Interferon-inducible cholesterol-25-hydroxylase restricts hepatitis C virus replication through blockage of membranous web formation

Authors

  • Anggakusuma,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Inés Romero-Brey,

    1. Department of Infectious Diseases, Molecular Virology, Heidelberg University, Heidelberg, Germany
    Search for more papers by this author
  • Carola Berger,

    1. Department of Infectious Diseases, Molecular Virology, Heidelberg University, Heidelberg, Germany
    Search for more papers by this author
  • Che C. Colpitts,

    1. Departments of Biochemistry and of Medical Microbiology and Immunology and Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Tujana Boldanova,

    1. Department of Biomedicine, University of Basel and Division of Gastroenterology and Hepatology, University Hospital Basel, Basel, Switzerland
    Search for more papers by this author
  • Michael Engelmann,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Daniel Todt,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Paula Monteiro Perin,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Patrick Behrendt,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    2. Department of Gastroenterology, Hepatology, and Endocrinology, Medical School Hannover, Hannover, Germany
    Search for more papers by this author
  • Florian W.R. Vondran,

    1. ReMediES, Department of General, Visceral, and Transplantation Surgery, Hannover Medical School, and German Centre for Infection Research, Hannover-Braunschweig, Hannover, Germany
    Search for more papers by this author
  • Shuting Xu,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Christine Goffinet,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Luis M. Schang,

    1. Departments of Biochemistry and of Medical Microbiology and Immunology and Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Markus H. Heim,

    1. Department of Biomedicine, University of Basel and Division of Gastroenterology and Hepatology, University Hospital Basel, Basel, Switzerland
    Search for more papers by this author
  • Ralf Bartenschlager,

    1. Department of Infectious Diseases, Molecular Virology, Heidelberg University, Heidelberg, Germany
    2. German Center for Infection Research, Heidelberg University, Heidelberg, Germany
    Search for more papers by this author
  • Thomas Pietschmann,

    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    Search for more papers by this author
  • Eike Steinmann

    Corresponding author
    1. Institute of Experimental Virology, Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany
    • Address reprint requests to: Dr. Eike Steinmann, Institute of Experimental Virology, Twincore Center for Experimental and Clinical Infection Research, Feodor-Lynen-Straße 7-9, 30625 Hannover, Germany. E-mail: eike.steinmann@twincore.de; tel: +49 511 220027 130; fax: +49 511 220027 139.

    Search for more papers by this author

  • Potential conflict of interest: T.P. has received consulting fees from Biotest AG and from Janssen Global Services.

  • Supported by grants from the Helmholtz Association (SO-024) and the Deutsche Forschungsgemeinschaft (PI 734/2-1 and SFB 900, Teilprojekt A6; both to T.P.); by an intramural young investigator award of the Helmholtz Centre for Infection Research and the Deutsche Forschungsgemeinschaft (STE 1954/1-1, to E.S.); by the Swiss National Science Foundation (grant 310030B_147089, to M.H.H.; grant 323530_145255, to T.B.); and by the Deutsche Forschungsgemeinschaft (FOR1202, TP1, to R.B.). L.M.S. is a Burroughs Wellcome Fund (BWF) Investigator in the Pathogenesis of Infectious Disease. This work was supported in part by the BWF and the Canadian Institutes of Health Research (CIHR).

Abstract

Hepatitis C virus (HCV) is a positive-strand RNA virus that primarily infects human hepatocytes. Infections with HCV constitute a global health problem, with 180 million people currently chronically infected. Recent studies have reported that cholesterol 25-hydroxylase (CH25H) is expressed as an interferon-stimulated gene and mediates antiviral activities against different enveloped viruses through the production of 25-hydroxycholesterol (25HC). However, the intrinsic regulation of human CH25H (hCH25H) expression within the liver as well as its mechanistic effects on HCV infectivity remain elusive. In this study, we characterized the expression of hCH25H using liver biopsies and primary human hepatocytes. In addition, the antiviral properties of this protein and its enzymatic product, 25HC, were further characterized against HCV in tissue culture. Levels of hCH25H messenger RNA were significantly up-regulated both in HCV-positive liver biopsies and in HCV-infected primary human hepatocytes. The expression of hCH25H in primary human hepatocytes was primarily and transiently induced by type I interferon. Transient expression of hCH25H in human hepatoma cells restricted HCV infection in a genotype-independent manner. This inhibition required the enzymatic activity of CH25H. We observed an inhibition of viral membrane fusion during the entry process by 25HC, which was not due to a virucidal effect. Yet the primary effect by 25HC on HCV was at the level of RNA replication, which was observed using subgenomic replicons of two different genotypes. Further analysis using electron microscopy revealed that 25HC inhibited formation of the membranous web, the HCV replication factory, independent of RNA replication. Conclusion: Infection with HCV causes up-regulation of interferon-inducible CH25H in vivo, and its product, 25HC, restricts HCV primarily at the level of RNA replication by preventing formation of the viral replication factory. (Hepatology 2015;62:702–714)

