Parkin-Mediated Mitophagy by TGF-β Is Connected with Hepatic Stellate Cell Activation

Abstract

Hepatic stellate cells (HSCs) are the main contributors to the development and progression of liver fibrosis. Parkin is an E3 ligase involved in mitophagy mediated by lysosomes that maintains mitochondrial homeostasis. Unfortunately, there is little information regarding the regulation of parkin by transforming growth factor-β (TGF-β) and its association with HSC trans-differentiation. This study showed that parkin is upregulated in fibrotic conditions and elucidated the underlying mechanism. Parkin was observed in the cirrhotic region of the patient liver tissues and visualized using immunostaining and immunoblotting of mouse fibrotic liver samples and primary HSCs. The role of parkin-mediated mitophagy in hepatic fibrogenesis was examined using TGF-β-treated LX-2 cells with mitophagy inhibitor, mitochondrial division inhibitor 1. Parkin overexpression and its colocalization with desmin in human tissues were found. Increased parkin in fibrotic liver homogenates of mice was observed. Parkin was expressed more abundantly in HSCs than in hepatocytes and was upregulated under TGF-β. TGF-β-induced parkin was due to Smad3. TGF-β facilitated mitochondrial translocation, leading to mitophagy activation, reversed by mitophagy inhibitor. However, TGF-β did not change mitochondrial function. Mitophagy inhibitor suppressed profibrotic genes and HSC migration mediated by TGF-β. Collectively, parkin-involved mitophagy by TGF-β facilitates HSC activation, suggesting mitophagy may utilize targets for liver fibrosis.

Keywords: Smad2; hepatic stellate cell; liver fibrosis; mitophagy; parkin.

Figure 1

Parkin overexpression in fibrotic liver samples. (A) Parkin and desmin immunostaining in cirrhotic patient and adjacent normal liver samples (magnification: 200×, Scale bar = 400 µm). White arrows indicate the colocalization of parkin and desmin. (B) Parkin and desmin immunostaining in carbon tetrachloride (CCl₄)-injected mice liver sections (magnification: 200×, Scale bar = 400 µm). White arrows indicate the colocalization of parkin and desmin. (C,D) Immunoblotting for parkin and PINK1 of liver samples from mice with fibrosis. For (C), the mice were injected with CCl₄ for two weeks. For (D), the mice underwent bile duct ligation (BDL). Protein levels of parkin or PINK1 were assessed by immunoblot analysis. GAPDH was used to assess equal protein loading. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant difference versus respective controls, * p < 0.05, ** p < 0.01, N.S. not significant).

Figure 2

Parkin upregulation upon TGF-β stimulation of hepatic fibrogenesis in HSCs. (A) Parkin expression in mouse primary hepatocytes and hepatic stellate cells (HSCs). Immunoblotting was performed on the cell lysates. β-actin was used to verify equal loading of proteins. Albumin (ALB) and α-smooth muscle actin (α-SMA) were detected as markers of hepatocytes and HSCs, respectively. Ponceau S staining represents equal loading of proteins. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant difference versus primary hepatocytes, ** p < 0.01). (B) Expression of parkin in primary HSCs from livers of mice administered a single dose (0.5 mg/kg) of CCl₄ for 24 h. The level of parkin or α-SMA was assessed by immunoblotting. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus primary HSCs isolated from vehicle-injected mice, ** p < 0.01). (C) Effect of transforming growth factor-β (TGF-β) on parkin expression in primary HSCs. Primary HSCs were isolated and treated with 2 ng/mL TGF-β for 12 h. The expression of parkin or α-SMA was evaluated by immunoblotting. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus vehicle-treated primary HSCs, * p < 0.05). (D) The time courses of parkin expression in TGF-β-treated LX-2 cells. Parkin protein was immunoblotted in the lysates of cells incubated with 2 ng/mL TGF-β for 0.5–24 h. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus control, * p < 0.05, ** p < 0.01). (E) The effect of various concentrations of TGF-β on parkin upregulation in LX-2 cells. Parkin was detected in the lysates of cells incubated with 0.5–2 ng/mL TGF-β for 12 h. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus control, ** p < 0.01).

