Bnip3 interacts with vimentin, an intermediate filament protein, and regulates autophagy of hepatic stellate cells

The expression of Bnip3 was increased in fibrotic liver tissues from clinical patients and Schistosoma japonicum infected mice

Expression of Bnip3 was firstly detected in fibrotic liver samples from clinical patients. Compared with non-fibrotic liver samples (Figure 1A), fibrotic liver tissues had obviously higher expression of Bnip3 (Figure 1B, 1E). Bnip3 was also detected in fibrotic liver tissues of mice infected with S. japonicum, which was recognized as an animal model of infective fibrosis [17–20]. It was shown that Bnip3 increased in fibrotic liver tissue from infected mice, as compared with normal liver tissues from non-infected mice (Figure 1C, 1D, 1F). From the morphological characteristics, it was suggested that Bnip3 positive cells were mainly non-parenchymal hepatic cells (Figure 1B, 1D). Collectively, Bnip3 was increasingly expressed in fibrotic liver tissues.

Figure 1. The expression of Bnip3 was increased in fibrotic liver tissues from clinical patients and Schistosoma japonicum infected mice. Expression of Bnip3 was identified in liver biopsy specimens from clinical patients, and mice infected with S. japonicum by immunohistochemistry. (A, B) Human samples from liver fibrosis were collected for Bnip3 detection by immunohistochemistry and liver samples of para-hemangioma from patients, were used as negative control. (A) Liver samples of para-hemangioma from patients (n=8). (B) Fibrotic liver samples from patients (n=8). Scale bar, 20 μm. (C, D) BALB/c female mice, 6–8 weeks old, were percutaneously infected with 25 cercariae of Schistosoma japonicum through the shaved abdomen, sacrificed at 8 weeks post-infection, and samples of liver were collected. (C) Non-infected liver samples from mice (n=4) and (D) fibrotic liver samples from S. japonicum infected mice (n=4). Scale bar, 20 μm. Representative images were shown. Positive cells were counted and the scores were multiplied with staining intensity. Products were introduced into SPSS 19.0 to analyze statistical difference of Bnip3 expression in normal liver tissues and fibrotic liver tissues. (E) Bnip3 in liver tissues from patients; (F) Bnip3 in liver tissues from S. japonicum infected mice. **P < 0.01.

The activation of hepatic stellate cells induced Bnip3 expression and its cytoplasmic translocation in vitro

We previously reported that human hepatic stellate cell line, LX-2, was activated upon hypoxia or LPS stimulation, along with the increase of Hif-1 [17–20]. Increased expression of Bnip3 was also confirmed in both hypoxia or LPS-stimulated LX-2 cells (Figure 2A). In immunofluorescence staining, it was observed that Bnip3 was localized in nucleus in non-stimulated control LX-2 cells, while in hypoxia and LPS-stimulated cells, Bnip3 displayed obvious cytoplasmic localization (Figure 2B). Consistent results were observed in primary HSCs isolated from mice. We have previously determined that in in-vitro culture, HSCs were naturally activated [19]. The increased expression of Bnip3 in activated primary HSCs was also confirmed by Western blot, which indicated that Bnip3 increasingly expressed as the primary HSCs were activated in-vitro cultured up to 4 or 6 days, as compared with cells cultured up to 2 days (Figure 2C). Nuclear Bnip3 expression was also observed in primary HSCs cultured up to day 2, while in cells cultured up to day 6, Bnip3 was mainly expressed in cytoplasm (Figure 2D). The results all above demonstrated that Bnip3 expressed increasingly and translocated from nucleus to cytoplasm during HSC activation.