Abbreviations
CH25H

cholesterol 25-hydroxylase

CHC

chronic hepatitis C

2′CMA

2′C-methyl-adenosine

DMV

double membrane vesicle

EFV

efavirenz

GFP

green fluorescent protein

HA

hemagglutinin

25HC

25-hydroxycholesterol

hCH25H

human CH25H

HCV

hepatitis C virus

IFN

interferon

ISG

interferon-stimulated gene

mRNA

messenger RNA

MW

membranous web

OSBP1

oxysterol-binding protein 1

PHH

primary human hepatocyte

SGR

subgenomic replicon

SREBP

sterol regulatory element binding protein

TCID50

50% tissue culture infective dose

TEM

transmission electron microscopy

Hepatitis C virus (HCV), a member of the Flaviviridae, is a positive-strand RNA virus that primarily infects human hepatocytes. Worldwide, an estimated 80 million people are chronically infected with HCV and are at high risk for developing severe liver damage, including hepatic steatosis, fibrosis, cirrhosis, and hepatocellular carcinoma.[1, 2] For the past 25 years, therapy has consisted of treatment with interferon (IFN)–alpha and the nucleoside analogue ribavirin.[3] Recently, the licensing of directly acting antivirals targeting HCV nonstructural proteins has improved cure rates profoundly, now reaching levels of over 90%.[4] However, many infected individuals have not been diagnosed, and a prophylactic vaccine is not available, which likely is required when considering global control and even eradication of HCV.

The first line of immune defense against HCV is based on cell-intrinsic innate immunity in the liver cells, which leads to the induction of type I and type III IFN systems. These cytokines induce a plethora of genes that exert a strong antiviral effect. In addition, the IFN system is required for efficient activation of the adaptive immune response.[5, 6] Only in a fraction of patients can HCV be cleared, and it becomes chronic in the majority of individuals. This has been linked primarily to excessive preactivation of the IFN response and dysfunctionality of T-cell responses.[5]

The production of IFN induces the expression of hundreds of IFN-stimulated genes (ISGs). Recent screening approaches have been used to identify a broader range of antiviral effectors against different viruses.[6-10] Although in some cases individual ISGs can have profound effects on the replication of certain viruses, such as the MX protein and influenza virus,[11] in the majority of cases the IFN system works in a combinatorial fashion, with multiple ISGs contributing to the antiviral response. The modes of action of only a few ISGs have been well elucidated, including protein kinase R, the 2-5 OAS/RNaseL system, viperin,[12] IFITM1,[13] IFI6,[14] and MX proteins.[15] Using an overexpression screening system for IFN-induced antiviral genes, Liu et al. discovered cholesterol-25-hydroxylase (CH25H), which converts cholesterol into the oxysterol 25-hydroxycholesterol (25HC), as an ISG with antiviral potency against murine gamma herpesvirus 6 (MHV6) and vesicular stomatitis virus.[16] In addition, it has been reported that macrophages and dendritic cells express CH25H in response to toll-like receptor activation.[17, 18] In a more recent study, Liu et al. also described the role of the hydroxylase enzymatic product, 25HC, as a broad-spectrum antiviral that blocks fusion of virions into target cells independently of its function as regulator of sterol regulatory element binding protein (SREBP).[19] Using a mass-spectrometry approach, 25HC was also shown to be up-regulated in murine macrophages as a result of viral infection or IFN treatment. It had antiviral activities against a panel of enveloped viruses in a liver X receptor–independent, but SREBP-dependent, mechanism.[20] The antiviral activity was reported to act mainly at a postentry step, suggesting that 25HC might interfere with multiple steps in the life cycle of viruses.[6, 20]

So far, the role of CH25H as an ISG and antiviral restriction factor has mainly been studied in murine model systems, and its effects on the HCV replication cycle have not been fully elucidated. Amino acid sequence alignment of murine and human cholesterol 25-hydroxylases shows 78% sequence identity, with the most notable difference being a 26–amino acid residue extension that is only present at the carboxy terminus of the murine enzyme.[21] In this study, we characterized the expression of human CH25H (hCH25H) in liver biopsies of chronic HCV-infected patients and elucidated its IFN dependence in primary human cells. The antiviral properties of hCH25H and the enzymatic product 25HC were also characterized against all HCV genotypes, in hepatoma cells and primary human hepatocytes (PHHs). Collectively, hCH25H was identified as an up-regulated ISG in human liver cells that, through its product 25HC, primarily targets early biogenesis of HCV-induced membranous replication factories.

Materials and Methods

Liver Biopsies and Informed Consent

The viral load, HCV genotype, and liver biopsies from patients with chronic hepatitis C (CHC, n = 34) and 10 HCV-negative liver biopsies were obtained in the context of routine diagnostic workup. Several liver biopsies have been reported.[22] Grading and staging of CHC were performed according to the METAVIR classification. All patients gave written informed consent in accord with local ethical committees.

