Dendritic cells take up viral antigens but do not support the early steps of hepatitis B virus infection


  • Potential conflict of interest: Nothing to report.


Dendritic cells (DC) of hepatitis B virus (HBV) carriers have been reported to exhibit functional impairment. Possible explanations for this phenomenon are infection of HBV by DC or alteration of DC function by HBV. We therefore analyzed whether DC support the different steps of HBV infection and replication: uptake, deposition of the HBV genome in the nucleus, antigen expression, and progeny virus release. When HBV genomes were artificially introduced into monocyte-derived DC by adenoviral vectors, low-level expression of hepatitis B surface antigen (HBsAg) and hepatitis B e antigen (HBeAg) but no HBV replication was detected. When monocyte-derived DC were subjected to wild-type HBV or a recombinant HBV expressing Renilla luciferase under a non–liver-specific promoter, intracellular HBV DNA was detected in a low percentage of cells. However, neither nuclear cccDNA was formed nor luciferase activity was detected, indicating that either uncoating or nucleocytoplasmic transport were blocked. To verify our observation in the in vivo situation, myeloid and plasmacytoid DC were isolated from blood of high viremic HBV carriers, and analyzed by quantitative polymerase chain reaction (PCR) and electron microscopy. Although circulating DC had in vivo been exposed to more than 104 HBV virions per cell, HBV genomic DNA was hardly detected, and no nuclear cccDNA was detected at all. By using electron microscopy, subviral particles were found in endocytic vesicles, but virions were undetectable as were viral capsids in the cytoplasm. In conclusion, circulating DC may take up HBV antigens, but neither support nucleocytoplasmic transport nor replication of HBV. (HEPATOLOGY 2006;43:539–547.)

Hepatitis B virus (HBV) infection represents a major health problem worldwide. Over 350 million individuals are chronically infected with HBV and are at high risk to develop liver cirrhosis or hepatocellular carcinoma. To eliminate the virus after infection, a strong humoral and cellular immune response is required.1 Thus, control of HBV infection is associated with a multispecific and polyclonal cytotoxic T-cell response and a strong type 1 T helper cell response.2, 3 In contrast, chronically infected patients display oligoclonal T helper cell responses with weak or undetectable cytotoxic T-cell activity.4

Dendritic cells (DC) are the most important professional antigen-presenting cells. They act as key players in initiating virus-specific T-cell responses.5 This is reflected by the fact that viruses can evade immune responses by impairing DC function.6–8 The role of DC during HBV infection was intensively studied during the last years, but whether total numbers of DC are reduced during chronic infection is discussed controversary.9–11 In some studies, functional deficits of these cells were reported on contact of moDC with HBV,14 whereas they remained minor in others.12

Whether the weak or absent T-cell response described in chronic hepatitis B patients results from a defect in the DC compartment, which is caused by the virus itself, is unclear. Theoretically, numbers or functionality of DC subsets could be affected by interaction of surface receptors on DC with either viral hepatitis B surface (HBsAg) or e antigen (HBeAg) or virus particles circulating in high amounts in patients' blood.

Alternatively, HBV might infect DC, and expression of viral antigens or replication of the viral genome could interfere with signaling pathways in DC. Successful infection of DC leading to gene expression from viral genomes and finally to virus replication, however, would require uptake of virions into an appropriate compartment, uncoating of viral capsids by fusion of viral with cellular membranes, transport of viral capsids to the nucleus, release of the viral genome into the nucleus, and filling up the viral genome to a covalently closed circular (ccc) DNA form serving as transcription template mandatory for HBV gene expression and productive infection.15

Detection of HBV-DNA in DC lysates by polymerase chain reaction (PCR) has been proposed to indicate HBV infection of DC.9, 10, 12, 13 However, this conclusion can be questioned, because the detection of HBV-DNA and even HBV-RNA in peripheral blood mononuclear cells (PBMC) can be explained by adsorbed virus and does not necessarily correlate with detection of cccDNA and thus with infection.16 Therefore, whether HBV can infect DC, induce expression of viral antigens, and initiate replication of the viral genome remains unknown.

