Enhancer of zeste homolog 2 promotes renal fibrosis after acute kidney injury by inducing epithelial-mesenchymal transition and activation of M2 macrophage polarization

Abstract

Long-term follow-up data indicates that 1/4 patients with acute kidney injury (AKI) will develop to chronic kidney disease (CKD). Our previous studies have demonstrated that enhancer of zeste homolog 2 (EZH2) played an important role in AKI and CKD. However, the role and mechanisms of EZH2 in AKI-to-CKD transition are still unclear. Here, we demonstrated EZH2 and H3K27me3 highly upregulated in kidney from patients with ANCA-associated glomerulonephritis, and expressed positively with fibrotic lesion and negatively with renal function. Conditional EZH2 deletion or pharmacological inhibition with 3-DZNeP significantly improved renal function and attenuated pathological lesion in ischemia/reperfusion (I/R) or folic acid (FA) mice models (two models of AKI-to-CKD transition). Mechanistically, we used CUT & Tag technology to verify that EZH2 binding to the PTEN promoter and regulating its transcription, thus regulating its downstream signaling pathways. Genetic or pharmacological depletion of EZH2 upregulated PTEN expression and suppressed the phosphorylation of EGFR and its downstream signaling ERK1/2 and STAT3, consequently alleviating the partial epithelial-mesenchymal transition (EMT), G2/M arrest, and the aberrant secretion of profibrogenic and proinflammatory factors in vivo and vitro experiments. In addition, EZH2 promoted the EMT program induced loss of renal tubular epithelial cell transporters (OAT1, ATPase, and AQP1), and blockade of EZH2 prevented it. We further co-cultured macrophages with the medium of human renal tubular epithelial cells treated with H₂O₂ and found macrophages transferred to M2 phenotype, and EZH2 could regulate M2 macrophage polarization through STAT6 and PI3K/AKT pathways. These results were further verified in two mice models. Thus, targeted inhibition of EZH2 might be a novel therapy for ameliorating renal fibrosis after acute kidney injury by counteracting partial EMT and blockade of M2 macrophage polarization.

Fig. 1. EZH2 is highly upregulated in kidney of patients with ANCA-associated glomerulonephritis and correlated positively with Masson’s trichrome-positive area and negatively with eGFR.

A The photomicrographs of EZH2 and H3K27me3 immunohistochemical staining, PAS, and Masson’s trichrome staining in normal paracancerous tissue from renal carcinoma patient (normal control, NC) and renal cortical tissue from patients with ANCA-associated glomerulonephritis (AAGN). B–K The correlation between positive areas of EZH2 as well as H3K27me3 and Masson’s positive area, serum creatinine, blood urea nitrogen (BUN), urine protein, and estimated glomerular filtration rate (eGFR) in all AAGN patients (n = 9). Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 2. EZH2 conditional knockout protects against renal dysfunction, suppresses renal fibrosis and loss of TEC transporters in I/R induced AKI-to-CKD transition mouse model.

The mouse model of AKI-to-CKD transition induced by I/R was established by clipping the bilateral renal arteries of mice for 30 min, and the animals were observed and fed for 4 weeks after operation. A Generation of conditional knockout mice in which EZH2 was specifically ablated in tubular epithelial cells by using Cre-LoxP recombination system. Genotyping was confirmed by tail preparation and PCR at 2 weeks of age. B Photograph showed the size, color, and texture of kidney in each group. C Serum creatinine of the mice in different groups. D Blood urea nitrogen (BUN) of the mice in different groups. E Photomicrographs showed the PAS staining of the kidneys. F Morphologic change of tubular injury was scored on the basis of PAS staining described in the Method section. G Photomicrographs showed the Masson’s trichrome staining of the kidneys. H The graph showed the positive areas (blue) of Masson’s trichrome staining. I Kidney tissue lysates from I/R mice were subjected to immunoblotting analysis with specific antibodies against EZH2, H3K27me3, Histone H3, and GAPDH. J, K Expression levels of EZH2 and H3K27me3 in different groups were quantified by densitometry and normalized with GAPDH and Histone H3 respectively. L Kidney tissue lysates from I/R mice were subjected to immunoblotting analysis with specific antibodies against α-SMA, Collagen I, E-cadherin, and GAPDH. M–O Expression levels of α-SMA, Collagen I, E-cadherin in different groups were quantified by densitometry and normalized with GAPDH. P Kidney tissue lysates from I/R mice were subjected to immunoblotting analysis with specific antibodies against OAT1, AQP1, ATPase, and GAPDH. Q–S Expression levels of OAT1, AQP1, and ATPase in different groups were quantified by densitometry and normalized with GAPDH. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 3. EZH2 conditional knockout protects against renal dysfunction, suppresses renal fibrosis and loss of TEC transporters in FA induced AKI-to-CKD transition mouse model.