Cell Culture

Human macrophages were obtained from human peripheral blood mononucleated cells which were plated in six-well plates at a density of 8 × 106 cells/well and kept in 10% AB-human serum and 10% fetal bovine serum containing Roswell Park Memorial Institute medium. At 10-17 days postseeding, undifferentiated cells and remaining lymphocytes were removed by thorough washing with prewarmed medium. The PHHs were isolated from liver specimens obtained after partial hepatectomy, plated at a density of 1.3 × 106 on collagen in P6 dishes, and kept in hepatocyte culture medium (Lonza) as described.[23] Huh7-Lunet cells containing persistent, selectable reporter subgenomic replicons (SGRs) of genotype 2a (LucUbiNeo_JFH1) have been described.[24] Huh7 cells containing a stable, hygromycin-selectable dengue virus SGR with Renilla luciferase reporter (dengue virus R2H; details of cloning and generation of the cell line will be described elsewhere) were used. Huh7.5 cells were cultured as described.[25]

Transmission Electron Microscopy

Cells were examined by transmission electron microscopy (TEM) as described.[26] Briefly, Huh7-Lunet/T7 or Huh7-Lunet cells transfected either with an HCV NS3-5B expression construct or with a full-length Jc1 genome were treated with three different concentrations of 25HC, chemically fixed, dehydrated, and resin-embedded. Ultrathin sections (70 nm) were contrasted with lead citrate and examined by TEM. Double membrane vesicle (DMV) numbers were counted in five cell profiles per condition, and a total of 90 DMVs were measured to obtain an estimation of their amount and diameter.

Additional material and methods are posted as Supporting Information.

Results

Human CH25H Messenger RNA is Up-Regulated in Liver Biopsies of CHC Patients and HCV-Infected PHHs

To study the regulation of hCH25H in response to HCV infection in humans, we compared messenger RNA (mRNA) expression levels in CHC liver biopsies from 34 patients with 10 liver biopsies from patients with non-HCV chronic liver diseases (Supporting Table S1 for patient data). Using this approach, we observed significantly higher mRNA expression levels of hCH25H in CHC liver biopsies compared to HCV-negative samples (Fig. 1A). Similar results were obtained for hIFIT3, a common marker for ISG induction (Fig. 1B).[8] Induction of hCH25H mRNA was independent of patient gender, HCV genotype, degree of liver inflammation and fibrosis, or viral load (Supporting Fig. S1). To confirm these findings in an ex vivo HCV infection model, PHHs were infected with cell culture–derived Jc1 virus and 48 hours later infection levels were determined by quantification of newly produced infectious virus with the nucleoside polymerase inhibitor 2′C-methyl-adenosine (2′CMA) serving as control (Fig. 1C). To analyze mRNA expression levels of hCH25H, total cellular RNA was extracted for quantitative real-time polymerase chain reaction analysis (Fig 1D,E). De novo produced viruses yielded titers of about 103 50% tissue culture infective dose (TCID50) per milliliter, whereas the 2′CMA treatment completely blocked HCV production (Fig. 1C). A significant up-regulation of hCH25H mRNA expression was detected and completely dependent on viral replication as 2′CMA treatment abrogated this effect (Fig. 1D). A similar phenomenon was again noted for hIFIT3 and hIFNβ (Fig. 1E,F).

Figure 1.

Messenger RNA expression of hCH25H in chronic hepatitis C liver biopsies and HCV-infected primary human hepatocytes (PHHs). Comparison of (A) hCH25H and (B) hIFIT3 relative mRNA expression in liver biopsies from patients with non-HCV chronic liver diseases (n = 10) or chronic hepatitis C (n = 34). The mRNA expressions were measured relative to human GAPDH mRNA as determined by quantitative real-time polymerase chain reaction. Asterisks indicate a significant difference as determined by the unpaired Student t test with Welch's correction (***P < 0.001). (C) Levels of HCV infection in PHHs at 48 hours postinfection were determined by the 50% tissue culture infectious dose per milliliter. The PHHs were infected with either mock or HCV Jc1 in the presence of dimethyl sulfoxide (M) or 10 µM of 2′CMA. Relative mRNA expression of (D) hCH25H, (E) hIFIT3, and (F) hIFNβ in mock treatment, HCV-infected, and treated PHHs at 48 hours postinfection. The mRNA expressions were measured relative to hGAPDH mRNA as determined by quantitative real-time polymerase chain reaction. Infection with HCV was performed with a multiplicity of infection of 10. Experiments C-E are depicted as mean values ± standard deviation from six independent donors. Asterisks indicate a significant difference as determined by a one-way analysis of variance adjusted with Bonferroni's multiple comparison test (*P < 0.05). Abbreviations: 2′CMA, 2′C-methyl-adenosine; hCH25H, human cholesterol 25-hydroxylase; HCV, hepatitis C virus; hGAPDH, human glyceraldehyde 3-phosphate dehydrogenase; hIFN, human interferon; mRNA, messenger RNA; n.s., nonsignificant; TCID50, 50% tissue culture infective dose.

CH25H Is an ISG in PHHs

The expression of CH25H is up-regulated in response to type I and type II IFNs in murine macrophages and dendritic cells[16, 17, 19, 20, 27] and in murine liver cells infected with vesicular stomatitis virus.[19] To validate and extend these findings to humans, we incubated PHHs and macrophages with types I, II, and III IFNs for different lengths of time. Human CH25H mRNA was strongly and significantly induced by type I IFN in PHHs (Fig. 2A) and macrophages (Fig. 2B) and to a lesser extent by types II and III IFN treatment. To demonstrate that the cells were in general responsive to types I, II, and III IFN treatment, hIFIT3 mRNA was quantified in parallel and found to be highly induced in hepatocytes (Fig. 2C) and macrophages (Fig. 2D). The induction of other ISGs known to be specifically stimulated by different types of IFN, like hMX1 (types I and III), hIFIT1 (types I and III), hIRF1 (type II), hCXCL10 (type II), and hPLSCR1 (types I, II, and III), was also observed to control the efficiency of each IFN treatment (Supporting Fig. S2).[28, 29] Of note, the type I IFN-dependent induction of hCH25H occurred rapidly already after 4 hours of IFN treatment, followed by a decrease of mRNA expression with longer treatment duration in both cell types (Fig. 2A,B). These results demonstrate that hCH25H is transiently induced by type I IFN in PHHs and macrophages.

