Not interferon, but interleukin-6 controls early gene expression in hepatitis B virus infection


  • Marianna Hösel,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Department of Internal Medicine I, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Maria Quasdorff,

    1. Department of Gastroenterology and Hepatology, University Hospital Cologne, Köln, Germany
    2. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Katja Wiegmann,

    1. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Dennis Webb,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Uta Zedler,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Mathias Broxtermann,

    1. Institute of Virology, Technische Universität München/Helmholtz Zentrum München, München, Germany
    Search for more papers by this author
  • Raindy Tedjokusumo,

    1. Institute of Virology, Technische Universität München/Helmholtz Zentrum München, München, Germany
    Search for more papers by this author
  • Knud Esser,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Institute of Virology, Technische Universität München/Helmholtz Zentrum München, München, Germany
    Search for more papers by this author
  • Silke Arzberger,

    1. Institute of Virology, Technische Universität München/Helmholtz Zentrum München, München, Germany
    Search for more papers by this author
  • Carsten J. Kirschning,

    1. Institute of Medical Microbiology, Immunology and Hygiene, Technische Universität München, München, Germany
    Search for more papers by this author
  • Anja Langenkamp,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Christine Falk,

    1. Immunomonitoring Unit, National Center for Tumor Diseases (NCT) / German Cancer Research Center (DKFZ) / Institute of Immunology, University Hospital Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Hildegard Büning,

    1. Center for Molecular Medicine Cologne (ZMMK), University Hospital Cologne, Köln, Germany
    2. Department of Internal Medicine I, University Hospital Cologne, Köln, Germany
    Search for more papers by this author
  • Stefan Rose-John,

    1. Department of Biochemistry, Christian-Albrechts-Universität zu Kiel, Medical Faculty, Kiel, Germany
    Search for more papers by this author
  • Ulrike Protzer

    Corresponding author
    1. Institute for Medical Microbiology, Immunology and Hygiene, University Hospital Cologne, Köln, Germany
    2. Institute of Virology, Technische Universität München/Helmholtz Zentrum München, München, Germany
    • Institute of Virology, Technische Universität München/Helmholtz Zentrum München, Trogerstr. 30, D-81675 München, Germany
    Search for more papers by this author
    • fax: +49-89-41406823

  • See Editorial on Page 1692.

  • Potential conflict of interest: Nothing to report.


With about 350 million virus carriers, hepatitis B virus (HBV) infection remains a major health problem. HBV is a noncytopathic virus causing persistent infection, but it is still unknown whether host recognition of HBV may activate an innate immune response. We describe that upon infection of primary human liver cells, HBV is recognized by nonparenchymal cells of the liver, mainly by liver macrophages (Kupffer cells), although they are not infected. Within 3 hours, this recognition leads to the activation of nuclear factor kappa B (NF-κB) and subsequently to the release of interleukin-6 (IL-6) and other proinflammatory cytokines (IL-8, TNF-α, IL-1β), but does not induce an interferon response. The activation of proinflammatory cytokines, however, is transient, and even inhibits responsiveness toward a subsequent challenge. IL-6 released by Kupffer cells after activation of NF-κB controls HBV gene expression and replication in hepatocytes at the level of transcription shortly after infection. Upon binding to its receptor complex, IL-6 activates the mitogen-activated protein kinases exogenous signal-regulated kinase 1/2, and c-jun N-terminal kinase, which inhibit expression of hepatocyte nuclear factor (HNF) 1α and HNF 4α, two transcription factors essential for HBV gene expression and replication. Conclusion: Our results demonstrate recognition of HBV patterns by nonparenchymal liver cells, which results in IL-6-mediated control of HBV infection at the transcriptional level. Thus, IL-6 ensures early control of the virus, limiting activation of the adaptive immune response and preventing death of the HBV-infected hepatocyte. This pattern recognition may be essential for a virus, which infects a new host with only a few virions. Our data also indicate that therapeutic neutralization of IL-6 for treatment of certain diseases may represent a risk if the patient is HBV-infected. (HEPATOLOGY 2009:50:1773–1782.)

The liver plays an important role in the host defense against pathogens. It consists of parenchymal cells, hepatocytes, and nonparenchymal cells (NPC): Kupffer cells, liver sinusoidal endothelial cells (LSEC), and stellate cells (Ito cells). They are involved in antigen processing and the induction of immune responses.1, 2 Activation of the innate immune system relies on recognition of molecular patterns associated with pathogens.

