Viral DNA-Dependent Induction of Innate Immune Response to Hepatitis B Virus in Immortalized Mouse Hepatocytes

ABSTRACT

Hepatitis B virus (HBV) infects hundreds of millions of people worldwide and causes acute and chronic hepatitis, cirrhosis, and hepatocellular carcinoma. HBV is an enveloped virus with a relaxed circular (RC) DNA genome. In the nuclei of infected human hepatocytes, conversion of RC DNA from the incoming virion or cytoplasmic mature nucleocapsid (NC) to the covalently closed circular (CCC) DNA, which serves as the template for producing all viral transcripts, is essential to establish and sustain viral replication. A prerequisite for CCC DNA formation is the uncoating (disassembly) of NCs to expose their RC DNA content for conversion to CCC DNA. We report here that in an immortalized mouse hepatocyte cell line, AML12HBV10, in which RC DNA exposure is enhanced, the exposed viral DNA could trigger an innate immune response that was able to modulate viral gene expression and replication. When viral gene expression and replication were low, the innate response initially stimulated these processes but subsequently acted to shut off viral gene expression and replication after they reached peak levels. Inhibition of viral DNA synthesis or cellular DNA sensing and innate immune signaling diminished the innate response. These results indicate that HBV DNA, when exposed in the host cell cytoplasm, can function to trigger an innate immune response that, in turn, modulates viral gene expression and replication.

IMPORTANCE Chronic infection by hepatitis B virus (HBV) afflicts hundreds of millions worldwide and is sustained by the episomal covalently closed circular (CCC) DNA in the nuclei of infected hepatocytes. Release of viral genomic DNA from cytoplasmic nucleocapsids (NCs) (NC disassembly or uncoating) is a prerequisite for its conversion to CCC DNA, which can also potentially expose the viral DNA to host DNA sensors and trigger an innate immune response. We have found that in an immortalized mouse hepatocyte cell line in which efficient CCC DNA formation was associated with enhanced exposure of nucleocapsid-associated DNA, the exposed viral DNA indeed triggered host cytoplasmic DNA sensing and an innate immune response that was able to modulate HBV gene expression and replication. Thus, HBV can, under select conditions, be recognized by the host innate immune response through exposed viral DNA, which may be exploited therapeutically to clear viral persistence.

INTRODUCTION

Hepatitis B virus (HBV) has infected approximately 2 billion people worldwide, with 350 million of them becoming chronically infected (1). Annually, 1 million fatalities are attributed to acute and chronic hepatitis, cirrhosis, and hepatocellular carcinoma (HCC) caused by HBV. HBV is a small, enveloped DNA virus that contains a 3.2-kb, partially double-stranded (DS), relaxed circular (RC) DNA genome and replicates via an RNA intermediate, the pregenomic RNA (pgRNA). Upon infection, HBV RC DNA is converted to covalently closed circular (CCC) DNA in the nuclei of infected human hepatocytes, which serves as the transcriptional template for all viral RNAs, including pgRNA, and thus is essential for establishing and maintaining viral infection (2, 3). Viral DNA replication starts with the assembly of a replication-competent nucleocapsid (NC) in the host cell cytoplasm by the viral core protein (HBc), incorporating pgRNA and the viral reverse transcriptase (RT). Subsequently, RT converts pgRNA first to a single-stranded (SS) (minus-strand) DNA and then to the DS RC DNA (4, 5). RC DNA-containing NCs can be enveloped and secreted extracellularly as complete virions. Alternatively, RC DNA synthesized de novo in cytoplasmic NCs can be recycled back to the nucleus via an intracellular amplification pathway to amplify and replenish the CCC DNA pool in the nucleus (6–8).