We investigated whether HBV can infect moDC in cell culture and whether viral antigens are expressed in DC. To be able to clearly distinguish between adsorption or unspecific uptake of HBV and infection of DC by HBV, we took advantage of recently developed HBV-based vectors.17 Recombinant HBV allow us to unambiguously assay the early steps of HBV infection18 because of their ability to express luciferase after successful uptake and nuclear transport.19 In addition, we investigated the formation of nuclear HBV cccDNA after infection with wild-type HBV. To determine whether HBV antigens are expressed in DC, we used an artificial transfer of HBV genomes into DC by adenoviral vectors20 and studied HBV gene expression. To verify our findings in the in vivo situation, we isolated mDC and pDC from highly viremic chronic carriers and tested them for HBV infection.


HBV, hepatitis B virus; DC, dendritic cells; moDC, monocyte-derived dendritic cells; mDC, myeloid dendritic cells; pDC, plasmacytoid dendritic cells; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; ccc DNA, covalently closed circular DNA; PCR, polymerase chain reaction; PBMC, peripheral blood mononuclear cells; PBS, phosphate-buffered saline; MAb, monoclonal antibody; HNF, hepatocyte nuclear factor; GFP, green fluorescent protein; e.u., expression units; wt, wild-type; rHBV, recombinant HBV; CMV, cytomegalovirus; rc DNA, relaxed circular DNA.

Materials and Methods

Preparation of Dendritic Cells and Hepatocytes.

PBMC were obtained from standard buffy coat preparations of healthy blood donors by Ficoll density gradient centrifugation (Pancoll, Pan Biotech, Aidenbach, Germany). Monocytes were isolated by adherence to plastic surfaces by incubation of PBMC at 5 × 106/mL in RPMI 1640 medium with 10% fetal calf serum (Biochrom, Berlin, Germany) for 2 hours. Subsequently, non-adherent cells were removed by washing with phosphate-buffered saline (PBS). Alternatively, monocytes were isolated by immunoselection with CD14 microbeads (Miltenyi, Bergisch Gladbach, Germany). Cells obtained were cultured in complete RPMI medium supplemented with 10 ng/mL interleukin 4 and 10 ng/mL granulocyte-macrophage colony-stimulating factor (R&D Systems, Wiesbaden, Germany) for 3 to 5 days to allow differentiation into immature DC. For further maturation, DC were stimulated by addition of 1 μg/mL lipopolysaccharide (Escherichia coli O26:B6, Sigma Aldrich, Germany), 50 ng/mL tumor necrosis factor alpha (R&D Systems), and 50 ng/mL prostaglandin E2 (Sigma Aldrich) for 24 to 48 hours. MoDCs were 70% to 98% pure; maturation state was determined by flow cytometry using mouse monoclonal antibodies (MAb) against HLA-DR, CD80, CD86 (BD, Pharmingen, San Diego, CA), and CD83 (Immunotech, Marseille, France).

Isolation of primary cells was approved by the local Ethics Committee. Circulating pDC and mDC were gained from PBMC of patients with chronic HBV infection and 1 healthy volunteer after informed consent. DC subsets were magnetically sorted using BDCA-4 and BDCA-1 cell isolation kits (Miltenyi) following the manufacturer's instructions. After 2 rounds of positive selection, purity was 76% to 96% as determined by flow cytometry MAbs against BDCA-2 and CD123 (Miltenyi) for pDC and streptavidin-PE to stain biotinylated MAb from the isolation procedure for mDC (BD Pharmingen). Contaminating cells were predominantly erythrocytes.

Primary human hepatocytes used as control were isolated from liver tissue samples from patients undergoing partial hepatectomy after informed consent was obtained. Cells were isolated, seeded and maintained as described.17, 21

Western Blot Analysis.

