Ginsenoside Rg3 restores hepatitis C virus–induced aberrant mitochondrial dynamics and inhibits virus propagation

CELL CULTURE AND VIRUS

Human hepatoma Huh7 and Huh7.5.1 cells were grown in high-glucose Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal bovine serum (Gibco), 1% minimal essential medium nonessential amino acids (Gibco), 100 units/mL penicillin, and 100 μg/mL streptomycin (Gibco). The R-1 HCV subgenomic replicon cells were maintained in Dulbecco's modified Eagle's medium with 0.5 mg/mL G418.13 Cell culture–derived HCV Japanese fulminant hepatitis 1 genotype 2a (HCVcc) was propagated and prepared as described.14 HCV infection was performed at a multiplicity of infection of 3.

IMMUNOFLUORESCENCE

HCV-infected cells and those posttreated with G-Rg3 (100 μM) were grown on coverslips and used in immunofluorescence assays as described.8, 9, 15 For monitoring mitophagic process, Huh7 cells transfected with phosphorylated signal peptide targeting mitochondria–red fluorescence protein–enhanced green fluorescence protein (p-mito-mRFP-EGFP) reporter were infected with HCVcc for 1 day and then treated with G-Rg3. At 2 days posttreatment, cells were immunostained with HCV core antibody. Confocal images were visualized under a 100 × oil objective using an Olympus FluoView 1000 confocal microscope or Zeiss LSM700 laser scanning confocal microscope. Analyses of colocalization of proteins and mitochondrial lengths were quantified by ImageJ software and FV10-ASW 3.0 viewer software (Olympus), respectively.

PLASMIDS, ANTIBODIES, AND REAGENTS

The p-mito-mRFP-RFP and pFLAG-CMV-HCV core DNA plasmids used in this study have been described.8, 15, 16 The primary antibodies used in this study include the following: rabbit monoclonal anti-Drp1 (Cell Signaling), rabbit monoclonal anti-phospho-Drp1 (Ser616; Cell Signaling), mouse monoclonal anti-HCV core (Thermo Scientific), rabbit polyclonal anti-β-actin (Cell Signaling), mouse monoclonal anti–translocase of outer mitochondrial membrane 20 (TOM20; BD Biosciences), human monoclonal anti-HCV E2,17 mouse monoclonal anti-HCV NS3 (Abcam), mouse monoclonal anti-mitofusin 2 (Mfn2; Abcam), and rabbit polyclonal anti–voltage-dependent anion channel 1 (VDAC1; Cell Signaling). The secondary antibodies used for western blot analysis were horseradish peroxidase–conjugated antimouse immunoglobulin G (IgG) and horseradish peroxidase–conjugated antirabbit IgG (both from Promega). The secondary antibodies for immunofluorescence were Alexa Fluor 350, 488, 594, or 647 donkey antimouse, rabbit, or goat IgG and Alexa Fluor 555 goat antihuman IgG (all from Molecular Probes). The chemical reagents used in this study were carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Sigma), sofosbuvir (Selleckchem), and ginsenoside compounds (Sigma).

WESTERN BLOT ANALYSIS

For western blot analysis, whole-cell lysates were extracted from cells, homogenized, subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and transferred to a nitrocellulose membrane (Thermo Scientific). Western blot analysis with antibodies against the indicated proteins was then performed as described.8, 9, 15 Densitometric graphs were generated using Image J (National Institutes of Health).18

REAL-TIME QUANTITATIVE RT-PCR

Real-time quantitative RT-PCR analysis of HCV RNA, mitochondrially encoded reduced nicotinamide adenine dinucleotide dehydrogenase subunit 2 (ND2) and cyclooxygenase 2 (COX2) levels was performed as described.9

CELL VIABILITY ASSAY

A previously described cell viability assay9 was used to assess the cytotoxic effects of ginsenoside compounds during HCV infection.

ΔΨm MEASUREMENT

Huh7 cells infected with HCVcc or treated with sofosbuvir, telaprevir, boceprevir, and CCCP at the indicated concentrations were stained with JC-1 reagent (Invitrogen) for 1 hour and immediately used for fluorescence-activated cell sorting analysis.

HUMAN LIVER BIOPSY SPECIMENS

Frozen human liver biopsy specimens (n = 6) were collected in the Soonchunhyang University Hospital Seoul from anti-HCV-negative patients (n = 3) and anti-HCV-positive patients (n = 3). All three anti-HCV-negative patients were nearly at the complete healing stage of drug-induced hepatitis, while all three anti-HCV-positive patients had active CHC. Liver biopsy samples were taken after obtaining written informed consent.