Figure 2.

Human CH25H is an early interferon-stimulation gene in PHHs and macrophages. Relative mRNA expression profile of hCH25H in (A) PHHs and (B) macrophages after treatment with interferon-α (1000 IU/mL), interferon-γ (1000 IU/mL), and interferon-λ1 (1000 ng/mL) at the indicated hours posttreatment. The hIFIT3 relative mRNA expression in (C) PHHs and (D) macrophages served as an interferon treatment control. At the indicated time points cells were harvested and total RNA was extracted. The mRNA expressions were measured relative to hGAPDH mRNA as determined by quantitative real-time polymerase chain reaction. All graph data are shown as mean values ± standard deviation from three independent donors. Asterisks indicate a significant difference as determined by a one-way analysis of variance adjusted with Bonferroni's multiple comparison test (***P < 0.001). Abbreviations: hCH25H, human cholesterol 25-hydroxylase; hGAPDH, human glyceraldehyde 3-phosphate dehydrogenase; h.p.t., hours posttreatment; MØ, macrophages; mRNA, messenger RNA; PHH, primary human hepatocyte.

Expression and Enzymatic Activity of hCH25H Exerts Antiviral Activity Against HCV Infection

To assess if hCH25H exerts an antiviral effect against HCV, we delivered and transiently expressed N-terminally hemagglutinin (HA)–tagged hCH25H or green fluorescent protein (GFP) as control through lentiviral gene transfer into permissive Huh7.5 cells and subsequently infected these cells with HCV (Fig. 3A). Infection with HCV and lentiviral transduction were evaluated at the single-cell level by monitoring HCV NS5A expression and GFP or HA-hCH25H expression using fluorescence microscopy. Treatment with the human immunodeficiency virus reverse transcriptase inhibitor efavirenz (EFV) served as a control for transduction and transgene expression. The analysis revealed that HCV Jc1 efficiently infected the Huh7.5 cells with transient GFP expression (first row of Fig. 3B). However, transient expression of hCH25H rendered the cells nonpermissive for viral infection (third row of Fig. 3B). This “viral exclusion” phenotype was specific for hCH25H-expressing cells as Huh7.5 cells treated with EFV could efficiently be infected by HCV Jc1 (fourth row of Fig. 3B). The EFV treatment itself had no effect on HCV infection as the compound-treated GFP-transduced cells could still be robustly infected by Jc1 despite the absence of GFP expression (second row of Fig. 3B). Transgene expression of hCH25H had no detectable effect on cell viability (Supporting Fig. S3A). Antiviral activity of hCH25H was also observed at the level of viral protein (NS3) accumulation (Fig. 3C) or in HCV infection experiments with reporter viruses upon multiplicity of infection–dependent transduction of hCH25H (Fig. 3D).

Figure 3.

Transient expression of hCH25H renders Huh7.5 cells resistant to HCV infection. (A) Schematic representation of the experimental setup used to transiently express hCH25H through lentiviral transduction into Huh7.5 cells. The absence or presence of EFV was used to control the gene transduction. (B) Detection of HCV antigen in Huh7.5 cells with transient expression of GFP or HA-hCH25H infected by HCV Jc1. At the endpoint of the experiment, cells were fixed and stained for HA-tag (green) and HCV-NS5A (red) expression for immunofluorescence analysis. Nuclear DNA was stained with 4′,6-diamidino-2-phenylindole (blue). (C,D) The effects of increasing lentiviral transduction dose on HCV infection. The Huh7.5 cells were transduced with different titers of lentivirus carrying GFP or HA-hCH25H (+ = low, ++ = middle, +++ = high titer doses), followed by HCV Jc1 or RLuc-Jc1 infection. At the endpoint of the experiment, the expression of GFP, HA-tag, and HCV-NS3 antigen from the HCV Jc1–infected cells was monitored by western blotting (C), whereas the infection level of RLuc-Jc1–infected cells was determined by Renilla luciferase activity assay (D). The luciferase result is shown as normalized percentage of infection values relative to mock GFP. (E,F) The effects of enzymatically inactive hCH25H expression on HCV infection. The Huh7.5 cells were transduced with equal titers of lentivirus carrying hCH25H wild type or the mutant form (QQ), followed by HCV Jc1 or RLuc-Jc1 infection. At the endpoint of the experiment, the expression of GFP, HA-tag, and HCV-NS3 antigen from the HCV Jc1–infected cells was monitored by western blotting (E), whereas the infection level of RLuc-Jc1–infected cells was determined by Renilla luciferase activity assay (F). The luciferase result is shown as normalized percentage of infection values relative to mock GFP. All HCV infections were performed with a multiplicity of infection of 0.05. Data are shown as mean values ± standard deviation of three independent experiments, and images are representative of two independent experiments. Asterisks indicate a significant difference as determined by a one-way analysis of variance adjusted with Bonferroni's multiple comparison test (**P < 0.01, ***P < 0.001). Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; EFV, efavirenz; GFP, green fluorescence protein; HA, hemagglutinin; hCH25H, human cholesterol 25-hydroxylase; HCV, hepatitis C virus; VSV, vesicular stomatitis virus; WT, wild type.