The release of HBV DNA from the protective NCs (NC disassembly or uncoating), a prerequisite for CCC DNA formation (9), may potentially expose the viral DNA to host DNA-sensing mechanisms. Indeed, foreign or mislocalized cellular DNA represents one of the major pathogen-associated molecular patterns (PAMPs) that are recognized by their corresponding cellular receptors or sensors, the pattern recognition receptors (PRRs) (10, 11). Thus, many DNA viruses are detected by DNA sensors in the host cell cytoplasm (cyclic GMP-AMP synthase [cGAS] and others) (12–14), endosomes (Toll-like receptor 9 [TLR9] in select immune cells) (10, 11), and even in the nucleus (gamma interferon [IFN-γ]-inducible protein 16 [IFI16]) (15, 16). DNA sensors are important innate immune factors in triggering early antiviral defenses, such as type I IFN production, and in regulating the adaptive immune response to clear viral infections. A major signal transducer in cytosolic DNA sensing is the endoplasmic reticulum (ER)-associated protein stimulator of interferon genes (STING), which acts downstream of DNA sensors like cGAS, although STING-independent pathways of DNA sensing have also been reported (11, 14, 17). Signaling through STING and downstream effectors, such as interferon-regulatory factor 3 (IRF3), leads to production of type I IFN, which has strong antiviral activities against a variety of viruses through the induction of a large number of interferon-stimulated genes (ISGs) (11, 18).

If and how HBV is detected by the innate immune system has remained an important yet unresolved issue. In contrast to many other viruses, HBV is generally thought to be a “stealth” virus because it does not induce a type I IFN response during natural or experimental infections (19–24). This is thought to be related to the unique replication cycle of the virus. In particular, the viral DNA is synthesized only after pgRNA packaging into the protective NCs, sequestered away from cytosolic DNA sensors. However, some recent reports have suggested that HBV may be able to trigger the innate immune response under certain conditions in infected hepatocytes or nonparenchymal liver cells, such as Kupffer cells, which are not productively infected (25–28). If innate immune sensing indeed occurs during HBV infection, the putative viral trigger (i.e., the PAMP) or host detector (i.e., the PRR) remains to be defined.

On the other hand, it is clear that a strong adaptive immune response against HBV infection is eventually triggered after an incubation period of weeks to months, when viral replication is actively spreading throughout the liver. This adaptive response is able to clear the virus in the vast majority (90 to 95%) of immunocompetent adults (29). In addition to cytotoxic effector lymphocytes that directly kill infected cells, soluble immune effectors (such as type I and type III IFN, tumor necrosis factor alpha [TNF-α], and interleukin 6 [IL-6]) elicit strong, noncytolytic antiviral effects targeting multiple stages of the HBV replication cycle, including transcriptional and posttranscriptional suppression of viral RNA expression, blocking of NC assembly, and destabilization of preformed NCs (24, 25, 30–37). Intriguingly, as demonstrated in an HBV transgenic-mouse model, IFN can also stimulate, rather than suppress, HBV gene expression and replication when viral replication levels are low (38), suggesting that HBV may have evolved to coopt the host antiviral response to enhance its own replication.

We recently developed an immortalized mouse hepatocyte line, AML12HBV10, that supports high levels of HBV replication in a tetracycline (Tet)-regulated manner and is highly responsive to HBV-suppressive effects of certain antiviral cytokines, like IFN (39). Furthermore, we have found that AML12HBV10 cells, in contrast to normal mouse hepatocytes in vivo, which are unable to support HBV CCC DNA formation despite accumulating abundant RC DNA (40), could support efficient HBV CCC DNA formation, which was likely facilitated by the rapid and efficient uncoating of the viral NCs to expose the genomic DNA for CCC DNA conversion in these cells (41). Here, we report that the increased exposure of RC DNA in AML12HBV10 cells led to the triggering of an innate immune response that was dependent on viral DNA and host DNA sensing and signaling mechanisms and was able to modulate viral gene expression and replication.

MATERIALS AND METHODS

Cell cultures.

The AML12HBV10 cell line (39), derived from an immortalized murine hepatocyte line (AML12) (42), and the HepAD38 cell line (43), derived from a human hepatoblastoma cell line (HepG2), were maintained in Dulbecco modified Eagle–F-12 medium supplemented with 10% fetal bovine serum (FBS), 50 μg/ml of penicillin-streptomycin, 400 μg/ml G418 (GIBCO), and 5 μg/ml of Tet until HBV replication induction. Both cell lines are induced to replicate HBV upon removal of Tet from the culture medium.

Time course of HBV gene expression and replication.