Cells were harvested in lysis buffer [200 mmol/L Tris-Cl pH 8.8; 5 mmol/L EDTA; 3% sodium dodecyl sulfate (SDS); 10% sucrose; 1.7% β-mercaptoethanol]. Proteins contained were separated using 10% SDS-PAGE, transferred to a nitrocellulose membrane for Western blot analysis, and detected using rabbit antisera against β-actin (42 kd) (Sigma, St. Louis, MO), hepatocyte nuclear factor (HNF) 1 (HNF1α 92 kd, HNF1β 61 kd) and 4 alpha (HNF4α 56 kd) (Santa Cruz Biotechnology, Santa Cruz, CA) and the “Enhanced Chemiluminescence detection” kit (Amersham Biosciences, Buckinghamshire, UK).

Adenoviral Vectors.

Adenoviral vector AdG-HBV1.3 contains a 1.3-fold over length HBV genome that is able to establish HBV replication in primary human hepatocytes and a green fluorescent protein (GFP) expression cassette to control transfer efficiency. Vectors were produced, grown, and purified as described previously.20, 22 The relative number of GFP-expressing cells 16 to 24 hours after infection determined the titer of infectious adenoviral vector particles as expression units (e.u.) per milliliter.

Hepatitis B Viruses.

Wild-type HBV (wtHBV) was concentrated from the medium of HepG2.2.15 cells cultivated in Williams E Medium with 5% fetal calf serum using centrifugal filter devices (Centricon Plus-70, Biomax 100.000, Millipore Corp., Bedford, MA), and stored in 10% glycerol at −80°C until further use.

HBV-based vectors (rHBV) were produced as described.17, 19, 21 In rHBV-rLuc, an HBV genome was packaged, in which all viral open reading frames were knocked out, and the gene encoding for the viral small envelope protein was replaced by a cDNA encoding renilla luciferase. In rHBV-cytomegalovirus (CMV)-rLuc, in addition, the HBV preS2/S-promoter has been replaced by a CMV immediate early promoter/enhancer. Titers of wtHBV and rHBV, measured as enveloped, DNA-containing viral particles (vp), were determined by sedimenting viral particles in a CsCl density-gradient and quantifying enveloped viral particles by dot blot analysis relative to an HBV standard.17

Infection of Cells.

Cells were incubated with wt HBV or rHBV-rLuc at moi 100 vp per cell in the presence of 5 % polyethylene glycol (PEG 6000) and—if indicated—neutralizing anti-HBs antibodies (1 IU/106 cells; Hepatect, Biotest Pharma GmbH, Dreieich, Germany). Alternatively, cells were incubated with indicated amounts of AdG-HBV1.3. After overnight infection, cells were washed 3 times with PBS before fresh medium was added.

Cell culture medium was collected from day 1 to day 4 postinfection, and HBsAg and HBeAg secretion were determined using commercial immuno assays (Axsym, HBeAg 2.0, HBsAg V2, Abbott Laboratories, Wiesbaden, Germany).

Detection of Luciferase Expression.

To detect luciferase expression, 106 cells were lysed on day 4 postinfection, and luciferase activity was determined using a commercially available assay (Promega Corp., Madison, WI) according to the manufacturer's instructions. Relative light units were measured using a luminometer (Berthold, Pforzheim, Germany) and normalized to the total protein content of the respective sample (Protein Assay, Bio-Rad Lab. Inc., Hercules, CA). Relative light units determined in mock infected controls were defined as 1, and remaining values were adapted proportionally.

Quantitative HBV-PCR.