To test if the hCH25H enzymatic activity was responsible for the antiviral property of hCH25H, an inactive form of the hydroxylase, in which two clustered histidine residues were replaced by glutamine (H242Q and H243Q or QQ mutant),[21] was generated and expressed in Huh7.5 cells. The detection of closely spaced double bands by western blot, particularly for the QQ mutant, is probably caused by differentially glycosylated forms of the hydroxylase protein as described (Fig. 3E).[21] Although the transient expression of the QQ mutant still could significantly prevent HCV infection (∼50% HCV inhibition), the loss of enzymatic activity severely attenuated the antiviral property of hCH25H (Fig. 3E,F). Thus, the expression of active hCH25H protects human liver cells against HCV infection.

The Enzymatic Product of hCH25H, 25HC, Exhibits Antiviral Activity Against All HCV Genotypes

The enzyme CH25H catalyzes the conversion of cholesterol into 25HC with the presence of molecular oxygen as an additional substrate and nicotinamide adenine dinucleotide phosphate hydrogen as a cofactor (Fig. 4A).[30] To investigate if hCH25H indeed produced a soluble antiviral factor against HCV, HEK 293T cells were transfected with either GFP, HA-hCH25H, or HA-hCH25H QQ mutant and the supernatant was used to cotreat HCV particles in an infection assay (Fig. 4B). The conditioned media from hCH25H-expressing 293T cells inhibited HCV infection by ∼80% compared to the GFP-control and QQ mutant (Fig. 4C), indicating that expression of hCH25H induces a soluble antiviral factor that is released into the culture supernatant and that this phenomenon requires the enzymatic function of the protein. To further analyze if 25HC is the effector molecule responsible for the antiviral activity against HCV, we incubated chimeric HCV reporter virus particles from all seven genotypes with increasing doses of 25HC for 4 hours, followed by a medium change without 25HC. Viral infectivity was measured 48 hours later. The addition of 25HC during infection of naive Huh7.5 cells with HCV resulted in a dose-dependent inhibition of infectivity with no overt cytotoxicity during these 4 hours (Fig. 4D; Supporting Fig. S2B). Of note, the presence of 25HC for a longer time period (48 hours) influenced cell viability at the highest doses of 25HC (Supporting Fig. S3B), which is in line with a previous report showing cytotoxicity and apoptosis induction of the oxysterol at higher concentrations.[27] The antiviral activity of 25HC against all major HCV genotypes was comparable, indicating that 25HC inhibits HCV independently of the genotype of the virus particles (Fig. 4D). The same inhibition was observed with HCV wild-type viruses without reporter in an immunofluorescence-based infection assay using HCV Jc1 and different concentrations of 25HC (Fig. 4E).

Figure 4.

The hCH25H enzymatic product, 25HC, inhibits HCV infection of Huh7.5 cells and primary human hepatocytes. (A) The CH25H catalyzes the synthesis of 25HC using cholesterol and molecular oxygen as substrates and nicotinamide adenine dinucleotide phosphate hydrogen as a cofactor. (B) Schematic representation of the experimental setup used to generate conditioned supernatant from HEK 293T cells expressing mock either GFP, HA-hCH25H, or HA-hCH25H QQ mutant. (C) Infection levels of HCV Rluc-Jc1 into Huh7.5 cells under the influence of the conditioned supernatants. Asterisk indicates a significant difference as determined by paired Student t test (**P < 0.01). (D) Infectivity of all genotypes of chimeric HCV into Huh7.5 cells in the presence of increasing concentrations of 25HC. The inoculum was removed 4 hours later, then monolayers were washed three times with phosphate-buffered saline and overlaid with fresh medium. Infected cells were fixed 2 days postinfection, and viral infectivity was determined by Renilla luciferase activity assay. All HCV infections were performed with a multiplicity of infection of 0.05. The result is shown as normalized percentage of infection values relative to mock dimethyl sulfoxide treatment. (E) Detection of HCV antigen in Huh7.5 cells infected by HCV Jc1 in the presence of the indicated amounts of 25HC. Infected cells were fixed 2 days postinfection, stained for HCV-NS5A (red) expression, and analyzed by immunofluorescence microscopy. (F) The infection level of HCV Jc1 (multiplicity of infection 10) on primary human hepatocytes in the presence of dimethyl sulfoxide, 25HC (12.5 µM), or 2′CMA (10 µM). The supernatants of infected primary human hepatocytes were collected at the indicated hours posttreatment and measured by TCID50 assay. The 2′CMA treatment served as a replication control. All graph data are shown as a representative experiment out of three independent repetitions, and images are representative of two independent experiments. Abbreviations: 2′CMA, 2′C-methyl-adenosine; DMSO, dimethyl sulfoxide; GFP, green fluorescence protein; HA, hemagglutinin; 25HC, 25-hydroxycholesterol; hCH25H, human cholesterol 25-hydroxylase; HCV, hepatitis C virus; NADPH, nicotinamide adenine dinucleotide phosphate; TCID50, 50% tissue culture infective dose; WT, wild type.