Kupffer cells, the liver-resident macrophages, account for approximately one-third of NPC in the liver and constitute more than 70% of the macrophage population of the body. They are part of the innate immune system and effectively remove, e.g., lipopolysaccharides (LPS) from the systemic circulation. LSEC rapidly clear antigens from the blood through receptor-mediated endocytosis by pattern recognition receptors (PRR) such as mannose-binding proteins and scavenger receptors.3 Isolated Kupffer cells and LSEC have been shown to produce proinflammatory cytokines interleukin (IL)-6, and tumor necrosis factor (TNF) α, but also IL-1β, IL-10, and interferons.4

IL-6 is a pleiotropic cytokine with well-known beneficial effects for the liver. IL-6 activates intracellular pathways by way of a heterodimeric IL-6/gp130 receptor and activating intracellular signal transducer and activator of transcription 3 (STAT3). IL-6 promotes liver regeneration and protects against a multitude of liver-damaging influences such as alcohol and carbon tetrachloride intoxication. Recently, the IL-6-gp130-STAT3 pathway has been demonstrated to protect hepatocytes from T-cell-mediated damage.5

In HBV infection, the role of the adaptive immune response is well established, but our knowledge about the innate immune response is limited (reviewed6). In HBV transgenic mice, natural killer (NK) cells and NKT-cells challenged with α-galactosylceramide inhibit HBV replication by secreting interferon (IFN) γ.7 Activation of Toll-like receptors (TLRs) on NPC inhibits HBV in vivo in an IFNα/β-dependent manner.4, 8 All these effects are mediated by antiviral cytokines, mainly IFNs, but also TNF-α.

Notably, most of the HBV DNA is cleared from the serum and the livers of experimentally infected chimpanzees prior to a detectable adaptive immune response in the liver, implying a contribution of the innate immune system.9 Microarray analyses of serial liver biopsies of experimentally infected chimpanzees, however, revealed no detectable changes in the expression profile of intrahepatic genes within the first weeks of infection.10 Despite this, a role for the innate immune response in the control of early HBV replication should not be dismissed, because expression of immune-response genes might occur below the level of detection in total liver RNA.6

Whereas the role of IFNs and TNF-α in HBV control has been intensively studied, little is known about the proinflammatory cytokines such as IL-6 and IL-1β. The purpose of this study was to examine whether host recognition of HBV activates an innate immune response in primary human liver cells, and what consequences the release of cytokines upon pattern recognition of HBV may have for both the host cell and the virus.


EMSA, electrophoretic mobility shift assay; ERK, exogenous signal-regulated kinase; HBeAg, hepatitis B early antigen; HBV, hepatitis B virus; HNF, hepatocyte nuclear factor; IFN, interferon; IL-1β, interleukin 1 beta; IL-6, interleukin 6; JNK, c-jun N-terminal kinase; LSEC, liver sinusoidal endothelial cells; MAPK, mitogen-activated protein kinases; NPC, nonparenchymal liver cells; NF-κB, nuclear factor kappa B; PHH, primary human hepatocytes; PRR, pattern recognition receptor; RT-PCR, reverse transcription polymerase chain reaction; STAT, signal transducer and activator of transcription; TLR, Toll-like receptor; TNF-α, tumor necrosis factor alpha.

Materials and Methods


The NF-κB inhibitor peptide, AAVALLPAVLLALLAPVQRKRQKLMP, was purchased from Santa Cruz Biotechnology (Santa Cruz, CA), mitogen-activated protein kinases (MAPK) inhibitors PD98059, SB203580, and SP600125 from Calbiochem (La Jolla, CA), Alexa Fluor 488-conjugated Escherichia coli from Invitrogen (Paisley, UK). Human recombinant IL-611 and mouse monoclonal anti-IL-6 antibody (IL-6ab) have been described.12

Isolation and Culture of Primary Human Liver Cells.

Primary human hepatocyte (PHH) and NPC cultures were prepared from surgical human liver biopsies after informed consent of patients after a standard two-step collagenase perfusion by serial differential centrifugation at 50g (PHH) and 250g (NPC), respectively. PHH were seeded onto plastic dishes coated with collagen type IV, NPC onto collagen-free plates in Williams medium supplemented as described.13

HBV Infection.

HBV inoculum was concentrated from HepG2.2.15 (mock from HepG2) cell culture medium using centrifugal filter devices (Centricon Plus-80, Biomax 100.000, Millipore, Bedford, MA).14 On day 3 after seeding, PHH or NPC cultures were inoculated with mock-inoculum or with HBV at the indicated multiplicity of infection (moi), both in the presence of 5% PEG 6000 for 24 hours postinfection (p.i.). Hepatitis B surface antigen (HBsAg) and hepatitis B early antigen (HBeAg) secretion were determined using commercial immunoassays (Abbott Laboratories, Wiesbaden, Germany).