Upon induction, AML12HBV10 cells were harvested every 3 days until day 18 days post-Tet removal. To test the effects of viral DNA synthesis or the TBK1 pathway on viral gene expression and replication over time, cells were treated with entecavir (ETV) (Moravek Biochemicals; 100 nM), AT-Br (a bromide-substituted phenylpropenamide derivative, synthesized by following the published procedure [44]; 20 μM), or BX795 (TBK1 inhibitor; Invivogen; 1 μM) during the entire time course.

Isolation of viral DNA.

Viral core DNA (NC-associated viral DNA) was isolated from induced HepAD38 or AML12HBV10 cells as previously described (8, 39, 45). In brief, cells were lysed in NP-40 lysis buffer (50 mM Tris-HCl [pH 8.0], 1 mM EDTA, 1% NP-40) containing a protease inhibitor cocktail (Roche) and briefly centrifuged at 14,000 rpm to remove the nuclei and cell debris. The resulting cytoplasmic lysate was incubated with micrococcal nuclease (MNase) (Roche; 150 units/ml) and CaCl₂ (5 mM) at 37°C for 90 min to degrade the nucleic acids outside NCs. The MNase was then inactivated by addition of 10 mM EDTA. Subsequently, the NCs were precipitated with polyethylene glycol (PEG) and disrupted by 0.5% sodium dodecyl sulfate (SDS). The NCs were then digested with 0.6 mg/ml of proteinase K (PK) at 37°C for 1 h. The viral DNAs were then recovered by phenol-chloroform extraction and ethanol precipitation. The viral DNAs were resolved on a 1.2% agarose gel and detected by standard Southern blot analysis using a ³²P-labeled HBV DNA probe.

RNA-packaging assay.

To detect viral RNA packaged inside NCs, NCs in MNase-treated cytoplasmic lysate were resolved on native agarose gels and detected by Northern blotting, without NaOH denaturation, before transfer to membranes, using a ³²P-labeled RNA probe specific for the HBV pgRNA, as previously described (9, 45). Under these conditions, we have shown previously that only pgRNA, and not plus-strand DNA, in NCs is detected (46, 47). The signal from ³²P-labeled viral RNA was detected using a Typhoon 9400 imager and analyzed with Quantity One software. HBV capsid proteins were detected using a mouse monoclonal antibody (MAb) specific for the N terminus of HBc (47). The chemiluminescent signal of core protein on Western blots was detected with the ChemiDoc MP system and analyzed with BioLab software.

siRNA knockdown.

On-Targetplus SMARTpool small interfering RNA (siRNA) against mouse cGAS and mouse STING and nontargeting siRNA were purchased from Dharmacon. AML12HBV10 cells were transfected with siRNA at a final concentration of 50 nM using Dharmafect 4 transfection reagent (Dharmacon) and maintained in medium without Tet for up to 16 days. To sustain the effect of siRNA knockdown, the cells were passaged at day 7 posttransfection preceding a second round of siRNA transfection on the following day. Cells were harvested every 2 days.

Western blot analysis.

Whole-cell extracts were separated by 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by Western blotting. Antibodies against STING (catalog number 13647), p-STAT1 (Y701; catalog number 7649), STAT1 (catalog number 9172), and β-actin (catalog number 4967) were purchased from Cell Signaling Technology. The antibody against cGAS was purchased from Millipore (catalog number ABF124). HBc was detected by a mouse monoclonal antibody specific for the N-terminal end of the core protein (47). The chemiluminescent signal of core protein on Western blots was detected with the ChemiDoc MP system and analyzed with BioLab software.

Measurement of capsid degradation and proteosome inhibitor treatment.

AML12HBV10 cells were cultured in Tet-free medium with dimethyl sulfoxide (DMSO) (0.1%) (mock treatment), ETV (100 nM), or BX795 (1 μM) for 6 days, and then the medium was replaced with Tet-containing medium supplemented with DMSO, ETV, or BX795. MG132 (1 μM) or epoxomicin (0.75 μM) was added to some mock treated cells at the same time as the Tet add back. Cells were harvested at 6 and 12 h post-Tet add back, and HBV core protein was detected by Western blotting.

Immunofluorescence.