Figure 3

Mechanistic study of the regulation of parkin induction by TGF-β. (A) RT-PCR analysis. LX-2 cells were treated with 2 ng/mL TGF-β for 1–12 h. The expression level of parkin mRNA was assessed by RT-PCR using GAPDH as an internal control. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus control, ** p < 0.01). (B) Effect of actinomycin D (ActD) on parkin induction by TGF-β in LX-2 cells. The cells were treated with 5 μg/mL of ActD with or without TGF-β treatment. The level of parkin was determined after 1 ng/mL TGF-β treatment for 12 h. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus respective controls, ** p < 0.01; significant versus TGF-β-treated cells, ## p < 0.01). (C) Involvement of Smad3 in TGF-β-induced parkin expression. LX-2 cells were transfected with pcDNA3.1 (MOCK) or Smad3 for 24 h. MOCK overexpressed cells were treated with 1 ng/mL TGF-β for 6 h. Parkin level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus MOCK-transfected cells, * p < 0.05, ** p < 0.01).

Figure 4

Role of TGF-β-induced parkin on mitophagy in LX-2 cells. (A) Immunostaining for parkin in TGF-β-stimulated LX-2 cells and their quantification. After treatment with 2 ng/mL TGF-β for 18 h, the cells were stained with MitoTracker™ Green (200 nM). CCCP was utilized as a positive control. Red: parkin, green: mitochondria, blue: DAPI, orange-yellow: merge. (white scale bar = 400 µm, red scale bar = 100 µm). The colocalization intensity of parkin with MitoTracker was evaluated by scanning densitometry. Ten different cells were randomly selected for each sample. The data represents the mean ± standard error (SE). (n = 3, significant different versus control, ** p < 0.01). (B) The effect of TGF-β on mitophagic activity. Level of parkin, LC3B, or p62 in LX-2 cells treated with 2 ng/mL TGF-β for various times (3–24 h) was detected by immunoblotting. Parkin or LC3BII level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus control, * p < 0.05, ** p < 0.01). (C) Effect of mitophagy inhibitor Mdivi-1 on TGF-β-mediated mitophagy. Expression of LC3B or p62 was measured after treatment with 2 ng/mL TGF-β sequential to 10 μM Mdivi-1 treatment. Parkin or LC3BII level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus control, * p < 0.05; significant versus TGF-β-treated cells, # p < 0.05, ## p < 0.01). (D) Representative image of mitophagy (left) and their quantification (right). LX-2 cells were infected with adeno-LC3B and then treated with MitoTracker™ (200 nM). Subsequently, cells were incubated with 2 ng/mL TGF-β for 12 h in the presence or absence of 10 μM Mdivi-1. Green: LC3B, red: MitoTracker™, blue: DAPI, orange-yellow: merge. Orange-yellow puncta represent mitochondrial digestion by autophagy (Scale bar = 16 µm). The colocalization intensity was evaluated by scanning densitometry. Ten different cells were randomly selected for each sample. The data represents the mean ± standard error (SE). (n = 3, significant different versus control, ** p < 0.01, significant versus TGF-β-treated cells, ## p < 0.01).

Figure 5

Effect of TGF-β exposure on mitochondrial function. (A) Effect of TGF-β on mitochondrial ROS. LX-2 cells were treated with or without 2 ng/mL TGF-β for 15 min, or 10 μM rotenone for 1 h, and then incubated with 10 μM MitoSOX™ for 30 min. Rotenone was used as a positive control. The cells were analyzed by flow cytometry (green: control, blue: TGF-β, sky-blue: rotenone). (B) Effect of TGF-β on mitochondrial membrane potential (MMP). MMP was measured by rhodamine-123 (Rho-123) staining and analyzed by flow cytometry. LX-2 cells were exposed to 2 ng/mL TGF-β or 10 μM rotenone for 18 h and then sequentially loaded with 0.05 ng/mL Rho-123 for 30 min. The inserted histogram exhibits a left shift of the histogram peak, illustrating the decrease of Rho-123 fluorescence intensity due to the loss of MMP. The percentage of cells with reduced fluorescence intensity was calculated.

Figure 6

Role of TGF-β-induced mitophagy on liver fibrogenesis and HSC migration in LX-2 cell (A) Effect of Mdivi-1 on hepatic fibrogenesis. LX-2 cells were pretreated with 10 μM Mdivi-1 for 0.5 h, and then were incubated with 1 ng/mL TGF-β for 12 h, and then liver fibrogenesis-related gene expression in the cell lysates was detected by immunoblotting. Parkin or PAI-1 level was assessed by scanning densitometry. The data represents the mean ± standard error (SE) (n = 3, significant different versus respective controls, * p < 0.05; significant versus TGF-β-treated cells, # p < 0.05). (B) Effect of Mdvi-1 on TGF-β-derived cellular migration by wound healing assay. LX-2 cells were treated with 10 µM Mdvi-1 with or without stimulation of TGF-β (2 ng/mL) after creating wounds.

Figure 7

Schematic illustration. Schematic diagram of the mechanism by which TGF-β-mediated mitophagy promotes profibrotic effect via parkin.