STATISTICAL ANALYSIS

Statistical analyses using the unpaired Student t test were performed using Sigma Plot software (Systat Software, San Jose, CA).

SOFOSBUVIR CAUSES ABERRANT MITOCHONDRIAL FISSION

DAAs, particularly sofosbuvir, are key components in the current IFN-sparing treatment regimens for CHC patients. Despite the high efficacy of sofosbuvir, its role in altering mitochondrial dynamics and physiology has not been studied, although mitochondrial toxicity caused by DAAs has been reported.19 Because sofosbuvir is an effective and one of the most widely used DAAs, we first investigated if sofosbuvir perturbs mitochondrial dynamics associated with modulation of mitochondrial structure and function. We examined the changes in ΔΨ_m in the presence of sofosbuvir. Treatment of human hepatoma Huh7 cells with sofosbuvir at the half-maximal effective concentration induced a decrease in the ΔΨ_m (Fig. 1A). Similar results were obtained in cells treated with the mitochondrial uncoupler CCCP, which causes mitochondrial depolarization (Fig. 1A; Supporting Fig. S1A).20 We also observed that telaprevir and boceprevir, which are HCV NS3 protease inhibitors, cause ΔΨ_m loss (Supporting Fig. S1A). An increase in depolarized mitochondria triggers the mitochondrial recruitment of phosphorylated Drp1 (Ser616), which induces mitochondrial fission.21 Sofosbuvir induced distinct mitochondrial fission in human hepatoma cells, whereas dimethyl sulfoxide–treated cells showed the typical tubular network of mitochondria (red color in the figures indicates TOM20-positive mitochondrial morphology) (Fig. 1B; Supporting Fig. S1B,C). Telaprevir and boceprevir also induced mitochondrial fragmentation (white color in Supporting Fig. S1B indicates MitoTracker-positive mitochondria morphology). Further, confocal microscopy revealed that the expression level of p-Drp1 (Ser616) was significantly higher in sofosbuvir-treated cells than in dimethyl sulfoxide–treated control cells as indicated by the p-Drp1 (Ser616) intensity immunostained green in Fig. 1B. More importantly, sofosbuvir promoted mitochondrial recruitment of p-Drp1 (Ser616), as evidenced by colocalization of p-Drp1 (Ser616) and TOM20 (yellow spots in Fig. 1B), which supported the increase in Drp1-mediated mitochondrial fission seen in sofosbuvir-treated cells. Telaprevir and boceprevir also induced mitochondrial translocation of p-Drp1 (Supporting Fig. S2). The sofosbuvir-induced increase in the expression level of p-Drp1 was further analyzed by western blot assays using whole-cell lysates extracted from hepatoma cells treated with sofosbuvir (accompanying graph in Fig. 1C). Together, these results indicate that HCV inhibitors such as sofosbuvir, telaprevir, and boceprevir induce loss of ΔΨ_m followed by mitochondrial translocation of p-Drp1 (Ser616), which promotes Drp1-mediated mitochondrial fission.

G-Rg3 INHIBITS HCV PROPAGATION

Cytoprotective effects of ginsenosides have been implicated in the treatment for many diseases including bacterial and viral infections.11 Furthermore, its effects on mitochondrial dynamics have also been reported.22, 23 We reasoned that these cytoprotective roles of ginsenosides may counter the adverse effects of antiviral DAAs seen in patients under treatment.

To investigate the antiviral effect of ginsenosides in HCV-infected cells, we first screened several representative ginsenoside compounds isolated from the root of Korean red ginseng (Panax ginseng C.A. Meyer).10 To conduct in vitro screening with ginsenoside compounds, we established an in vitro HCV infection system using HCVcc24 and Huh7.5.1, a highly permissive cell line for HCV infection.25 The infectivity of HCVcc in Huh7.5.1 cells was confirmed using confocal microscopy and immunostaining with an antibody specific to HCV core protein (Fig. 2A).

These analyses revealed that G-Rg3 remarkably suppresses the level of HCV RNA, as determined by real-time quantitative RT-PCR with primers specific to the HCV 5′-untranslated region (Fig. 2B). Also, western blot and cell viability assays showed that G-Rg3 reduces the expression level of HCV core protein in HCV-infected cells without cellular cytotoxicity (Fig. 2C,D). However, treatment of HCV-infected cells with G-Rh2, which is a protopanaxadiol type of ginsenoside like G-Rg3, induced very high cytotoxicity (Fig. 2E). These results suggest that G-Rg3 effectively inhibited HCV propagation.