Figure 2. The activation of hepatic stellate cells induced Bnip3 expression and its cytoplasmic translocation in vitro. LX-2 cells were treated with 100 μM CoCl₂ or 2 μg/ml LPS for 8 h. (A) Cells were collected at indicated time and cell lysates were subjected to detect Bnip3 with Western blot. Densitometric analysis for Western blot was performed and data were expressed as mean ± SD, *P < 0.05. (B) Immunofluorescence assay was performed to detect Bnip3 (Cy3) in CoCl₂- or LPS-treated LX-2 cells. Culture-activated primary HSCs from mice were cultured up to 2 days, 4 days or 6 days. (C) Cells were collected at indicated time and cell lysates were subjected to detect Bnip3 with Western blot. Densitometric analysis for Western blot was performed and data were expressed as mean ± SD, *P < 0.05, **P < 0.01. (D) Immunofluorescence assay was performed to detect Bnip3 (Cy3) and images were captured by confocal microscope.

Autophagy occurred when hepatic stellate cells were activated, along with increase of Hif-1

We have previously reported that Hif-1 regulates autophagy to activate hepatic stellate cells, using LX-2 cells, human HSC cell line, as the cell model [17]. LC3 (microtubule-associated protein light chain 3), a mammalian homolog of yeast Atg8, is a widely recognized protein for monitoring autophagic activity. There are three isoforms of LC3 in mammalian cells including LC3A, LC3B and LC3C, among which, LC3B is a widely used autophagy marker [21]. After synthesis, LC3 is cleaved to generate 16-kDa cytosolic LC3 and 14-kDa lipidated LC3. 14-kDa lipidated LC3 is the membrane component involved in formation of autophagosomes and autophagolysosomes [22, 23]. In this study, formation of autophagosomes indicated by LC3B punctate particles and 14-kDa lipidated LC3B were used to detect autophagy activity. In LX-2 cells, it was observed that CoCl₂ or LPS stimulation resulted in apparent accumulation of LC3B (Figure 3A), which is consistent with our previous study [17]. Molecular markers of autophagy, including LC3B punctate particles (Figure 3B) and 14-kDa lipidated LC3B (Figure 3C) were respectively detected with confocal microscopy and Western blot in primary HSCs. Expression of Hif-1α, α-SMA, molecular marker of HSC activation, was also detected (Figure 3C). As the primary HSCs were cultured in vitro, with the extending of in-vitro culture time, increase of α-SMA indicated the activation of primary HSCs and increase of Hif-1α demonstrated the activation of Hif-1 signaling cascade in primary HSCs (Figure 3C). Increase of 14-kDa lipidated LC3B and accumulation of LC3B punctate particles in primary HSCs cultured in vitro up to day 6, indicated that autophagic degradation enhanced along with the activation of primary HSCs (Figure 3B, 3C). Bnip3 was previously reported to be associated with both autophagy and apoptosis. Apoptosis was detected in hypoxia/LPS-stimulated LX-2 cells and the results indicated that apoptosis did not occur in our in-vitro culture system (Supplementary Figure 1). The above results demonstrated the enhanced autophagy in activated HSCs, with the activation of Hif-1 signaling.

Figure 3. Autophagy occurred when hepatic stellate cells were activated. (A) LX-2 cells were treated with 100 μM CoCl₂ or 2 μg/ml LPS for 8 h. (B) Culture-activated primary HSCs from mice were cultured up to 2 days or 6 days. Immunofluorescence assay was performed to detect LC3B (FITC) in CoCl₂- or LPS-treated LX-2 cells (A) and culture-activated primary HSCs from mice cultured up to 2 days or 6 days (B). Images were captured by confocal microscope. (C) Culture-activated primary HSCs from mice were cultured up to 2 days, 4 days or 6 days. Cells were collected at indicated time and cell lysates were subjected to detect Hif-1α, α-SMA and LC3B with Western blot. Densitometric analysis for Western blot was performed and data were expressed as mean ± SD, *P < 0.05, ***P < 0.001.

Inhibition of Hif-1α suppressed the increased expression of Bnip3 in hepatic stellate cells

Bnip3 was previously reported to be regulated by Hif-1 in cells upon hypoxia [24], however, the relationship between Hif-1 and Bnip3 was still unclear in hepatic stellate cells. YC-1, chemical inhibitor of Hif-1α, was used to inhibit Hif-1α expression. In primary HSCs, as Hif-1α expression was efficiently inhibited by YC-1, Bnip3 expression was suppressed in in-vitro cultured and activated primary HSCs (Figure 4A, 4B). Similar results were observed in LX-2 cells. LX-2 cells were pretreated with YC-1 and then stimulated with CoCl₂ or LPS. The results from Western blot (Figure 4C) and immunofluorescence staining (Figure 4D) showed that the inhibition of Hif-1 expression led to the suppress of Bnip3 in hypoxia or LPS activated LX-2 cells. Collectively, these results indicated that Bnip3 expression was regulated by Hif-1 during the activation of hepatic stellate cells.