Total DNA was purified from lysates of moDC at day 4 postinfection or from mDC/pDC directly after isolation using the DNA MinElute Kit (Qiagen, Hilden, Germany) and eluted with 25 μL H2O. HBV-DNA was quantified by real-time PCR relative to an external plasmid DNA standard on a Light Cycler instrument using LightCycler FastStart DNA Masterplus SYBR Green I (Roche, Mannheim, Germany) and primers HBV1745fw (5′-GTTGCCCGTTTGTCCTCTAATTC-3′) and HBV1844rev (5′-GGAGGGATACATAGAGGTTCCTTGA-3′) hybridizing in the HBV surface gene to detect HBV relaxed circular (rc) genomes (100-bp fragment). To detect HBV ccc DNA, we designed primers HBVccc2760fw (5′-GACTCTCTCGTCCCCTTCTC-3′) and HBVccc156rev (5′-ATGGTGAGGTGAACAATGCT-3′) (580-bp fragment), which span the gap and the nick in the rc form of the HBV genome as principally described earlier.16 Using optimized PCR conditions on a Light Cycler instrument (95°C for 5′, 45 cycles at 95°C for 15″, 60°C for 4″, 72°C for 25″, and detection at 88°C for 2″ after each cycle), we determined the specificity to amplify ccc DNA over rc DNA to be 104 to1.

Electron Microscopy.

After isolation, mDC and pDC were pelleted in a 1.5-mL microfuge tube at 1,000g. Supernatant was removed, and the cell pellet was fixed in 2.5% glutaraldehyde for 12 hours at 4°C, washed 3-fold in PBS, and incubated in 2% OsO4 and 1% K3Fe (CN)6 for 1 hour. After 3 further PBS washes, pellets were dehydrated and infiltrated in a 1:1 mixture of propylene oxide (Polybed 812 and epoxy resin; Polysciences, Warrington, PA) and embedded in resin. Ultrathin (50-nm) sections were collected on 200-mesh copper grids and stained with 6% uranyl acetate in bidestilled water for 10 minutes followed by 1% lead citrate for 15 minutes. Sections were viewed using a Zeiss 902 transmission electron microscope at 50 kV.


HBV Antigen Expression and Replication in moDC.

Adenoviral vector AdG-HBV1.3, which contains a linear 1.3-fold HBV genome and a GFP expression cassette, efficiently initiates HBV antigen expression and replication in hepatocytes.20 We used AdG-HBV1.3 to study whether immature and mature moDC express HBV antigens under HBV-specific promoters and whether they support HBV genome replication.

Using a moi ≥ 3, more than 90% of all cells in primary human hepatocyte cultures were transduced as monitored by GFP expression. Human hepatocytes secreted high levels of HBeAg and HBsAg after transduction (Fig. 1A). Cells not transduced by AdG-HBV1.3 most probably were nonparenchymal liver cells, which make up for approximately 10% of all cells in primary human hepatocyte cultures.21, 23

Figure 1.

Transduction of primary human hepatocytes (PHH) and moDC by adenoviral vectors carrying a replication competent 1.3-fold HBV genome. (A) PHH were transduced using AdG-HBV1.3 at moi 1, 3, or 10 e.u./cell. A representative experiment is shown. (B) Immature DC and (C) mature DC were transduced using AdG-HBV1.3 at moi 10, 30, 100, 300, 1,000, or 3,000 e.u./cell. Efficiency of the adenoviral gene transfer was controlled by flow cytometry counting GFP-positive cells. HBsAg and HBeAg secreted at day 4 post-transduction were measured in the cell culture medium. Median values and standard deviations determined in moDC obtained from 3 different donors are given. moDC, monocyte-derived dendritic cells; HBV, hepatitis B virus; e.u., expression units; GFP, green fluorescent protein; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; S/N, signal to noise; S/Co signal to control.

Maturation status of moDC was determined by flow cytometry revealing significant upregulation of HLA-DR, CD80, CD83, and CD86 in matured DC (Table 1). Immature and mature moDC were infected with increasing mois of AdG-HBV1.3 to reach transduction efficiency similar to that of human hepatocytes (Fig. 1B-C, left panel). Accordingly, 1,000 e.u. per cell were required to transduce 80% of moDC. The amount of HBeAg and HBsAg secreted correlated with the amount of transduced cells (Fig. 1). Compared with hepatocytes transduced to 80%, immature moDC secreted 4.3-fold less, and mature moDC secreted 1.6-fold less HBsAg per cell. HBeAg, which is expressed under a more liver-specific promoter, showed 17-fold lower expression in immature moDC and 24.5-fold lower expression in mature moDC (Fig. 1 middle and right panel). To analyze whether HNF1 and HNF4 necessary for efficient transcription from HBV preS1 and precore/core promoters, respectively, are expressed in DC, we performed Western blot analysis. Whereas HNF1 α and β as well as HNF 4 α were easily detected in lysates from human hepatoma HepG2 cells used as positive control, we could detect them neither in immature nor in mature moDC (Fig. 2).