Quantification of Gene Expression by Real-Time Reverse-Transcription Polymerase Chain Reaction (RT-PCR).

Total RNA was extracted using Trizol reagent, DNase-digested, and transcribed into cDNA using First-Strand Synthesis Supermix (both Invitrogen). HBV pregenomic RNA (pgRNA) was detected as described.15 Real-time PCRs were performed using the LightCycler system. For messenger RNA (mRNA) detection, exon-exon-spanning primers were used (Supporting Table 1) and the results were normalized to a dilution series of calibrator complementary DNA (cDNA) using the Relative Quantification Software (all Roche Diagnostics, Mannheim, Germany). For IFN gene expression the RT2Profiler PCR Array (SA Biosciences, Frederick, MD) was used.

Analysis of Proteins.

For the simultaneous quantification of cytokines in culture supernatants, Luminex-based multiplex protein array was utilized (BioRad Laboratories, Hercules, CA). IL-6, IL-1β, IL-8, TNF-α, and IFNα were confirmed by enzyme-linked immunosorbent assay (ELISA). Intracellular proteins were analyzed by western blot.16 (For details, see Supporting Materials.)

Detection of NF-κB Activation.

For the electrophoretic mobility shift assay (EMSA), nuclear extracts from PHH were incubated with 32P-labeled double-stranded oligonucleotides containing the κB site of the HIV-1 LTR enhancer, separated by 6% nondenaturing polyacrylamide gel electrophoresis (PAGE) and detected by autoradiography. In addition, the activation of NF-κB was measured in nuclear proteins using the EZ-Detect Transcription Factor kit for NF-κB p50 (Pierce, Rockford, IL).


HBV Stimulates IL-6 Secretion from NPC Prior to HBV Replication in Hepatocytes.

To determine if HBV or its secretory antigens activate PRR on human liver cells, we prepared PHH and NPC cultures and measured cytokine induction upon incubation with HBV.

PHH cultures contained ≥85% hepatocytes and 3%-15% of LSEC and Kupffer cells.13 NPC cultures contained ≥85% NPC, mainly Kupffer cells (Fig. 1A,B). Real-time RT-PCR analysis confirmed a 5-fold higher expression of the macrophage-specific CD68 mRNA in NPC than in PHH. Kupffer cells phagocytosed fluorescent-labeled E. coli, which induced a phenotypic change from stellate to rounded cells within 2 hours (Fig. 1C,D).

Figure 1.

Kinetics of IL-6 secretion in mock- and HBV-infected PHH or NPC cultures. Phase contrast microscopy of PHH (A) and NPC (B) cultures on day 3 after seeding. The arrow indicates a Kupffer cell in the PHH culture. (C,D) Fluorescent microscopy of NPC at 2 hours after addition of green-fluorescent E. coli (A-C: 400-fold magnification; D: 100-fold magnification). (E) Cells were mock- or HBV-infected or inoculated with fluorescent bacteria on day 3 after seeding. IL-6 was measured by ELISA in cell culture supernatants at indicated timepoints. Median values ± SD of three measurements each of two independent experiments are given. (F) PHH cultures were either mock- or HBV-infected on day 3 after seeding. On day 6 p.i., cells were left untreated (−) or inoculated again with HBV or with fluorescent E. coli. IL-6 was measured by ELISA in cell culture supernatants at 24 hours after treatment. Values represent median ± SD of three measurements.

Both PHH and NPC rapidly started to secrete IL-6 within the first 5 hours after inoculation with HBV at an moi of 200 virions per cell, with maximal levels at 24 hours p.i. IL-6 induction was transient, equaling background levels again on day 3 p.i. E. coli-treated NPC secreted IL-6 with comparable kinetics (Fig. 1E). Mock-inocula did not induce IL-6.

Maximum levels of IL-6 were markedly higher in HBV-inoculated NPC (4,464.6 ± 294.5 pg/mL) than in PHH (2310.7 ± 128.6 pg/mL), although cell density in NPC cultures was lower. On day 4 p.i. we detected newly synthesized HBeAg (21.4 ng/mL) in HBV-infected PHH but not in NPC (detection limit 0.2 ng/mL; data not shown), confirming HBV gene expression and beginning virus replication in PHH, but not in NPC.

