µm. D, E The expression of fibronectin and α-SMA was assessed in HK-2 cells after various treatments. Representative Western blotting (D) and quantitative data (E) were presented. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus controls (n = 6); ^††P < 0.01, ^†††P < 0.001 versus rhMMP-10 treatment alone (n = 6). F In vivo experimental design. The timing of UUO surgery, plasmid injection, and oral gavage of erlotinib at dose 40 mg/kg/d or 80 mg/kg/d is indicated, respectively. G Western blotting confirmed the expression of Flag and MMP-10 after Flag-tagged MMP-10 plasmid injection. H Representative micrographs also confirmed the expression and localization of Flag and MMP-10 in three groups as indicated. Scale bar, 50 µm. Representative Western blotting (I) and quantitative data (J) showed the expression levels of p-EGFR (Tyr845), p-ERK (Thr202/Tyr204), p-GSK-3β (Ser9), and active β-catenin in different groups. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001 versus UUO plus pcDNA3.1 group (n = 5); ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus UUO plus pFlag-MMP-10 group (n = 5). K, L Representative micrographs (K) and quantitative data (L) showed immunohistochemical staining for p-EGFR (Tyr845) and β-catenin. Scale bar, 50 µm. Data are presented as the mean ± SEM. *P < 0.05, ***P < 0.001 versus sham controls (n = 5); ^††P < 0.01 versus UUO plus pcDNA3.1 group (n = 5); ^##P < 0.01, ^###P < 0.001 versus UUO plus pFlag-MMP-10 group (n = 5).

Fig. 7. Inhibition of EGFR signaling blocks exogenous MMP-10-induced fibrotic lesions in UUO.

A Immunohistochemical staining confirmed the co-localization of MMP-10, p-EGFR, and KIM-1 in serial sections of the obstructed kidney after UUO. Arrows indicate positive staining. Scale bar, 50 µm. B, C The expression levels of renal KIM-1 and MMP-7 in different groups were assessed by Western blotting. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^†P < 0.05, ^†††P < 0.001 versus UUO plus pcDNA3.1 group (n = 5); ^#P < 0.05, ^###P < 0.001 versus UUO plus pFlag^-MMP-10 group (n = 5). D, E Representative Western blotting (D) and quantitative data (E) showed renal expression of fibronectin and α-SMA in different groups. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^††P < 0.01 versus UUO plus pcDNA3.1 group (n = 5); ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus UUO plus pFlag-MMP-10 group (n = 5). F, G Representative micrographs (F) and quantitative data (G) showed renal expression of KIM-1 and fibronectin and fibrotic lesions by Sirius red staining. Scale bar, 50 µm. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus sham controls (n = 5); ^††P < 0.01 versus UUO plus pcDNA3.1 group (n = 5); ^##P < 0.01, ^###P < 0.001 versus UUO plus pFlag-MMP-10 group (n = 5).

Fig. 8. HB-EGF mediates MMP-10-triggered β-catenin transactivation.

A, B Western blot analysis showed the expression of cleaved HB-EGF in HK-2 cells after various treatments. Data are presented as the mean ± SEM. ***P < 0.001 versus negative controls (n = 6); ^†††P < 0.001 versus the group treated with rhMMP-10 (n = 6). C, D In vitro experiments showed that knockdown of HB-EGF effectively blocked rhMMP-10-induced expression of p-EGFR, p-ERK1/2, p-GSK-3β, active β-catenin and MMP-7. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 versus negative controls (n = 6); ^†P < 0.05, ^††P < 0.01, ^†††P < 0.001 versus the group treated with rhMMP-10 (n = 6). E Representative micrographs demonstrated β-catenin activation and nuclear translocation induced by rhMMP-10, which was abolished by knocking down of HB-EGF. Scale bar, 75 µm. F, G Western blotting showed the expression of fibronectin and α-SMA in HK-2 cells after various treatments as indicated. Knockdown of HB-EGF abolished rhMMP-10-induced fibronectin and α-SMA expression. Data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 versus controls (n = 6); ^†P < 0.05 versus the group treated with rhMMP-10 (n = 6). H Representative immunofluorescence images illustrated the expression levels of fibronectin across all treatment groups. Scale bar, 75 µm. I Re-analysis of an independent study through data mining (GSE193282) revealed that knockout of tubular β-catenin in mice abolished renal MMP-10 expression after UUO. Data are presented as the mean ± SEM. *P < 0.05 versus wild-type controls (n = 3). J, K Transient transfection with a N-terminus-truncated, constitutively activated β-catenin expression vector (pDel-β-cat) induced MMP-10 expression in HK-2 cells. Data are presented as the mean ± SEM. ***P < 0.001 versus pcDNA3 controls (n = 6). L Schematic diagram depicts the potential mechanism of MMP-10 action in kidney fibrosis. Upregulated MMP-10, as a response to injury, cleaves the HB-EGF, leading to its activation, which then activates EGFR and subsequently phosphorylates ERK1/2 and GSK-3β. This cascade of event causes β-catenin accumulation and translocation into the nucleus. Intranuclear β-catenin binds to TCF/LEF and promotes the transcription of its target genes including MMP-10, forming a vicious cycle. In addition, MMP-10 induces the expression of MMP-7 via β-catenin, which enzymatically degrades E-cadherin, liberating more β-catenin and further promoting kidney fibrosis.