Table 1. Flow Cytometry Analysis of Monocyte-Derived DC (moDC) Before (Immature) and After (Mature) In Vitro Differentiation: Percentage of Cells Staining Positive for Surface Antigens HLA-DR, CD80, CD83, and CD86
 Immature moDCMature moDC
HLA-DR21.6% ± 5.2%92.2% ± 3.4%
CD8032.5% ± 9.9%90.5% ± 8.4%
CD833.1% ± 1.9%88.0% ± 4.0%
CD864.4% ± 4.7%96.6% ± 2.6%
Figure 2.

Western blot analysis of transcription factors required for HBV gene expression. Cellular lysates of human hepatoma HepG2 cells as well as mature and immature moDC isolated from 2 different donors were subjected to Western blot analysis using antibodies against hepatocyte nuclear factor (HNF) 1α and β (92 and 61 kd), HNF 4 (56 kd), and β actin (42 kd) to control equal loading. HBV, hepatitis B virus; moDC, monocyte-derived dendritic cells.

Culture media of the cells were subject to CsCl density-gradient centrifugation, and HBV progeny secretion into the culture medium was assayed by dot blot analysis using HBV- and adenovirus-specific 32P-labeled probes, which allow detecting 8 pg of the respective DNA. In contrast to human hepatocytes, no HBV particles were detected in the cell culture medium of DC (data not shown).

From these experiments, we concluded that immature and mature DC because of the lack of transcription factors are only capable of promoting HBV gene expression from HBV promoters at low levels. Missing progeny HBV indicated that HBV does not replicate in DC.

Infection of moDC With HBV.

To determine whether moDC support the early steps of HBV infection, we used HBV-based vectors rHBV-rLuc and rHBV-CMV-rLuc. These recombinant HBV (rHBV) are replication deficient, but after entry and nuclear transport of rHBV genomes, they express renilla luciferase under control of the HBV preS2/S-promoter or a CMV-promoter, respectively19 (Fig. 3A).

Figure 3.

Infection of hepatocytes and monocyte-derived DC with recombinant HBV expressing Renilla luciferase. Primary human hepatocytes (PHH) or monocyte-derived DC were infected with recombinant HBV (rHBV) transferring a renilla luciferase gene under control of the HBV preS2/S-promoter (rHBV-rLuc) or a CMV promoter (rHBV-CMV-rLuc). (A) Schematic depiction of the constructs pCH-rLuc and pCH-CMV-rLuc used for generation of recombinant HBV (rHBV) rHBV-rLuc and rHBV-CMV-rLuc, respectively. In both constructs, all HBV open reading frames are knocked out (C, P, L, M, X−) In rHBV-rLuc, luciferase is expressed under control of HBV preS2/S-promoter (rHBV-rLuc), in rHBV-CMV-rLuc under control of a CMV promoter. B-F) Infection with indicated rHBV at moi 100 vp/cell in the presence and absence of neutralizing antibodies (nAb). 106 cells were lysed on day 4 p.i. and assayed for luciferase expression. Infection of (B) primary human hepatocytes, (C, E) immature, or (D, F) mature DC. In (C) and (D), rHBV-rLuc was used for infection, in (E) and (F) rHBV-CMV-rLuc. A, B, and C represent cells obtained from three different donors. Relative light units (median and standard deviation) normalized to total protein content of the lysates measured in 3 independent infection experiments are given. For DC, infection experiments were performed in triplicate with cells from each donor.