As PHHs more closely resemble the natural host for HCV infection in humans, we next performed HCV infection experiments in PHHs. Cells were infected for 6 hours with HCV Jc1 in the presence of 25HC, and newly released infectious virus particles were quantified by TCID50 assay 24 and 48 hours later. At both time points, 25HC reduced de novo production of infectious particles by ∼10-fold (Fig. 4F). The NS5B polymerase inhibitor 2′CMA was used as a control for blocking HCV RNA replication and thus infectious virus particle production.

25HC Acts on the Target Cells but Does not Affect Virion Integrity

To analyze whether the inhibitory effect of 25HC is caused by an alteration of the target cells rendering it HCV-resistant, by an effect on the virus particles, or by a postentry effect, we administered 25HC before (pretreatment), during (cotreatment), or after (posttreatment) infection of Huh7.5 cells with HCV reporter viruses. Pretreatment of cells inhibited HCV infection efficiently, and the same was found with co- or postadministration in a dose-dependent manner (Fig. 5A), indicating that 25HC acts on the host cell and presumably inhibits HCV primarily at a postentry step. We next corroborated this conclusion using HCV pseudoparticles, which are retroviral cores harboring the HCV glycoproteins in their envelope. Using this approach, 25HC showed only a slight inhibition of HCV pseudoparticle cell entry, which was, however, not dose-dependent and much less efficient compared to anti-CD81 antibodies serving as a positive control (Fig. 5B). To test the effects of 25HC solely on virus envelope–cell membrane fusion, R18-labeled HCV particles were exposed to 25HC or vehicle prior to mixing with Huh7.5 cells. Fusion was evaluated by fluorescence dequenching of R18.[31] The 25HC did not inhibit HCV fusion when the virions were preexposed, but fusion was partially inhibited when the cells were preexposed to 25HC (Fig. 5C). Fusion was more strongly inhibited when the cells were treated after virion binding (Fig. 5C), suggesting that 25HC acts on cellular, and not viral, membranes.

Figure 5.

The oxysterol 25HC acts on the target cells restricting hepatitis C viral (HCV) infection. (A) The Huh7.5 cells were pretreated before HCV RLuc-Jc1 (multiplicity of infection 0.05) infection, cotreated during infection, or posttreated after infection with the indicated concentrations of 25HC for 4 hours. Each step was always followed by cell washing with phosphate-buffered saline to remove the compound. After each final step, the cells were incubated in fresh medium and Renilla luciferase activity was determined 2 days later. The result is shown as normalized percentage of infection values relative to mock dimethyl sulfoxide treatment. (B) Effect of 25HC treatment on HCV pseudoparticles. The Huh-7.5 cells were infected with pseudoparticles bearing A-MLV, HCV genotype 1a (H77), 1b (Con1), or 2a (J6) envelope glycoprotein in the presence of different concentrations of 25HC. The inoculum was removed 4 hours later, and monolayers were washed and then overlaid with fresh medium. Infected cells were fixed 2 days later, and Firefly luciferase activity was determined. (C) The R18-labeled HCV JFH-1 virions or Huh7.5 cells were preexposed to 20 μM 25HC for 10 minutes at 37°C. Alternatively, virions were allowed to attach to Huh7.5 cells at 4°C for 1 hour, followed by exposure to 20 μM 25HC for 10 minutes at 37°C. Fusion was triggered by increasing the temperature and lowering the pH and monitored by fluorescence dequenching of R18 at the indicated time points. All graph data are shown as a representative experiment out of three independent repetitions. Abbreviations: 25HC, 25-hydroxycholesterol; MLV, murine leukemia virus; PP, pseudoparticles.

25HC Inhibits HCV RNA Replication by Blocking Membranous Web Biogenesis

To investigate the postentry effect of 25HC and its influence on HCV RNA replication, Huh7.5 cells were transfected with a reporter HCV-Con1 (genotype 1b) or -JFH1 (genotype 2a) SGR, followed by 48-hour incubation in the presence of increasing but nontoxic concentrations of 25HC. Replication efficiency was assessed using luciferase activity assays at the end time point. The 25HC efficiently inhibited HCV RNA replication in a dose-dependent manner (Fig. 6A,B). As HCV-Con1 and -JFH1 SGR replicate in different kinetics, we also performed a time-dependent experiment with 25HC against each SGR RNA replication, which was then assessed using direct quantitative real-time polymerase chain reaction on the RNA in comparison to the luciferase assay results (Supporting Fig. S4). Interestingly, HCV replication in stable cell lines was several orders of magnitude more sensitive to 25HC compared to dengue virus replicon–containing cells and hepatitis E virus replication (Supporting Fig. S5). In line with reduced RNA replication, 25HC treatment inhibited the production of released and cell-associated infectious HCV particles in a similar manner (data not shown).

Figure 6.