To prove that IL-6 was induced independently of HBV replication, we pretreated the HBV inoculum with 100 mJ of UV-light before infection of PHH. This reduced transcription of HBV pregenomic (pg) RNA by 86%, but secretion of IL-6 was not affected (1,069.5 ± 15.5 and 1153 ± 6 pg/mL for UV-treated and untreated HBV, respectively), indicating that neither an intact viral genome nor its transcription nor replication were required.

To exclude LPS contamination of our preparations, we incubated TLR4-MD-2-transfected 293 cells with our HBV or mock inocula, which did not lead to cell activation as compared to LPS used as positive control (data not shown). In order to analyze potential LPS-like tolerance effects of HBV, mock- or HBV-infected PHH on day 6 p.i. were inoculated again with HBV or challenged with fluorescent-labeled E. coli particles. Mock-infected PHH secreted IL-6 at levels comparable to those in 3-day-old PHH stimulated with HBV or E. coli (Fig. 1E), but PHH previously infected with HBV secreted barely detectable amounts of IL-6 (Fig. 1F). Thus, once activated after HBV infection, NPC seem not to be capable of pathogen recognition and IL-6 production any more.

Proinflammatory Cytokines, but Not Interferons Are Induced After Recognition of HBV.

To determine whether other cytokines besides IL-6 were induced upon cellular HBV recognition, we analyzed cell culture supernatants from HBV-infected PHH or NPC for the presence of proinflammatory cytokines and interferons by ELISA and Luminex-based multiplex protein array technique (Supporting Table 2), and determined their mRNA expression by quantitative RT-PCR (Table 1, and Supporting Table 3). IL-8 (PHH: 77,600 ± 320 pg/mL, NPC: 22,100 ± 270 pg/mL), IL-1β (PHH: 157.4 ± 22.2 pg/mL, NPC: 1,862.7 ± 122.4 pg/mL) and TNF-α (PHH: 262.5 ± 23.6 pg/mL, NPC: 1,189.5 ± 171.5 pg/mL), but no IFN, were secreted by NPC with kinetics very similar to that of IL-6 (data not shown). IL-6, IL-1β, and TNF-α gene expression was up-regulated mainly in NPC, IL-8 mainly in PHH (Table 1). After contact with HBV, a ≤2-fold induction of IFN or IFN-related genes was detected (Supporting Table 3). In HBV-infected PHH, IFNβ, IFN-inducible 2′5′-oligoadenylatsynthetase (2′5′OAS) and IFN-inducible protein 10 (IP10) genes were 6-, 5.9-, and 6.3-fold down-regulated (Table 1), whereas they were 3- to 47-fold up-regulated upon exposure to E. coli (data not shown).

Table 1. Alteration of Cellular Gene Expression in HBV-Infected Primary Liver Cells
Cell CultureFold Change in Gene Expression, HBV Versus Mock
  1. ↑, up-regulated gene; ↓, down-regulated gene; nc, no change in gene expression detected.

PHH74 ↑nc11.5 ↑8 ↑24 ↑6 ↓5.9 ↓6.3 ↓
NPC137 ↑7 ↑36 ↑5.5 ↑nc≤2 (↑)nc≤2 (↑)

Taken together, IL-6, IL-1β, and TNF-α were induced rapidly after contact with nonhepatocytes, most probably Kupffer cells, with HBV, whereas it was neither induced by nor depended on HBV replication. In contrast, HBV induced no interferon response, and interferon-regulated genes were even down-regulated.

Activation of NF-κB in Primary Liver Cells Early After HBV Infection.

Our data indicated that NPC might recognize HBV particles by PRR. Because this led to release of primarily NF-κB-regulated cytokines, we examined whether NF-κB was activated upon contact with HBV.

Nuclear extracts from HBV-infected, but not from mock-infected PHH cultures exhibited NF-κB binding activity by EMSA on day 1 but not on day 4 p.i. (Fig. 2A). In cells inoculated at low moi (20 virions/cell), DNA binding of NF-κB was 5-fold increased 3 hours p.i. and 24 hours p.i., whereas at later timepoints and in mock samples no activated NF-κB was detected by NF-κB-DNA binding assay. NF-κB competitor oligonucleotides confirmed specificity of the assay (Fig. 2B). Addition of NF-κB inhibitory peptide (16.6 μg/mL) to PHH 2.5 hours before HBV infection resulted in an 81% reduced activation of NF-κB and 50% reduced secretion of IL-6 (data not shown).

Figure 2.