After infection of primary human hepatocytes with rHBV-rLuc and rHBV-CMV-rLuc, luciferase expression was 200-fold above background although only approximately 5% of the hepatocytes were infected (moi 100 vp/cell). If neutralizing antibodies were present during infection, luciferase expression was abrogated (Fig. 3B), showing that these antibodies completely blocked specific uptake of HBV.

By adenoviral vector titration experiments in DC and by titration experiments in hepatocytes, we determined that 1 to 2.5 positive cells per 104 cells were enough to detect luciferase expression (data not shown).

MoDC were isolated from three different donors (A, B, C) and matured in vitro if indicated. After incubation of moDC with rHBV-rLuc (Fig. 3C-D) or rHBV-CMV-rLuc (Fig. 3E-F) at moi 100 vp/cell, cells were lysed and luciferase assays were performed. Neither in immature DC (Fig. 3C,E) nor in mature DC (Fig. 3D,F), luciferase expression was detected, irrespective of whether neutralizing antibodies were present. Using cells isolated from six other donors or independent recombinant virus stocks, identical results were obtained (data not shown).

To exclude that wild-type HBV behaves differently, we incubated moDC with wild-type HBV at moi 100 vp/cell for 18 hours, extensively washed and trypsinized the cells to remove attached virus, further cultivated them for 24 hours, and then isolated DNA from cellular lysates. By specific real-time PCRs (see Materials and Methods), we assayed for the presence of HBV rcDNA, representing intracellular HBV genomes, and HBV cccDNA, representing the transcription template established after nuclear entry (Fig. 4). In 3 independent infection experiments, between 0 and 20 copies of HBV rcDNA per 103 cells were detected in immature and mature moDC incubated with wtHBV. However, no HBV cccDNA was detected with a detection limit of 1 copy in 103 cells. Neutralizing antibodies present during incubation with wtHBV had no major influence on data obtained with immature moDC (data not shown), whereas they reduced uptake of wtHBV into mature moDC (Fig. 4).

Figure 4.

PCR analysis of monocyte-derived DC subjected to wild-type HBV. In vitro matured monocyte-derived DC (moDC) were incubated with wild-type HBV (+HBV) at moi 100 vp/cell. Uninfected DC (mock) and primary human hepatocytes (+) served as controls. Specific primer pairs were used to detect HBV relaxed circular (rc) DNA genomes and HBV covalently closed circular (ccc) DNA formed in the nucleus of infected cells. M = DNA size marker.

These experiments showed uptake of HBV particles into in vitro differentiated DC, but no establishment of a nuclear transcription template indicating that either uncoating or nucleocytol plasmic transport of viral capsids was blocked. We therefore concluded that neither immature nor mature moDC support the early steps of HBV infection.

Analysis of Blood-Derived DC From High Viremic Chronic Hepatitis B Patients.

DC circulating in the blood of HBeAg-positive HBV carriers are permanently subjected to extremely high amounts of HBV (Table 2, “HBV input”). We therefore isolated mDC and pDC from the blood of four high viremic HBsAg-/HBeAg-positive chronic hepatitis B patients. All patients had mild to moderate hepatitis with alanine aminotransferase levels reaching from the upper limit of normal to 6-fold elevation. None of the patients had advanced cirrhosis. By real-time PCR, we assayed isolated DC for the presence of HBV genomic DNA (rcDNA) and of HBV cccDNA, indicating nuclear transport of HBV genomes and thus infection of respective cells. For ccc DNA analysis, DNA from hepatocytes in vitro infected with HBV at moi 100 vp/cell served as positive and DNA isolated from patients' own sera and a serum from an HBV-infected chimpanzee (1.25 × 10e8 HBV copies/mL) served as negative controls (Table 2, Fig. 5A). The detection limit of our assay was 0.2 HBV rcDNA copies, and 1 HBV cccDNA copy per 103 cells.