Cells cultured on glass coverslips (18 by 18 mm) were washed with phosphate-buffered saline (PBS) three times before fixation using freshly prepared 4% paraformaldehyde in PBS for 15 min at room temperature. The fixed cells were washed three times with 30 mM glycine in PBS (pH 7.4), permeabilized with 0.1% Triton X-100 in PBS for 5 min at room temperature, and then washed again with PBS three times. After being blocked in 3% bovine serum albumin (BSA) in PBST (PBS with 0.1% Tween 20) for 1 h at room temperature or overnight at 4°C, the fixed cells were incubated with the rabbit anti-HBc (Dako) or rabbit anti-IRF3 (Abcam) antibody at 4°C overnight. The next day, the fixed cells were washed three times with PBS and incubated with goat anti-rabbit IgG conjugated with Alexa Fluor 488 for 2 h at room temperature. Then, the cells were washed three times with PBS, and the coverslips were subsequently mounted on slides in mounting medium containing 4′-6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). Images were collected using a Leica SP8 confocal microscope.

Northern blot analysis.

The RNaqueous kit (Life Technologies) was used for RNA extraction according to the supplier's directions. The RNA quality and concentration were assessed by spectrophotometry (Nanodrop; Thermo Scientific), and subsequent steps were normalized by RNA weight. The NorthernMax kit (Life Technologies) was used for detection of transcripts. Briefly, RNA was separated by size in a 1% agarose gel, transferred to a nylon membrane, UV cross-linked, and hybridized to ³²P-labeled HBV DNA probe. Signal detection was performed on a Typhoon 9400 imager and analyzed with ImageQuant 5.2 software.

Quantitative PCR.

RNA was extracted as described above, treated with DNase I (Promega) according to the supplier's directions, and then reverse transcribed using SuperScript III (Life Technologies) with oligo(dT) according to the supplier's directions. The cDNA quality and concentration were assessed by Nanodrop. Quantitative PCR (qPCR) was performed on three technical replicates in adhesive-sealed 384-well plates on a 12K Flex machine (Life Technologies) in 10-μl volumes consisting of 1× Power SYBR green master mix or 1× TaqMan gene expression master mix (Life Technologies), 200 nM each primer for SYBR green or 150 nM each primer with 250 nM probe for TaqMan, and 100 ng cDNA. The cycling conditions were 95°C for 10 min, 60°C for 15 s, and 55 cycles of 95°C for 15 s and 60°C for 1 min. The data were validated by verifying specificity in a melting curve analysis for SYBR green experiments or by running several PCR products on a 3% agarose gel for TaqMan experiments. Analysis was performed with Expression Suite software (Applied Biosystems, version 1.0) using the 2^−ΔΔCT method. Primer details are listed in Table 1 for SYBR green-based primer sets. β-Actin or TATA box binding protein (Tbp) was used for normalization of HBV RNA or cytokine RNAs, respectively. TaqMan-based primer-probe sets were from Life Technologies (Mm00439552 for IFN-β [79% efficiency] and Mm00446973 for Tbp [64% efficiency]). Tbp was used for normalization of IFN-β.

TABLE 1.

PCR primer information

Parametera	Value for gene:
	ActB	HBV mRNAs	Il-6	Tbp
GenBank accession no.	NM_007393.3	V01460.1	NM_031168.1	NM_013684.3
Amplicon length (bp)	160	96	146	145
In silico specificity screen with BLAST	Specific vs NCBI mouse RefSeq mRNAs	Specific vs NCBI genomic HBV sequences; not specific vs NCBI mouse RefSeq mRNAs	Specific vs NCBI mouse RefSeq mRNAs	Specific vs NCBI mouse RefSeq mRNAs
Location of each primer	Exon 6	2058 to 2153 (within the HBcAg coding region, before Pol)	Exon 4 to exon 5	Exon 6 to exon 7
Forward primer	CTCCTGAGCGCAAGTACTCTGTG	TACTGCACTCAGGCAAGCAA	AGCCAGAGTCCTTCAGAGA	GACTTCAAGATCCAGAACATGGT
Reverse primer	TAAAACGCAGCTCAGTAACAGTCC	AGGTCTCTAGACGCTGGATCT	TCCTTAGCCACTCCTTCTGT	TTCTGGGTTTGATCATTCTGTAGA
Amplicon Tm (°C)	83	78	75	80
Efficiency (%)	96	104	93	98

^aT_m, melting temperature.