Activation of NF-κB upon HBV infection of primary liver cells. (A) Electro-mobility shift assay of nuclear extracts (5 μg protein per lane) prepared from mock- or HBV-infected PHH cultures at day 1 and 4 p.i. using 32P-labeled oligonucleotides containing the NF-κB binding site of the HIV-1 LTR. Nuclear extracts from HeLa cells stimulated with 10 ng/mL TNF-α were used as a positive control (+). (B,C) chemiluminescent ELISA-based NF-κB assay of 2 μg of nuclear proteins isolated at timepoints indicated from mock- or HBV-infected PHH or NPC cultures set up in parallel. Forty pmol of a competitor oligonucleotide duplex (+C) or a mutant competitor (+M) were added to ensure signal specificity. 10 μL of NER-protein extraction buffer were used to measure background luminescence (buffer). Nuclear extracts from E. coli-treated NPC were used as a positive control. Values are shown as median ± SD of three measurements.

NPC inoculated with HBV at moi 20 showed a 10.5-fold increase of NF-κB binding 3 hours after inoculation, whereas fluorescent-labeled E. coli (100 particles per cell) activated NF-κB only 6.6-fold.

Taken together, our data demonstrate a transient activation of NF-κB in primary human liver cell cultures, mainly in NPC within the first 3 hours p.i. by HBV, which subsequently led to induction of inflammatory cytokines.

HBV Envelope Proteins Induce NF-κB and Secretion of Cytokines by NPC.

To prove HBV-specific activation of PRR, we added 0.5 or 2.5 international units (I.U.) of neutralizing human anti-HBs antibodies (Hepatect, Biotest Pharma, Dreieich, Germany) to the HBV inoculum prior to infection of PHH (moi 20 virions/cell). Although equal levels of HBeAg in the inoculum indicated equal infectious doses, the addition of anti-HBs antibodies neutralized HBsAg and subsequently reduced the release of IL-6 in a dose-dependent fashion (Table 2). Equal amounts of unspecific human IgG had no influence (data not shown). Accordingly, pretreatment of HBV inocula with 0.5 or 2.5 I.U. of anti-HBs reduced NF-κB activation by 26% and 78%, respectively (Table 2). Therefore, HBV envelope proteins largely contributed to pattern recognition and induction of NF-κB-regulated cytokines in human NPC early after infection.

Table 2. Inhibition of NF-κB Activation and IL-6 Secretion After Treatment of HBV Inoculate with Anti-HBs Neutralizing Antibodies
SampleHBeAg S/COHBsAg S/NIL-6 [pg/mL]NF-κB Activity, × 104RLU
  1. Results are shown as mean of triplicates ± SEM; levels of HBeAg and HBsAg in the inoculum are shown as S/CO, signal-to-control, and S/N, signal-to-noise ratio, respectively; nd, not detected; RLU, relative light units.

PHH/HBV161.5 ± 4.839.8 ± 3.12119.7 ± 15.610.6 ± 4.51
PHH/HBV + anti-S 0.5 I.U.158.9 ± 6.314.6 ± 0.41245.7 ± 67.57.82 ± 3.51
PHH/HBV + anti-S 2.5 I.U.159.4± 9.9nd724.7 ± 25.32.29 ± 1.55

IL-6 Inhibits HBV Transcription and Replication.

Because NF-κB activation was transient and thus was abolished before virus replication started, we wondered whether the proinflammatory cytokines—without interferon induction—were able to control HBV gene expression and replication early after infection.

Addition of supernatants collected from NPC 24 hours p.i. with HBV inhibited infection of PHH. When supernatants were preincubated with IL-6-ab the effect was gone (Fig. 3A). Addition of IL-6ab to PHH prior to HBV infection resulted in a 2-fold increase of HBeAg and HBV progeny secretion (Fig. 3A,C), proving that the IL-6 released by NPC upon contact with HBV blocked HBV gene expression and replication in hepatocytes.

Figure 3.