Table 2. Quantitative Analysis of HBV DNA Forms in DNA Extracted From Blood-Derived DC and Primary Human Hepatocytes
SampleHBV Viremia per mLCellsHBV Input per CellHBV Genomes per 103 CellscccDNA Copies per 103 Cells
  1. Abbreviations: mDC, myeloid dendritic cells; pDC, plasmacytoid dendritic cells; −, negative PCR; PHH, primary human hepatocytes; cccDNA, covalently closed circular DNA

18.2 × 10e7mDC5,46011.6
27.9 × 10e7mDC5,260
32.6 × 10e9mDC173,300
44.8 × 10e8mDC32,000
Figure 5.

PCR and electron microscopy analysis of blood-derived DC from high viremic HBV carriers. (A) Myeloid (mDC) and plasmacytoid (pDC) dendritic cells were isolated from an HBsAg-/HBeAg-positive HBV carrier with a viremia of 8.2 × 107 HBV per milliliter. DNA extracted from 38,000 to 50,000 cells was subjected to real-time PCR analysis to detect total HBV genomes (rcDNA) and HBV cccDNA. DNA extracted from patient serum, from serum of an HBV-positive chimpanzee and from primary human hepatocytes (PHH) in vitro infected with HBV served as controls. M = DNA size marker. (B) Transmission electron microscopy of ultrathin sections (50 nm) of freshly isolated myeloid DC.

In mDC and pDC from only 1 patient each, HBV genomic rcDNA was detected, and in none of the samples analyzed was HBV ccc DNA detected (Table 2, Fig. 5A). From this we concluded that circulating DC in HBV carriers do rarely contain HBV DNA at all and do not support establishment of the viral genome in the nucleus. Thus, they are not infected with the virus to a significant proportion.

To analyze whether HBV viral or subviral particles were taken up by circulating DC in patient's blood, we performed electron microscopy of mDC and pDC. Subviral particles were detected in endocytic vesicles in mDC (Fig. 5B), but neither complete virions nor free capsids were detected in the cytoplasm or at the nuclear membrane. This indicates that subviral particles circulating in high amounts in the blood are taken up by DC and thus might interact with surface receptors on these cells.


Because DC dysfunction in chronic hepatitis B patients might be caused by HBV infection of DC, we intended to discriminate the multiple steps of HBV infection in DC and analyzed the replication cycle in detail to understand at which level HBV may affect DC functionality. Although DC expressed HBsAg and at very low levels also HBeAg under control of HBV viral promoter/enhancer elements, we were not able to detect HBV infection of DC. This finding was confirmed by cccDNA PCR analysis and electron microscopy of circulating mDC and pDC from high viremic HBV carriers ex vivo.

To test whether HBV antigens are expressed in DC in the context of an HBV genome, moDC were artificially transduced with HBV genomes using adenoviral vectors.20 Eighty-four percent of the immature DC and 75% of the mature DC received an HBV genome as determined by co-expression of a GFP marker gene. Adenoviral transfer of HBV genomes into DC resulted in low-level expression of HBV antigens, determined by secretion of HBsAg and HBeAg (Fig. 2). HBsAg levels per cell were 4.3- and 1.6-fold, and HBeAg levels were 17- and 24.5-fold lower in immature and mature DC, respectively, than in hepatocytes. Although HBV antigens were expressed in DC at low amounts, no production of HBV particles was detected in our assays.

Because transcription from the precore-/core-promoter is highly restricted to liver cells,15, 24 missing nuclear factors in DC might explain the low expression level of HBeAg and lacking progeny virus release. In Western blot analysis of DC lysates, we could detect expression of neither HNF 1 nor HNF 4α, which have been reported to be essential for transcription of preS1 transcripts and HBV pregenomes, respectively.24–26 In contrast, the preS2-/S-promoter only shows limited liver specificity27 and is constitutively active in a wide range of cells.28, 29 The differences in HBsAg expression might therefore either result from differences in promoter activity or from amplification of HBV covalently closed circular DNA (cccDNA) serving as an additional transcription template in hepatocytes,20 but not in DC.