Figure 4. Inhibition of Hif-1 suppressed increased expression of Bnip3 in activated hepatic stellate cells. Culture-activated primary HSCs from mice were cultured up to 2 days, 4 days or 6 days and Hif-1 chemical inhibitor YC-1 (50 μM) was added as cells were cultivated to 2 days. (A) Cell lysates were subjected to detect Hif-1α and Bnip3 with Western blot. Densitometric analysis was performed and data were expressed as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001. (B) Immunofluorescence assay was performed to detect Bnip3 (Cy3) by confocal microscopy. (C) LX-2 cells were stimulated by 100 μM CoCl₂ or 2 μg/ml LPS either alone or after YC-1 pre-treatment (50 μM). Cell lysates were subjected to detect Hif-1α and Bnip3 with Western blot. Densitometric analysis was performed and data were expressed as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001. (D) Immunofluorescence assay was performed to detect Bnip3 (Cy3) by confocal microscopy.

Bnip3 was partially co-localized with autophagosomes in activated HSCs and inhibition of Bnip3 led to the blockage of the autophagic flow and activation of HSCs

Bnip3 and LC3B were co-stained to observe whether Bnip3 was co-localized with autophagosomes in activated HSCs. It was shown that in both hypoxia/LPS-stimulated LX-2 cells (Figure 5A) and culture-activated primary HSCs from mice (Figure 5B), Bnip3 was partially co-localized with LC3B punctate particles. Autophagic flux was further detected in primary HSCs transfected with Bnip3 siRNA. Primary cells were firstly transfected with Bnip3 siRNA at day 2 and GFP-RFP-LC3B plasmid was transfected 24h later. Cells were further cultured until day 6 and formation of autophagosomes was observed. In control cells transfected with non-specific siRNA, punctate particles of both GFP-LC3B (autophagosomes) and RFP-LC3B (autophagolysosomes) were formed, when merged, non-degraded autophagic bodies (yellow dots) and degraded autophagic bodies (red dots) were clearly visible, which indicated the functional autophagy in control cells (Figure 5C). However, in Bnip3 interfered primary HSCs, signals of both GFP-LC3B (autophagosomes) and RFP-LC3B (autophagolysosomes) were attenuated. Especially, the signals of RFP-LC3B (autophagolysosomes) were obviously weaker than the signals of GFP-LC3B (autophagosomes) (Figure 5C), which indicated the blockage of autophagic flow due to Bnip3 interference.

Figure 5. Bnip3 was partially co-localized with autophagosomes in activated HSCs and inhibition of Bnip3 led to the blockage of the autophagic flow. LX-2 cells were treated with 100 μM CoCl₂ or 2 μg/ml LPS for 8 h. Culture-activated primary HSCs from mice were cultured up to 2 days or 6 days. Immunofluorescence assay was performed to detect Bnip3 (Cy3) and LC3B (FITC) in LX-2 cells (A) and culture-activated primary HSCs from mice (B) by confocal microscopy. (C) Primary HSCs were isolated from mice and seeded in coverslips. Cells were transfected with specific siRNA targeting Bnip3 as cells were cultivated up to day 2 and GFP-RFP-LC3B plasmid was transfected into cells 24h later. Cells were fixed at day 6 and images were captured by confocal microscope to observe the formation of autophagosomes.

Molecular marker of HSC activation, α-SMA, was detected in CoCl₂/LPS-stimulated LX-2 cells (Supplementary Figure 2A, 2B) and primary HSCs (Supplementary Figure 2C) transfected with Bnip3-specific siRNA. It was shown that Bnip3 interference inhibited increase of Bnip3 and α-SMA. Collectively, the results above demonstrated that Bnip3 played an important role in the autophagy and activation of HSCs.