Effect of IL-6 on HBV transcription, gene expression, and replication and on expression of hepatocyte-enriched transcription factors. Supernatant from HBV-inoculated NPC (NPCsupe) or recombinant IL-6 (rIL-6) with or without neutralizing IL-6ab were added to HBV-infected cells on days 1 and 3 p.i. To neutralize endogenous IL-6, IL-6ab (50 ng/mL in B,E; 200 ng/mL in A,C) was added prior to HBV. Cells and supernatants were harvested on day 5 p.i.; levels in untreated, HBV-infected PHH were set to 100%. (A) HBV progeny were quantified by real-time PCR of HBV DNA in cell culture medium. Mean ± SD of three measurements from two independent experiments are shown. (B) Relative gene expression levels of CRP, HNF4α, HNF1α, and HBV pgRNA determined by real-time RT-PCR and (C) dot blot analysis of HBeAg. One representative out of three experiments is shown. (D) Expression levels of HBV pgRNA after treatment with increasing amounts of rIL-6. Mean ± SD of three measurements from two independent experiments are shown. (E) 30 μg of nuclear proteins were analyzed for transcription factors HNF4α, HNF1α, and β-lamin as control. Seventy μg of cytoplasmic proteins were analyzed for HBV core and envelope proteins L, M, and S and β-actin by western blot.

For detailed analysis of IL-6 effects, we added increasing doses of recombinant IL-6 (rIL-6) to HBV-infected cells every 48 hours, after endogenous IL-6 release had ceased. C-reactive protein (CRP) expression indicated proper activation of the IL-6 signaling pathway (Fig. 3B). A dose-dependent decline of HBV pgRNA (Fig. 3D), as well as a reduction of HBeAg secretion (Fig. 3C), DNA replication intermediates (Supporting Fig. 1) and HBV progeny release (Fig. 3A) were detected. Intracellular HBV core protein was reduced by 68% and HBV envelope proteins were diminished by 90% (L), 79% (M), and 60% (S) (Fig. 3E), but restored by IL-6ab.

To examine the effects of IL-6 on the state of HBV-infected hepatocytes, we determined expression levels of the antiapoptosis genes cIAP2, Mn-SOD, and IGFBP1, which we recently found to be regulated by HBV infection (Hösel and Protzer, unpubl. results). Treatment with rIL-6 resulted in a 3.8-, 3.9-, and 24.6-fold up-regulation, respectively. In contrast, INFβ- or IFN-inducible genes were either not affected or even down-regulated (Table 3).

Table 3. Alteration of Gene Expression in HBV-Infected PHH Cultures After Treatment with Recombinant IL-6
Fold Change in Gene Expression: PHH/HBV Versus PHH/HBV+rIL-6
  1. ↑, up-regulated gene; ↓,down-regulated gene; nc, no change in gene expression detected.

1,017 ↑3.8 ↑3.9 ↑24.6 ↑ncnc7.4 ↓

Thus, IL-6 is the key factor controlling HBV early after infection and suppresses transcription of pgRNA, antigen expression, and replication. In addition, IL-6 activated a set of cellular survival factors.

MAPK Activated by IL-6 Down-regulate Hepatocyte Nuclear Factor (HNF) 1α and HNF4α.

We next searched for the mechanism of IL-6-mediated control of HBV transcription. Hepatocyte enriched transcription factors HNF1α and HNF4α determine HBV pgRNA and HBV gene expression in a concerted action in hepatocytes.16 Addition of rIL-6 decreased HNF1α and HNF4α mRNA expression 48% and 56% (Fig. 3B), and protein amounts by 55% and 58% (Fig. 3E), respectively, but not expression levels of HNF3β or γ (data not shown). Accordingly, neutralization of endogenous IL-6 released after HBV infection increased HNF1α and HNF4α gene expression by 24% and 54% (Fig. 3B). Therefore, endogenously produced IL-6 as well as added rIL-6 controlled expression of HNF1α and HNF4α, two essential transcription factors driving HBV gene expression and replication.

Because the activated MAPKs exogenous signal regulated kinase (ERK) 1/2 and jun N-terminal kinase (JNK) have been reported to control HNF4α expression, and IL-6-type cytokines may activate members of the MAPK-family, we examined whether MAPK-family members ERK1/2, p38, and/or JNK were activated after HBV infection in PHH. Levels of phosphorylated ERK and JNK, but not p38, were increased upon HBV infection (Fig. 4A). This activation was obviously mediated by IL-6 because pretreatment of PHH with neutralizing IL-6ab abolished activation of ERK and JNK during HBV infection (Fig. 4B). This was confirmed by treatment with rIL-6, which also activated ERK and JNK (Fig. 4B), but not p38 (data not shown). Upstream inhibitors PD98059 or SP600125 added prior to rIL-6 inhibited ERK phosphorylation to a large extent, whereas they blocked JNK phosphorylation (Fig. 4C). Accordingly, inhibition of JNK-activation completely and that of ERK-activation partially overcame rIL-6-mediated down-regulation of HNF4α and HNF1α (Fig. 4D) and control of HBV progeny release, whereas pp38 inhibitor showed no effect (Fig. 4E).