G-Rg3 RESTORES HCV-INDUCED ABERRANT MITOCHONDRIAL FISSION

We have recently shown that HCV induces Drp1-mediated mitochondrial fission, which promotes robust HCV propagation.8 To examine an inhibitory mechanism of G-Rg3 in robust HCV infections, we analyzed the role of G-Rg3 in ΔΨ_m loss caused by HCV infection6, 26 because the HCV-induced loss of ΔΨ_m leads to mitochondrial fission.6, 8, 26 It is known that G-Rg3 inhibits the opening of mitochondrial permeability transition pores by free radical scavenging action.27 Consistent with our previous study,6 HCV infection decreased ΔΨ_m compared to uninfected cells (Fig. 3A).28, 29 Further, the HCV-induced loss of ΔΨ_m was remarkably restored by G-Rg3 treatment (third panel and accompanying graph in Fig. 3A). We also observed that G-Rg3 restored the ΔΨ_m loss caused by DAA treatment (Supporting Fig. S1A).

We next investigated if G-Rg3 affects HCV-induced aberrant mitochondrial fission. Consistent with our previous results,8 HCV infection induced mitochondrial fission (second panel in Fig. 3B), whereas confocal microscopic analysis of uninfected cells using immunostaining with MitoTracker displayed the typical tubular network of mitochondria (first panel in Fig. 3B). Interestingly, G-Rg3 treatment inhibited HCV-induced mitochondrial fission formation and displayed a normal tubular mitochondrial network (third panel in Fig. 3B). This effect is similar to silencing the mitochondrial fission factor or Drp1 and results in mitochondrial fusion status.8 The accompanying graph in Fig. 3B indicates the restoring effect of G-Rg3 using quantification data of the mitochondrial length during HCV infection. The present data demonstrate that G-Rg3 prevented the HCV-induced loss of ΔΨ_m and inhibited HCV-induced aberrant mitochondrial fission.

G-Rg3 SUPPRESSES HCV PROPAGATION BY DOWN-REGULATION OF Drp1

We have shown that HCV induces mitochondrial translocation of p-Drp1, leading to mitochondrial fission.8 Therefore, we used confocal microscopy to further examine whether G-Rg3 inhibits mitochondrial translocation of p-Drp1. As indicated by the yellow spots in the zoomed image of the third panel in Fig. 4A, HCV-infected cells showed activation and translocation of p-Drp1 onto fragmented mitochondria. In contrast, HCV-infected cells treated with G-Rg3 displayed no translocation of p-Drp1 on the tubular mitochondrial network (red mitochondrial morphology in the zoomed image of the fourth panel in Fig. 4A). The accompanying graph in Fig. 4A shows the quantitative colocalization of p-Drp1 and mitochondrial marker TOM20.

Drp1 translocation to mitochondria requires that Drp1 is phosphorylated at the Ser616 position. We analyzed the expression level of Drp1 (S616) using whole-cell lysates prepared from HCV-infected cells treated with G-Rg3. Analysis of these lysates shows that HCV-infected cells treated with G-Rg3 display decreased levels of Drp1 phosphorylation (Fig. 4B). There was also a modest reduction in the expression levels of Drp1 protein in these lysates. Together, these data suggest that G-Rg3 regulates Drp1 phosphorylation and subsequent mitochondrial translocation, thereby preventing HCV-induced abnormal mitochondrial fission.

Because HCV-induced mitochondrial fission is associated with virus secretion,8 we investigated if G-Rg3 affects HCV secretion. We first determined whether G-Rg3 inhibits viral replication in R-1 cells harboring HCV subgenomic replicon, supporting only HCV replication.13 G-Rg3 did not inhibit HCV replication of HCV subgenomic replicon (Fig. 4C). Thus, it is likely that the inhibitory effect of G-Rg3 is associated with viral secretion but not with replication. We further examined if G-Rg3 has any synergistic anti-HCV effect in combination with known HCV antivirals which inhibit HCV. We chose a 50 μM concentration of Rg-3, which does not affect viral replication, to evaluate its combinatorial synergic effect. Sofosbuvir and other nucleotide analogues used in this study inhibit viral replication. Rg3 treatment produced a synergic anti-HCV effect with sofosbuvir, telaprevir, and boceprevir and suppressed viral propagation (Fig. 4D).