Bnip3 interacted with vimentin in activated hepatic stellate cells

Proteins interacted with Bnip3 were further screened in hypoxia-stimulated LX-2 cells with mass spectrometry (Supplementary Tables 1, 2). It was reported from mass spectrometry that Bnip3 might have interaction with a kind of intermediate filament (IF) protein, vimentin, which has highly abundant expression in HSCs [25]. Antibody specific to Bnip3 was used as the bait to detect vimentin with co-immunoprecipitation and the interaction of Bnip3 and vimentin was confirmed in both hypoxia/LPS stimulated LX-2 cells and culture-activated primary HSCs (Figure 6A). As vimentin has multiple phosphorylation sites at serine residues, phosphorylated vimentin at serine residue 56, serine residue 72, serine residue 82 and serine residue 38, was also detected in Bnip3 interacted complex. In hypoxia/LPS stimulated LX-2 cells, as compared with control cells, phosphorylations at serine residue 72 and 82 were enhanced, while phosphorylations at serine residue 56 and 38 were reduced (Figure 6A). In primary HSCs cultured up to day 6, as compared with cells cultured up to day 2, phosphorylation at serine residue 72 was enhanced, phosphorylations at serine residue 56 and 38 were reduced, while phosphorylation at serine residue 82 showed no change (Figure 6A). Apparent co-localization of Bnip3 and vimentin was captured by confocal microscopy in both hypoxia/LPS stimulated LX-2 cells (Figure 6B) and primary HSCs cultured up to day 6 (Figure 6C). Withaferin A, a chemical inhibitor targeting vimentin [26], which induces aggregation and fragmentation of vimentin, was used to inhibit vimentin activity in LX-2 cells. It was shown that inhibition of vimentin re-organization with Withaferin A inhibited increased expression of Bnip3 in activated LX-2 cells. Autophagy marker LC3B was also detected and 14-kDa lipidated LC3B was inhibited with Withaferin A treatment in LX-2 cells (Supplementary Figure 3). Collectively, the above results demonstrated that in activated HSCs, Bnip3 regulated the autophagy and activation of HSCs, possibly via interacting with vimentin, a kind of IF protein with highly abundant expression in HSCs.

Figure 6. Bnip3 interacted with vimentin in activated hepatic stellate cells. LX-2 cells were treated with 100 μM CoCl₂ or 2 μg/ml LPS for 8 h. Culture-activated primary HSCs from mice were cultured up to 2 days or 6 days. (A) 300μg cell lysates of LX-2 cells and 150μg cell lysates of primary HSCs from mice were collected and immunoprecipitated with anti-Bnip3, followed by immunoblotting with anti-vimentin, anti-phosphorylated vimentin S56, S72, S82 and S38. (B, C) Immunofluorescence assay was performed to detect Bnip3 (Cy3) and vimentin (FITC) in LX-2 cells (B) and culture-activated primary HSCs from mice (C) by confocal microscopy.

Ethics statement and human studies

Animal experiments were approved by the Committee on Animal Research of Tongji Medical College (TMC), Huazhong University of Science and Technology, Hubei Province, PR China. Human studies were approved by the institutional review board of TMC. All patients provided written informed consent. Eight cases of liver tissues from patients with liver fibrosis diagnosed by liver biopsy and normal liver tissues surrounding hepatic hemangioma excised from surgical operation were collected.

Animal study

BALB/c female mice, 8 weeks old, were randomly divided into two groups: the infected group and the control group. S. japonicum cercariae were shed from the snails. Each anaesthetized mouse in the infected group was percutaneously infected with 25 cercariae through the shaved abdomen. The mice were sacrificed at 8 weeks post-infection, and samples of liver were collected [17].