To test whether DC are infected by HBV, we used recombinant HBV, which express luciferase either under control of the HBV preS2-/S-promoter or of a CMV promoter on establishment of their genome in the nucleus of infected cells. Renilla luciferase seemed to be an optimal marker because luminescence of its substrate is detected with high sensitivity over a broad linear range, cells display a minimal luminescence, and the input virus does not give a background signal.30 These recombinant HBV used allowed sensitive and specific detection of HBV infection of permissive cells, and infection could be completely blocked by neutralizing antibodies. However, no luciferase expression was detected in DC.

We estimated that the detection limit of our luciferase assay is approximately 0.1 positive cells out of 103 DC. Therefore, sensitivity and specificity of this assay is even higher than that of PCR assays preferentially amplifying cccDNA. Although this allows very sensitive detection of HBV infection, infection of DC at a level of ≤1:10−4 cannot be excluded. However, infection at such a low frequency is unlikely to cause alteration of the virus-specific T-cell response.

In previous studies, possible HBV infection of DC was mostly studied by PCR methods detecting HBV rcDNA genomes9, 10, 12 or by PCR in situ hybridization.31 These methods, however, hardly distinguish input viruses from newly synthesized particles. Because nuclear entry of viral DNA is a prerequisite for successful HBV infection of a cell and results in the formation of HBV ccc DNA,15, 32 Köck et al.16 developed a PCR assay that amplifies cccDNA more efficiently than relaxed circular DNA present in HBV particles, and tested whether PBMC can be infected by HBV. They found no evidence that PMBC can be infected with HBV and concluded that adsorbed virus can explain detection of HBV DNA or HBV RNA by PCR.33, 34 We optimized detection of HBV cccDNA by PCR on a Light Cyler™ instrument (see Materials and Methods). When this assay was used, nuclear HBV cccDNA was detected neither in lysates of moDC infected with recombinant HBV nor with wild-type HBV. This confirmed that moDC were not infected with HBV, although the presence of low amounts of HBV genomes indicated that an uptake of virions was possible. The lack of nuclear cccDNA formation, and the lack of gene expression from HBV genomes taken up, however, points to an unspecific uptake, for example, by phagocytosis.

To verify our findings in the physiological situation in HBV carriers, we analyzed circulating mDC and pDC isolated from four HBV-infected patients with high viremia ex vivo for HBV infection. With a detection limit of 1 cccDNA copy per 103 cells, we did not detect HBV cccDNA, although each cell had been exposed to up to 260,000 virions per cell in the blood before isolation (see Table 2). In agreement with molecular analyses, we were able to detect neither HBV virions nor free capsids, but endocytosed subviral particles in circulating DC from infected patients by electron microscopy. However, because DC are very heterogenous in vivo, and we were not able to study, for example, liver-derived DC, we cannot exclude that there are DC population behaving differently from those investigated.

So far, no correlation was observed between the amount of HBV-DNA present in mDC or pDC,10 or the presence of viral antigens in moDC and the allostimulatory capacity of DC.12 Thus, to which extent and how the function of DC is altered during HBV infection remains unclear.

Taking into consideration our own and data reported in the literature,35 we hypothesize that contact with circulating viral antigens rather than infection with HBV may alter the function of DC in HBV carriers. In particular, subviral particles, which circulate during HBV infection in high amounts and can be detected as HBsAg, may alter the antiviral cytotoxic T cell response.36, 37 Alternatively, an altered microenvironment in lymphoid organs during HBV infection or the microenvironment in the liver could be responsible for the impaired T-cell activation observed in chronic hepatitis B patients.4

From the results obtained in this study using in vitro matured as well as ex vivo blood-derived DC, we can exclude that HBV infects DC at a frequency sufficient to explain a functional impairment of virus-specific T cells in chronic hepatitis B patients. The fundamental question, however, remains to what extent HBV or HBV antigens interact with dendritic cells, and whether this plays any role in the pathogenesis of HBV-related disease.


The authors thank the Medical Faculty of the University of Cologne for establishing the Junior Research Group “Molecular Infectiology” and Martin Kroenke for critical discussion and continuous support. We are grateful to Eva Varus for help with electron miscroscopy and Roland Kaiser for serological analyses.