G-Rg3 INHIBITS HCV-INDUCED DEGRADATION OF CYTOPLASMIC p21

HCV stimulates the activity of cyclin-dependent kinase 1 (CDK1), which promotes Drp1 phosphorylation at Ser616 in the cytoplasm.8 The subsequent mitochondrial translocation of p-Drp1 mediates mitochondrial fission during HCV infection.8, 21, 30 CDK1 activity is central to Drp1 phosphorylation at Ser616 in the cytosol during HCV infection.30 To examine how HCV promotes CDK1 activity, we analyzed the expression level of the cytoplasmic p21 protein, which directly inhibits CDK1 activity.31 The expression level of cytosolic p21 protein was strikingly lower in liver biopsy materials than in the control materials (see the light brown color in the cytoplasmic area in Fig. 5A). Cytoplasmic p21-positive cells are quantified in Fig. 5B, which clearly shows that HCV down-regulates cytoplasmic p21 expression to promote CDK1 activity.

To substantiate the role of G-Rg3 in down-regulated cytoplasmic p21 during HCV infection, we prepared a cytosolic fraction from HCV-infected cells and analyzed by western blot assays using an antibody specific to p21 protein. Figure 5C shows that HCV infection caused a decrease in the expression level of cytoplasmic p21. Surprisingly, this reduction was rescued by G-Rg3 treatment.

HCV core protein is a strong inhibitor of p21 expression.32 To further investigate the role of G-Rg3 in HCV core protein–dependent regulation of cytoplasmic p21 expression during HCV infection, we isolated the cytosolic fraction from cells transiently expressing HCV core protein and then analyzed the expression level of cytoplasmic p21 protein. Similar to the HCV-induced decrease in cytoplasmic p21 expression shown in Fig. 5C, HCV core protein dramatically reduced the expression level of cytosolic p21 (Fig. 5D). This reduction was again rescued by G-Rg3 treatment. These data collectively suggest that restoration of cytosolic p21 by G-Rg3 during HCV infection (Fig. 5C) was not due to inhibition of HCV propagation by G-Rg3 but that G-Rg3 reversed the reduction in HCV core protein–dependent cytoplasmic p21 expression during HCV infection. In summary, G-Rg3 increased p21 and decreased CDK1 activity, which has a direct bearing on Drp1 phosphorylation and mitochondrial fission.

G-Rg3 INHIBITS HCV-INDUCED ABNORMAL MITOPHAGY

Mitochondrial fission usually triggers mitophagy orchestrated by activation of PINK1 and Parkin recruitment to the mitochondria.33 To investigate the effect of G-Rg3 on HCV-induced mitophagy, we used a novel vector containing mRFP-enhanced GFP (EGFP) dual luciferase gene fused in frame with a mitochondrial targeting sequence (mito-mRFP-EGFP), as described (Fig. 6A).8, 15, 34 The hybrid protein displays a yellow image when expressed in mitochondria. However, if mitochondria fuse with lysosomes to complete the process of mitophagy, only RFP (red) is expressed as EGFP (green) is unstable in the acidic environment of lysosomes and quenched off.35, 36 In cells undergoing complete mitophagy, red puncta are visible, indicating expression of this hybrid protein in the lysosomes.35, 36 Consistent with our previous observation,8, 15 HCV infection induced mitochondrial fission and mitophagy (Fig. 6B, middle panel, and C). However, in the presence of G-Rg3, mitophagy was inhibited (Fig. 6B, lower panel). During HCV infection, Parkin, the key mediator of mitophagy, recruited to the mitochondria ubiquitinates and degrades outer mitochondrial substrates like VDAC1 and Mfn2.9 To determine the function of G-Rg3 in mitochondrial Parkin activity during HCV infection, we performed western blot analysis of cellular lysates using specific VDAC1 and Mfn2 antibodies. G-Rg3 prevented ubiquitin degradation of VDAC1 and Mfn2 proteins (Fig. 6D and accompanying graph). We further show that G-Rg3 inhibited the HCV-induced decreases in the gene expression of mitochondrial DNA, ND2, and COX2 proteins (Fig. 6E).

The effect of G-Rg3 on innate immune responses during HCV infection was also examined (Supporting Fig. S3). G-Rg3-treated cells showed comparable antiviral responses with nontreated control during HCV infection. Therefore, it is highly likely that G-Rg3 suppresses HCV propagation not by immune responses but by performing a protective role against HCV-induced mitophagy. These results collectively suggest that G-Rg3 rescues HCV-infected cells from mitophagy.