ELISA.

Cell culture supernatants were collected, cleared of cell debris by centrifugation, and stored at −80°C until testing. All enzyme-linked immunosorbent assays (ELISAs) were prepared according to the suppliers' directions and read on a Synergy H1 plate reader (BioTek). The products used were the mouse IFN-α Platinum ELISA kit (eBioscience; BMS6027), the VeriKine-HS mouse IFN-β kit (PBL Assay Science; 42410), the mouse IL-28 (IFN-λ) ELISA kit (RayBiotech; ELM-IL-28), and TNF-α and IL-6 with a custom mouse Multi-Analyte ELISArray (Qiagen).

Statistical analysis.

Experiments were repeated at least three times, and the differences of the means between the groups were analyzed by independent t tests using SPSS version 22. A P value of <0.05 was considered statistically significant.

RESULTS

HBV gene expression and replication could not be maintained in AML12HBV10 cells.

In our efforts to better characterize HBV gene expression and replication in the mouse hepatocyte line AML12HBV10, we compared the viral gene expression and replication in these cells to those in the well-characterized human hepatoma cell line HepAD38 (43). Both cell lines express the HBV pgRNA under the regulation of a Tet-repressible promoter (39, 43). Analysis of HBV replication in the AML12HBV10 and HepAD38 cells over a time course of 15 days after Tet removal revealed a dramatic difference in viral replication kinetics: in contrast to the continuing accumulation of viral core DNA in the HepAD38 cells, core DNA declined sharply after reaching a peak level around days 7 to 9 postinduction in AML12HBV10 cells (Fig. 1A and B). Similarly, the viral core protein levels continued to increase in HepAD38 cells over the entire time course, whereas those in AML12HBV10 cells declined sharply after days 7 to 9 postinduction (Fig. 1C to E). Thus, HBV gene expression and replication could not be maintained in AML12HBV10 cells over time, even though the viral genome is stably integrated into the cellular genome.

FIG 1.

Time course of HBV expression and replication in AML12HBV10 and HepAD38 cells. AML12HBV10 cells (A, C, and E) and HepAD38 cells (B and D) were cultured without Tet for up to 15 days. Viral core DNA was isolated and detected by Southern blotting (A and B), and the HBV core protein was detected by Western blotting (C and D) or immunofluorescence (E, green staining) using the anti-HBc antibody. (C and D) The level of β-actin was used as a loading control. (E) DAPI was used to stain the nuclei (blue staining). RC, relaxed circular; SS, single stranded.

Blockade of viral DNA synthesis resulted in sustained viral gene expression.

As we reported recently (41), HBV NCs are porous in AML12HBV10 cells and are unable to protect their RC DNA content from exogenous nuclease, and this is associated with enhanced CCC DNA formation in the nucleus and exposure of viral DNA in the cytoplasm. The failure of AML12HBV10 cells to maintain HBV gene expression and replication, coupled with the destabilization of viral NCs in these cells, prompted us to hypothesize that NC destabilization may actually be responsible for the shutdown of viral replication in AML12HBV10 cells. Specifically, we postulated that the exposure of viral DNA due to NC destabilization in the cytoplasm might trigger innate immune sensing of the viral DNA and subsequent immune signaling to shut down viral gene expression and replication.