Figure 4.

Involvement of MAP kinase pathways in down-regulation of HNF1α, HNF4α, and HBV replication by IL-6. (A) Western blot analysis of 70 μg total proteins isolated from mock- or HBV-infected PHH 24 hours p.i. Phosphorylated and total ERK, JNK, and p38 were stained by specific antibodies. (B) Western blot analysis of pERK, ERK, pJNK, and JNK using total proteins (40 μg) isolated from HBV-infected PHH without treatment (−), treated with rIL-6 with or without IL-6ab, or from cells preincubated with IL-6ab prior to HBV infection. (C) pERK and pJNK in HBV-infected PHH untreated or treated with 50 μM PD98059 inhibiting pERK or SP600125 inhibiting pJNK for 30 minutes prior to stimulation with rIL-6. (D) 20 μg nuclear proteins from mock- or HBV-infected cells (not treated (−) or stimulated with rIL-6 or preincubated with PD98059 and SP600125 before rIL-6 stimulation) were analyzed for HNF4α, HNF1α, and lamin B as control. (E) HBV progeny DNA were quantified by real-time PCR in medium of untreated, HBV-infected cells (100%). Cells were treated with MAPK inhibitors PD98059, SP600125, or SB203580 prior to rIL-6. Mean ± SD from three measurements are given.

These results demonstrated that IL-6 activation of the MAPK ERK and JNK and subsequent down-regulation of HNF4α and HNF1α were responsible for the negative control of HBV replication by IL-6 in primary human hepatocytes.


In this study we describe that contact of primary human liver cells with HBV activated NF-κB and induced proinflammatory cytokines within 5 hours after inoculation, whereas we detected no induction of type I IFN at this early timepoint. Using primary human liver cell cultures, we identified NPC, most probably Kupffer cells, to recognize HBV patterns and to activate NF-κB prior to virus replication in hepatocytes.

We identified IL-6 to be responsible for suppression of HBV early after infection of target hepatocytes. IL-6 activated the MAPK ERK and JNK, and thus down-regulated expression of HNF4α and HNF1α, the key transcription factors regulating HBV gene expression and replication in a concerted action.16

Careful controls excluded that activation of PRR was due to contamination of our virus stocks with, e.g., co-purified lipoproteins. Neutralization of the HBV inoculum with anti-HBs antibodies, but not UV-inactivation of the virus dose-dependently prevented NF-κB activation and IL-6 secretion. This implied that HBV envelope proteins are recognized by NPC, and confirmed that this recognition does not depend on HBV replication.

Patients with acute HBV infection have elevated plasma levels of proinflammatory cytokines.17 In addition, HBeAg seems to regulate TLR-2 expression.18 A recent study by Fisicaro et al.19 confirmed that the innate immune system is able to sense HBV infection in humans, as shown by the early development of natural killer and T-cell responses, and speculate that this contributes to contain the HBV infection and to allow timely induction of adaptive immune responses. Our results help to explain these observations and demonstrate that IL-6 is responsible for the early control of HBV infection.

Notably, the activation of NF-κB and proinflammatory cytokines was transient in our experiments and was not induced by newly synthesized virus or viral antigens, confirming in human cells that HBV or its secretory antigens tolerize hepatic NPC and thus prevent continuous stimulation by newly synthesized virus.20 An interference of HBsAg expression with NF-κB signaling pathways21 may explain why the virus hardly induces any detectable innate immune responses later during infection,10 although huge amounts of viral antigens circulate. As in chimpanzees,10 we found no IFN response; IFN and IFN-inducible genes were even down-regulated.

Although transient, induction of NF-κB and proinflammatory cytokines was sufficient to affect hepatocellular gene expression. Expression of CRP was up-regulated in PHH as well as transcription of the NF-κB- and STAT3-regulated antiapoptotic genes.22 In additional experiments, we showed activation of STAT3 in HBV-infected hepatocytes (Hösel and Protzer, unpubl.). The IL-6/STAT-3 pathway is essential for liver regeneration, inhibits hepatocyte apoptosis,23 and induces the synthesis of acute phase proteins in hepatocytes, which also seem to protect the cells (reviewed24). Thus, we suggest that HBV recognition and subsequent NF-κB and IL-6 activation ensure cellular homeostasis and support survival of the infected hepatocyte because HBV obviously needs an intact cell for propagation.