The oxysterol 25HC inhibits HCV subgenomic RNA replication and Jc1-induced membranous web formation. The effect of 25HC treatment is shown against HCV Con1 (genotype 1b) (A) or JFH1 (genotype 2a) (B) SGR replication in Huh7.5 cells. The cells were transfected with the SGR RNA encoding a Firefly luciferase reporter, and 4 hours later 25HC or 2′CMA was added. Replication of HCV RNA was determined by luciferase assays 2 days later. The result is shown as normalized percentage of replication values relative to mock dimethyl sulfoxide treatment. All graph data are shown as mean values ± standard deviation of three independent experiments. (C) Scheme of the experiment to analyze the effect of 25HC on the formation of DMVs, the main constituent of the membranous web. Two days after electroporation with JC1, Huh7-Lunet cells were treated with increasing concentrations of 25HC for 12 hours. (D) Abundance of NS5A was determined by western blot. (E) Electron micrographs of cells treated with dimethyl sulfoxide (left panel) or with 6.25 µM 25HC (middle panel) or 12.5 µM 25HC (right panel). Lower panels show higher magnification images of the areas highlighted in the upper panels with a dashed white square. (F) Quantification of the number of DMVs per square micrometer and DMV diameter in mock-treated cells versus cells treated with 6.25 and 12.5 µM 25HC. The surface of five cell profiles was measured, and the number of DMVs within every cell profile was determined (upper panel). The size of a total of 90 DMVs/condition was measured (lower panel). P values shown in the graphs were calculated using an unpaired Students t test, ∗P < 0.01, ∗∗P < 0.001, ∗∗∗P < 0.0001. Abbreviations: 2′CMA, 2′C-methyl-adenosine; DMSO, dimethyl sulfoxide; DMVs, double membrane vesicles; ER, endoplasmic reticulum; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; 25HC, 25-hydroxycholesterol; HCV, hepatitis C virus; LD, lipid droplet; m, mitochondria; n.s., nonsignificant; SGR, subgenomic replicon.

It is known that 25HC is a ligand of the oxysterol-binding protein 1 (OSBP1),[32, 33] which, together with HCV NS5A and the host cell phosphatidylinositol-4 kinase IIIα,[34, 35] is a crucial component for the formation of the membranous HCV replication factory, termed the membranous web (MW).[36] Therefore, using TEM we analyzed the effect of 25HC on the biogenesis of the MW, which is mainly composed of DMVs.[35] In the initial set of experiments (Fig. 6C), we used Jc1-infected Huh7.5 cells that were treated with 25HC for time periods that did not affect the abundance of viral proteins (Fig. 6D). In mock-treated and HCV-infected hepatoma cells, high numbers of DMVs could be observed (Fig. 6E, left panel). However in HCV-infected cells treated with 25HC (Fig. 6E, middle and right panel), the amounts and diameter of DMVs were decreased in a dose-dependent manner (Fig. 6F, upper and lower panel). These results suggested that 25HC affects the biogenesis and/or integrity of the MW.

We have shown that expression of an NS3-5B polyprotein induces an MW independent of RNA replication with a morphology indistinguishable from the one found in HCV-infected cells.[35] To exclude that the effect on the MW we observed in Jc1-infected Huh7 cells was caused by replication inhibition rather than by a direct effect on web integrity, we treated NS3-5B expressing cells with 25HC and analyzed them by TEM (Fig. 7A). The 25HC treatment did not alter the HCV protein levels (Fig. 7B). In agreement with our assumption, 25HC treatment profoundly altered MW morphology, causing a significant reduction in number and size of DMVs and clustering these vesicles in confined areas (Fig. 7C,D). In summary, these data suggest that 25HC inhibits HCV by blocking MW formation in a replication-independent manner.

Figure 7.

Effect of 25HC treatment on the formation of HCV-induced double membrane vesicles. (A) Schematic representation of the workflow: Huh7-Lunet cells stably expressing T7 RNA polymerase (Huh7-Lunet/T7) were transfected with pTM_NS3-3′ construct encoding the replicase proteins and containing the 3′ untranslated region of HCV. Six hours later, cells were mock-treated either with dimethyl sulfoxide or with increasing concentrations of 25HC. (B) Abundance of HCV protein was determined by NS5A western blotting. (C) Representative electron micrographs of cells treated with dimethyl sulfoxide (left panel) or 12.5 µM 25HC (right panel). Lower panels show higher-magnification images of the areas highlighted with a dashed white square in the upper panels. (D) Quantification of DMVs per square micrometer (upper panel) and DMV diameter (lower panel) in mock-treated cells versus cells treated with 12.5 µM 25HC, respectively, as described in the legend of Fig. 6. Abbreviations: DMSO, dimethyl sulfoxide; DMVs, double membrane vesicles; ER, endoplasmic reticulum; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; 25HC, 25-hydroxycholesterol; HCV, hepatitis C virus; LD, lipid droplet; m, mitochondria.