To test specifically the role of HBV DNA in triggering the putative antiviral response, AML12HBV10 cells were treated with ETV, an inhibitor of the viral RT that blocks viral DNA synthesis but not viral RNA or protein expression (48), or with AT-Br, an analogue of AT-61 that targets HBc and can inhibit the encapsidation of pgRNA into NCs (49) and consequently can also block viral DNA synthesis, but via a different mechanism and target. Cells were harvested every 3 days postinduction over a time course of 18 days, and viral gene expression and replication were monitored by determining the levels of viral RNAs, core proteins, capsid assembly, RNA packaging, and DNA synthesis. As anticipated, ETV was able to block viral DNA synthesis to undetectable levels (Fig. 2A). Moreover, AT-Br also could efficiently block viral DNA synthesis, as anticipated (Fig. 2B). Upon ETV or AT-Br treatment, viral RNA, core protein, and capsid assembly were all maintained at high levels without any sign of decline over the 18-day period (Fig. 2A, B, and D). Viral RNA packaging was also maintained during ETV treatment (Fig. 2A and D), but not during AT-Br treatment due to its known effect on RNA packaging (Fig. 2B). These results suggested that viral DNA, but not viral RNA or protein, was most likely responsible for the induction of the putative antiviral response in AML12HBV10 cells that shut down HBV gene expression and replication. Interestingly, when viral DNA synthesis was inhibited by either ETV or AT-Br, viral RNA and core protein levels were lower at the beginning of the time course (especially day 6), when the mock-treated cells reached peak viral levels (Fig. 2A, B, and D). These results suggested that the host response triggered by the viral DNA might have actually facilitated viral gene expression and replication when the viral levels were low at the beginning of the time course, even though it clearly inhibited viral gene expression and replication at the later points of the time course (see below).

FIG 2.

Effects of HBV RT inhibitor, viral-RNA-packaging inhibitor, and host TBK1 inhibitor on HBV gene expression and replication in AML12HBV10 cells. (A to C) AML12HBV10 cells were maintained in Tet-free medium with 0.1% DMSO (mock-treated control [Mock]) (A, B, and C), the HBV RT inhibitor (ETV; 100 nM) (A), the viral-RNA-packaging inhibitor (AT-Br; 20 μM) (B), or the TBK1 inhibitor (BX795; 1 μM) (C) for 18 days, and the cells were harvested every 3 days. Northern blotting was used for detection of viral RNAs (A, top blot). GAPDH (glyceraldehyde 3-phosphate dehydrogenase) was used as an RNA-loading control (A, 2nd blot from the top). For SDS-PAGE, cells were lysed with SDS sample buffers and resolved on SDS-10% polyacrylamide gels. HBc and β-actin were detected by Western blotting (A, 3rd and 4th blots from the top; B and C, 1st and 2nd blots, respectively). For native agarose gel analysis, cytoplasmic lysates prepared with NP-40 lysis buffer were resolved on agarose gels, and the packaged viral RNA (A, 5th blot from the top; B and C, 3rd blot from the top), DNA (A, 6th blot from the top; B, 4th blot from the top), and assembled capsid (A, 7th blot from the top; B, 5th blot from the top; C, 4th blot from the top) were detected by the RNA-packaging assay, Southern blotting, or Western blotting, respectively. 3.5 kb and 2.4/2.1 kb, HBV RNA species of the indicated sizes; NC-DNA, nucleocapsid-associated DNA or core DNA. (D) Quantification of viral gene expression and replication over a time course under different conditions. Relative expression levels of viral RNA (3.5 kb or 2.4/2.1 kb), core protein or HBc, capsid assembly, and RNA packaging during mock, ETV, AT-Br, and BX795 treatments were compared throughout the time course. The chemiluminescent signal of the core protein on Western blots was detected with the ChemiDoc MP system and analyzed with BioLab software. The signal from ³²P-labeled viral RNA was detected using a Typhoon 9400 imager and analyzed with Quantity One software. RNA signals from the Northern blots were normalized to GAPDH or β-actin, and HBc protein signals on Western blots were also normalized to β-actin. All signals were compared to the signal of mock-treated cells at day 6, which was set to 1. Individual experiments were performed at least three times, and the differences of the means between the groups were analyzed by an independent t test using SPSS version 22. *, P < 0.05. The shades of the asterisks refer to the corresponding experimental conditions as represented by the shades of the lines. The error bars indicate standard deviations.

Blockade of host cell DNA sensing and innate immune signaling resulted in maintenance of viral gene expression.

The above-mentioned results strongly implicated HBV DNA, which was exposed due to NC destabilization in AML12HBV10 cells, as a putative PAMP that was responsible for triggering the innate antiviral response observed. To ascertain if cGAS, a major DNA sensor involved in cytosolic DNA sensing, and STING, a major signal transducer in DNA-induced antiviral response, played a role in the induction of the antiviral response in AML12HBV10 cells, we reduced the expression of cGAS or STING by siRNA knockdown. The efficiency of knockdown was confirmed by Western blotting (Fig. 3). Knockdown of either cGAS or STING indeed allowed sustained expression of the viral core protein, capsid assembly, and RNA packaging (Fig. 3). As observed above, with inhibition of viral DNA synthesis, the levels of viral core protein, capsid assembly, and RNA packaging were decreased by cGAS or STING knockdown at earlier time points when the mock-treated cells reached peak viral levels (Fig. 3), again suggesting that these host factors might actually be involved in stimulating viral gene expression at the earlier time points.