Isolation of primary cells and cell culture

Primary hepatic stellate cells were isolated from healthy BALB/c mouse by enzymatic digestion and Percoll density gradient centrifugation [19]. Primary HSCs were cultured in DMEM (Hyclone, USA) supplemented with 16% inactivated fetal bovine serum (Gibco, USA) as appropriate. LX-2, human HSC line, was grown in DMEM supplemented with 10% inactivated fetal bovine serum (Sciencell, USA) as appropriate. The antibiotics penicillin G (100 U/mL) and streptomycin sulfate (100 μg/mL) were added.

Transfection, cell stimulation and chemical inhibitors

Cells were transfected with 50nM siRNA specific to Bnip3 or non-specific siRNA using Ribo FECTTM CP transfection kit following manual instruction (RiboBio, China). Lipofectamine 2000 (Invitrogen, USA) was used to transfect GFP-RFP-LC3B plasmid into cells. Cells were stimulated by 100μM CoCl₂ (Sigma, C8661, USA) or 2μg/ml LPS (Sigma, L-2880, USA). YC-1 (MCE, 170632-47-0, USA) was used as Hif-1 chemical inhibitor and Withaferin A (MCE, HY-N2065, USA) was used as chemical inhibitor of vimentin re-organization.

Antibodies

Primary antibodies used for Western blot, immunohistochemistry and immunocytochemistry are listed as followed: Hif-1α (79233, Cell signaling, USA), α-SMA (ab32575, Abcam, USA), Bnip3 (ab10433, Abcam, USA; ab38621, Abcam, USA), LC3B (83506, cell signaling, USA; 12741, cell signaling, USA), vimentin (ServiceBio, GB12192, China), phosphor-vimentin Ser56 (ab217673, Abcam, USA), phosphor-vimentin Ser38 (ab52942, Abcam, USA), phosphor-vimentin Ser72 (ab52944, Abcam, USA), phosphor-vimentin Ser82 (ab52943, Abcam, USA), caspase 3 (ab197202, Abcam, USA), cytochrome C (sc-13156, Santa Cruz, USA), GAPDH (GB12002, Servicebio, China).

Western blot

Protein concentration from cell lysates was valued using BCA Protein Assay Kit (P0011, Beyotime, China). Protein samples were separated by SDS/PAGE and transferred onto PVDF membrane (Millipore, USA). After blocking in 5% BSA, membranes were incubated with primary antibodies and then corresponding secondary antibodies. Immunoreactive bands were tested with SuperSignal TM ELISA Femto Maximum Sensitivity Substrate (37075, Thermo, USA).

Immunohistochemistry and immunocytochemistry

The formalin-fixed and paraffin-embedded liver tissues were cut into 4-μm sections and the expression of Bnip3 was determined in tissue of normal and fibrotic livers. Cells seeded on coverslips were fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and counter-stained with antibodies against Bnip3, LC3B or vimentin. The coverslips were mounted onto slides in anti-fade mounting medium (H-1000, Vectorlabs, USA) and fluorescent images were captured using confocal microscopy (FV3000, Olympus, Japan).

Immunoprecipitation and mass spectrometry

Cells were lysed in 4° C pre-cooled RIPA buffer and 150-300 μg of cell lysate was incubated with 1 μg Bnip3 monoclonal antibody at 4° C overnight with continuous agitation. Protein A+G Agarose (P2012, Beyotime, China) was added and incubated for additional 2 hours at 4° C. The beads were washed and precipitated proteins were eluted by boiling the beads in 2×SDS-PAGE sample buffer for 5 minutes. The samples were analyzed by Western blot or LC-MS/MS-based analysis.

Flow cytometry analysis of apoptosis

Cells were stained with Annexin V-FITC/PI detection kit according to the manual instruction (AD-10, Dojindo, Japan). Apoptotic cells were detected using flow cytometry (AFC2, Life Technologies, Singapore) and apoptosis was calculated with percentage of Q2 cells (late apoptosis) and Q4 cells (early apoptosis).

Statistical analysis

All data are expressed as mean±SD. Differences between experimental and control groups were assessed by one-way ANOVA or unpaired, two-sided Mann–Whitney U-test using SPSS 19.0, *P < 0.05, **P < 0.01, ***P < 0.001.