Discussion

In this study, we characterized the ISG hCH25H and its enzymatic product 25HC in liver cells as an antiviral restriction factor against HCV. By investigating the role of this enzyme in human liver biopsies, we found a significant elevation of CH25H gene expression in CHC patients compared to HCV-negative controls (Fig. 1A). Because hCH25H's main biological role is to catalyze the production of the oxysterol 25HC, these findings are supported by a recent report showing increasing serum oxysterol concentrations, particularly 25HC, in patients with CHC infection that, however, were significantly reduced during antiviral therapy.[37] Human CH25H is normally poorly expressed in the healthy liver,[38] but the elevation of hCH25H during HCV infection observed in the liver biopsies could be also seen in HCV-infected PHH cultures (Fig. 1B). Interferons are the central cytokines responsible for the induction of an antiviral state in virus-infected cells and for the activation and regulation of the cellular components of innate immunity.[39] The rapid IFN induction of CH25H was initially identified in murine bone marrow– and lung-derived dendritic cells and macrophages in which toll-like receptor–mediated induction of this enzyme was observed through a mechanism that is dependent on signaling through the IFN-α/-β receptor and STAT1.18 We observed in primary human macrophages and hepatocytes an early and transient induction of hCH25H primarily after type I IFN stimulation (Fig. 2), in line with data from murine systems demonstrating that CH25H was more potently activated by type I IFN.[16, 19, 20] The short duration of this transcriptional stimulation upon IFN treatment, together with the significantly higher levels in CHC patients, argues for a constant stimulation in the infected liver. Moreover, the transient nature of hCH25H stimulation in PHH indicates the presence of a negative feedback regulation. Future studies are required to understand the underlying mechanism in more detail.

Ectopic expression of hCH25H by lentiviral gene transfer protected human hepatoma cells against HCV infection, and the enzymatic activity was crucial for the antiviral activity (Figs. 3A-D, 4). Another recent report described similar findings with hCH25H suppressing HCV infection and suggested that the mutant form of hCH25H could still inhibit HCV replication independently of its enzyme activity by directly interacting with and inhibiting NS5A dimerization.[40] As we observed blockage of MW formation by treatment with 25HC without hCH25H gene expression, the binding and disruption of NS5A dimerization by hCH25H seem to be an independent and additional mechanism of antiviral restriction. We also found a significant reduction of HCV replication by the hCH25H mutant, which was, however, significantly lower compared to enzymatically active hCH25H (Fig. 3E-F), suggesting that the 25HC plays the more important role in antiviral activity against HCV. Very recently, Xiang et al. also demonstrated that the induction of hCH25H represents an important host innate response against HCV infection and highlighted the role of 25HC as the host lipid regulator responsible for the anti-HCV activity.[41]

The product 25HC belongs to a diverse class of endogenous oxysterols that possess complex biological roles including inflammation.[42, 43] It regulates the activation of genes under the control of the SREBP transcription factor and the nuclear hormone receptor liver X receptor. Down-regulation of low-density lipoprotein receptor as one of the proteins regulated through the SREBP and liver X receptor pathway has been suggested to be the reason for the antiviral activity of 25HC against HCV.[44] However, the strong inhibitory effect observed under posttreatment conditions (Fig. 5A) and on SGRs (Fig. 6 and Chen et al.[40]) and the minor effect on HCV entry (Fig. 5B) suggest that the main mode of action is the inhibition of HCV RNA replication. Similar observations were initially reported by Pezacki et al., who described a down-regulation of many key genes involved in the mevalonate pathway by 25HC, leading to cholesterol depletion and an antiviral state against HCV Con1 replicons.[45] Using electron microscopy, in this study we observed a collapse of HCV-induced DMV formation upon 25HC treatment (Figs. 6, 7). It is known that HCV extensively remodels intracellular membranes, giving rise to an MW that is composed predominantly of endoplasmic reticulum–derived DMVs with an average diameter of ±150 nm.[35] Induction of DMVs does not require viral RNA replication but can be triggered by the sole expression of the HCV replicase proteins NS3-5B.[35, 36] It has been shown that NS5A and NS5B recruit phosphatidylinositol-4 kinase IIIα to endoplasmic reticulum–derived membranes, leading to a massive accumulation of phosphatidylinositol-4–phosphate at these sites.[46, 47] Phosphatidylinositol-4–phosphate enrichment likely facilitates the recruitment of cellular lipid transporters such as FAPP-2 and OSBP1, delivering sphingolipids and cholesterol to these membranes, respectively.[36, 48-50] Cholesterol in particular is thought to provide membrane stability to DMVs as cholesterol depletion induces shrinking of these membranes whereas knock-down of phosphatidylinositol-4 kinase IIIα causes in addition DMV clustering.[36, 51, 52] Interestingly, Strating et al. recently identified an inhibition of OSBP1-dependent cholesterol transfer activity caused by 25HC in in vitro liposomal assays.[53] These results suggest that 25HC acts as an OSBP ligand, preventing the delivery of cholesterol to DMVs and thus biogenesis and integrity of the MW, reminiscent to what has been described for poliovirus replication.[36, 49, 54] This hypothesis is also supported by the fact that 25HC showed only a minor inhibitory effect on dengue virus replication (Supporting Fig. S3), another plus strand RNA virus that requires host cholesterol but does not require OSBP1 for membrane rearrangements.[49, 55, 56]

In summary, we show that HCV infection causes up-regulation of the ISG hCH25H in PHHs and CHC patients. The product produced by this enzyme, 25HC, profoundly blocks HCV RNA replication by affecting the biogenesis of the membranous HCV replication factory.

Acknowledgment

We are grateful to Takaji Wakita and Jens Bukh for JFH1 and J6CF isolates, respectively; to Charles Rice for Huh-7.5 cells and the 9E10 monoclonal antibody; to Suzanne Emerson for the hepatitis E virus p6 clone; and to Wolfgang Fischl for establishing the subgenomic dengue virus replicon. Moreover, we thank Stephanie Pfaender for critical reading of the manuscript and all members of the Institute of Experimental Virology, Twincore, for helpful support, suggestions, and discussions.

Ancillary