The mice in FA model were injected with a single dose of folic acid (250 mg/kg, dissolved in 300 mM NaHCO₃, i.p.), while the control group was injected with an identical voluminal vehicle (300 mM NaHCO₃, i.p.) for 4 weeks. A Photograph showed the size, color, and texture of kidney in each group. B Serum creatinine of the mice in different groups. C Blood urea nitrogen (BUN) of the mice in different groups. D Photomicrographs showed the PAS staining of the kidneys. E Morphologic change of tubular injury was scored on the basis of PAS staining described in the Method section. F Photomicrographs showed the Masson’s trichrome staining of the kidneys. G The graph showed the positive areas (blue) of Masson’s trichrome staining. H Kidney tissue lysates from FA mice were subjected to immunoblotting analysis with specific antibodies against EZH2, H3K27me3, Histone H3 and GAPDH. I, J Expression levels of EZH2 and H3K27me3 in different groups were quantified by densitometry and normalized with GAPDH and Histone H3 respectively. K Kidney tissue lysates from FA mice were subjected to immunoblotting analysis with specific antibodies against α-SMA, Collagen I, E-cadherin, and GAPDH. L–N Expression levels of α-SMA, Collagen I, E-cadherin in different groups were quantified by densitometry and normalized with GAPDH. (O) Kidney tissue lysates from FA mice were subjected to immunoblotting analysis with specific antibodies against OAT1, AQP1, ATPase and GAPDH. P–R Expression levels of OAT1, AQP1, and ATPase in different groups were quantified by densitometry and normalized with GAPDH. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 4. EZH2 conditional knockout upregulates the expression of PTEN, thus blocking the activation of EGFR/ERK1/2/STAT3 signaling pathway in I/R induced AKI-to-CKD transition mouse model.

A Kidney tissue lysates from I/R mice were prepared and subjected to immunoblotting analysis with antibodies against PTEN, p-EGFR, EGFR, and GAPDH. B–D Expression levels of PTEN, p-EGFR, EGFR in different groups were quantified by densitometry and normalized with GAPDH and EGFR respectively. E Photomicrographs showed the immunofluorescent staining of PTEN in different groups. F Kidney tissue lysates from I/R mice were prepared and subjected to immunoblotting analysis with antibodies against p-ERK1/2, ERK1/2, p-STAT3, STAT3, and GAPDH. G–J Expression levels of p-ERK1/2, ERK1/2, p-STAT3, STAT3 in different groups were quantified by densitometry and normalized with GAPDH, ERK1/2, and STAT3, respectively. K Kidney tissue lysates from I/R mice were prepared and subjected to immunoblotting analysis with antibodies against Snail, H3pSer10, and GAPDH. L, M Expression levels of Snail and H3pSer10 in different groups were quantified by densitometry and normalized with GAPDH. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 5. Genetic or pharmacological depletion of EZH2 inhibits H₂O₂ induced partial EMT of cultured human tubular epithelial cells.

Starved HK2 cells were exposed to H₂O₂ (0.5 mM) in the presence of 3-DZNeP (5 μM) for 24 h before cell harvesting. A Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against EZH2, H3K27me3, Histone H3, α-SMA, E-cadherin, and GAPDH. B–E Expression levels of EZH2, H3K27me3, α-SMA, E-cadherin were quantified by densitometry and normalized with GAPDH and Histone H3. F Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against OAT1, AQP1, ATPase, and GAPDH. G–I Expression levels of OAT1, AQP1, ATPase were quantified by densitometry and normalized with GAPDH. J Photomicrographs showed the immunofluorescent co-staining of EZH2 and α-SMA in the HK2 cells of each group. HK2 cells were transfected with EZH2 siRNA and scrambled siRNA for 6 h, and then incubated with or without H₂O₂ (0.5 mM) for an additional 24 h before being harvested for analysis. K Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against EZH2, H3K27me3, Histone H3, α-SMA, E-cadherin, and GAPDH. L–O Expression levels of EZH2, H3K27me3, α-SMA, E-cadherin were quantified by densitometry and normalized with GAPDH and Histone H3. P Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against OAT1, AQP1, ATPase, and GAPDH. Q–S Expression levels of OAT1, AQP1, ATPase were quantified by densitometry and normalized with GAPDH. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 6. EZH2 binds to the PTEN promoter and regulates its transcription, thus increasing the activation of EGFR/ERK1/2/STAT3 signaling pathway in cultured human tubular epithelial cells.

A Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against PTEN, p-EGFR, EGFR, and GAPDH. B, C Expression levels of PTEN and p-EGFR were quantified by densitometry and normalized with GAPDH and EGFR, respectively. D Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against p-ERK1/2, ERK1/2, p-STAT3, STAT3, and GAPDH. E, F Expression levels of p-ERK1/2 and p-STAT3 were quantified by densitometry and normalized with ERK1/2 and STAT3, respectively. G Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against Snail, H3pSer10, and GAPDH. H, I Expression levels of Snail and H3pSer10 were quantified by densitometry and normalized with GAPDH. J 0.5 mM H₂O₂-treated HK2 cell lysates were subjected to immunoprecipitation with IgG or EZH2 antibody, followed by EZH2 and H3K27me3 immunoblotting. K Content of PTEN promoter enriched by IgG or H3K27me3 antibody in HK2 cells treated with or without 0.5 mM H₂O₂ by CUT & TAG analysis. L Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against PTEN, p-EGFR, EGFR, and GAPDH. M, N Expression levels of PTEN and p-EGFR were quantified by densitometry and normalized with GAPDH and EGFR respectively. O Photomicrographs showed the immunofluorescent staining of PTEN in the HK2 of each group. P Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against p-ERK1/2, ERK1/2, p-STAT3, STAT3, and GAPDH. Q, R Expression levels of p-ERK1/2 and p-STAT3 were quantified by densitometry and normalized with ERK1/2 and STAT3, respectively. S Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against Snail, H3pSer10, and GAPDH. T, U Expression levels of Snail and H3pSer10 were quantified by densitometry and normalized with GAPDH. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 7. Genetic or pharmacological depletion of EZH2 inhibits M2 macrophage polarization through STAT6 and PI3K/AKT pathways in Raw264.7 cells.

Starved Raw264.7 cells were cultured with or without 10% (vol/vol) pre-collected cell culture media from HK2 cells treated with 0.5 mM H₂O₂ in the presence of 3-DZNeP (5 μM) for 24 h before cell harvesting. A Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against EZH2, H3K27me3, Histone H3, and GAPDH. B, C Expression levels of EZH2 and H3K27me3 were quantified by densitometry and normalized with GAPDH and Histone H3. D Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against CD163, Arginase-1, and GAPDH. E, F Expression levels of CD163 and Arginase-1 were quantified by densitometry and normalized with GAPDH. G Cell lysates treated with 3-DZNeP were subjected to immunoblotting analysis with antibodies against p-STAT6, STAT6, p-PI3K, PI3K, p-AKT, AKT, and GAPDH. H–J Expression levels of p-STAT6, p-PI3K, and p-AKT were quantified by densitometry and normalized with STAT6, PI3K, and AKT. K Raw264.7 cell lysates cultured with 10% (vol/vol) pre-collected cell culture media from HK2 cells treated with 0.5 mM H₂O₂ were subjected to immunoprecipitation with IgG or EZH2 antibody, followed by EZH2 and STAT6 immunoblotting. RAW264.7 cells were transfected with EZH2 siRNA and scrambled siRNA for 6 h, and then incubated with or without 10% (vol/vol) pre-collected cell culture media from HK2 cells treated with 0.5 mM H₂O₂ for an additional 24 h before being harvested for analysis. L Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against EZH2, H3K27me3, Histone H3, and GAPDH. M, N Expression levels of EZH2 and H3K27me3 were quantified by densitometry and normalized with GAPDH and Histone H3. O Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against CD163, Arginase-1, and GAPDH. (P, Q) Expression levels of CD163 and Arginase-1 were quantified by densitometry and normalized with GAPDH. R Cell lysates transfected with EZH2 siRNA were subjected to immunoblotting analysis with antibodies against p-STAT6, STAT6, p-PI3K, PI3K, p-AKT, AKT, and GAPDH. S–U Expression levels of p-STAT6, p-PI3K, and p-AKT were quantified by densitometry and normalized with STAT6, PI3K, and AKT. V Photomicrographs showed the immunofluorescent staining of CD163 in the RAW264.7 of each group. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.

Fig. 8. EZH2 conditional knockout inhibits M2 macrophage polarization through STAT6 and PI3K/AKT pathways in I/R or FA induced AKI-to-CKD transition mice models.

A Photomicrographs showed the immunofluorescent staining of CD163 in different groups. B Kidney tissue lysates from I/R mice were prepared and subjected to immunoblotting analysis with antibodies against CD163, Arginase-1, MMP9, and GAPDH. C–E Expression levels of CD163, Arginase-1, MMP9 in different groups were quantified by densitometry and normalized with GAPDH. F Kidney tissue lysates from I/R mice were prepared and subjected to immunoblotting analysis with antibodies against p-STAT6, STAT6, p-PI3K, PI3K, p-AKT, AKT, and GAPDH. G–I Expression levels of p-STAT6, p-PI3K, and p-AKT were quantified by densitometry and normalized with STAT6, PI3K, and AKT. J Kidney tissue lysates from FA mice were prepared and subjected to immunoblotting analysis with antibodies against CD163, Arginase-1, MMP9, and GAPDH. K Kidney tissue lysates from FA mice were prepared and subjected to immunoblotting analysis with antibodies against p-STAT6, STAT6, p-PI3K, PI3K, p-AKT, AKT, and GAPDH. L–N Expression levels of p-STAT6, p-PI3K, and p-AKT were quantified by densitometry and normalized with STAT6, PI3K, and AKT. O The role and mechanisms of EZH2 in AKI-to-CKD transition. Data were expressed as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. N.S., statistically not significant, with the comparisons labeled. All scale bars = 50 μm.