It is important to note that we do not challenge the significance of other cytokines controlling HBV later during infection. In addition to type I IFN, IFNγ and TNF-α play an important role in controlling HBV infection in a noncytopathic fashion in HBV-transgenic mice25 and in HBV-infected chimpanzees.9 Both cytokines lead to the elimination of HBV RNA-containing capsids from the cytoplasm of infected hepatocytes.26, 27

As demonstrated here, IL-6 acts at the level of HBV gene expression and thereby controls replication. Levels of HBV pgRNA, HBeAg, core, and envelope proteins as well as viral progeny release were markedly reduced by rIL-6 (Fig. 3). Moreover, neutralization of endogenous IL-6 induced by HBV pattern recognition increased HBV gene expression and progeny release (Fig. 3). In accordance with our findings in PHH, administration of rIL-6 to HBV-transgenic mice suppresses hepatic HBV steady-state mRNA expression.28 However, in HepG2.2.15 hepatoma cells, IL-6 stimulated HBV transcription because STAT-3 interacting with HNF3 bound to the HBV enhancer I.29 Thus, primary human or mouse hepatocytes seem to react differently than hepatoma cells.

Galun et al.30 observed that addition of IL-6 to human liver tissue before HBV infection and transplantation into SCID mice increased numbers of HBV DNA-positive animals. This may be explained by the improvement of survival of hepatocytes by IL-6, because IL-6 does not directly interact with HBV.31

HBV transcription by host RNA-polymerase II is regulated by a number of hepatocyte-enriched transcription factors including HNF1, HNF3, or HNF4. HNF1α has been shown to be required for the efficient transcription from HBV pre-S1 promoter,32 HNF4α is necessary for the stimulation of the HBV preC/C promoter33 and binds to the viral enhancer I.34 Recently, we have shown that a concerted action of HNF4α and HNF1α, which also determines morphological and functional differentiation of hepatocytes, mediates efficient HBV transcription as a prerequisite of its replication.16

IL-6 affected HBV at the level of transcription and down-regulated expression of HNF1α and HNF4α. Neutralizing IL-6ab antibodies restored the effect (Fig. 3). This demonstrated that IL-6-mediated down-regulation of the essential transcription factors HNF1α and HNF4α is responsible for reduced HBV gene expression and pregenome transcription, and thus controls HBV replication at the level of transcription. Recently, the helioxanthin analog 8-1 has been described to suppress HBV replication by down-regulating HNF4α in virus harboring cells,35 whereas other cytokines control HBV replication usually at a posttranscriptional step.25–27, 36, 37

Finally, we tried to identify a mechanism underlying IL-6-mediated down-regulation of HNF1α and HNF4α expression. It has been reported that the family of IL-6-type cytokines besides STAT-3 activates MAPK ERK1/2, p38, and JNK (reviewed38). In this study, we show that HBV infection as well as IL-6 treatment activates MAPK JNK and ERK, but not p38 (Fig. 4A,B).

By inhibition of MAPK, we showed that upon IL-6 stimulation, activated JNK and, to a lesser extent, ERK down-regulated HNF1α and HNF4α in primary hepatocytes. Our results thus explain why activation of ERK1/2 suppressed HBV replication in HBV-transfected hepatoma cells at a transcriptional level.39 Because HNF4α also controls expression of HNF1α, it remains open whether there is a direct effect of MAPK on HNF1α, or whether this is secondary to the suppression of HNF4α.

Given the clinical benefit shown by the humanized neutralizing IL-6R monoclonal antibody tocilizumab as a treatment for Crohn disease, rheumatoid arthritis, and Castleman disease, it will be important to keep in mind that neutralization of the IL-6 pathway could represent a risk for infected patients.

Taken together, we provide strong evidence for recognition of HBV envelope proteins by a yet-unspecified PRR. This recognition occurs in or on NPC and results in control of HBV gene expression and transcription of the HBV pregenome in infected hepatocytes. For this control, IL-6 but not IFN plays a major role. By activating MAPK JNK and ERK, IL-6 controls expression of HNF1α and HNF4α, two transcription factors essential for HBV promoter activity. This represents a novel mechanism by which a cytokine controls HBV at the transcriptional level. Thus, IL-6, well known to stabilize the hepatocyte, helps to minimize early induction of immune responses and thus is advantageous for the virus. If IL-6 is blocked for therapeutic purposes, the clinical course of hepatitis B should be carefully monitored.


The authors thank Dirk Stippel for patient information, Gregor Ebert for preparing PHH, and Benjamin Yazdanpanah for perfoming the IL-1β ELISA. We thank Martin Krönke for continuous support.