FIG 3.

Effect of cGAS or STING knockdown on HBV replication. (A and B) Knockdown of cGAS (A) or STING (B) was achieved by siRNA transfection, as described in Materials and Methods. For SDS-PAGE, cells were lysed with SDS sample buffers and resolved on SDS-10% polyacrylamide gels. HBc (A and B), cGAS (A), STING (B), and β-actin (A and B) were detected by Western blotting. For native agarose gel analysis, cytoplasmic lysates prepared with NP-40 lysis buffer were resolved on agarose gels, and the packaged viral RNA and assembled capsid were detected by the RNA-packaging assay or Western blotting, respectively. (C) Quantification of the kinetics of HBV core protein expression, capsid assembly, and RNA packaging in AML12HBV10 cells upon STING or cGAS knockdown. Relative levels of the HBc protein, capsid, and packaged RNA following mock treatment or cGAS or STING siRNA treatment were normalized to the scrambled siRNA control signals at day 4, which were set to 1. Quantifications of the different signals were done as described in the legend to Fig. 2D. HBc protein signals on Western blots were normalized to β-actin. Individual experiments were performed at least three times, and the differences of the means between the groups were analyzed by an independent t test using SPSS version 22. *, P < 0.05. The shades of the asterisks refer to the corresponding experimental conditions as represented by the shades of the lines. The error bars indicate standard deviations.

To understand further the signaling pathways involved in downregulating HBV replication in AML12HBV10 cells, we treated the cells with an inhibitor of TBK1 (BX795), which is a major cellular kinase downstream of STING in mediating the innate response to cytosolic DNA, as well as other PAMPs (18, 50). Consistent with an involvement of TBK1 in mediating the induction of the antiviral response in AML12HBV10 cells, treatment with BX795 allowed sustained viral gene expression; levels of viral RNAs, core protein, capsid assembly, and RNA packaging at the end of the time course (day 18 postinduction) were approximately the same or increased relative to earlier time points (Fig. 2C and D). As observed (see above) with inhibition of viral DNA synthesis or cGAS and STING knockdown, the levels of viral gene expression were generally lower in BX795-treated cells at earlier time points (Fig. 2C and D), again suggesting that the innate response blocked by these inhibitors may have enhanced viral gene expression at the earlier time points (see Discussion).

TBK1 is known to induce the nuclear translocation of the transcription factor IRF3 to induce type I IFN, as well as other antiviral genes (18). To test whether HBV DNA synthesis could activate the IRF3 pathway, AML12HBV10 cells were maintained in Tet-free medium with or without ETV, AT-Br, or BX795 for 6 days, and the subcellular localization of IRF3 was determined by immunofluorescence staining and confocal microscopy (Fig. 4). Indeed, viral replication resulted in strong IRF3 nuclear translocation (2nd row) compared to the Tet⁺ condition (no viral induction) when IRF3 was located in the cytoplasm (1st row). Viral-replication-induced IRF3 nuclear translocation could be abrogated by blockade of viral DNA synthesis, either with the RT inhibitor ETV (3rd row) or the RNA-packaging inhibitor AT-Br (4th row). As expected, blockade of TBK1 phosphorylation by BX795 could also prevent IRF3 nuclear translocation (5th row). These results thus indicated that viral DNA synthesis could indeed induce IRF3 nuclear translocation following TBK1 activation.

FIG 4.

Detection of IRF3 subcellular localization by immunofluorescence. AML12HBV10 cells were induced to replicate HBV in Tet-free medium for 6 days in the presence of 0.1% DMSO (mock) (2nd row) or the presence of ETV (100 nM) (3rd row), AT-Br (20 μM) (4th row), or BX795 (1 μM) (bottom row). The cells were then harvested for immunofluorescence assay to determine IRF3 (green) subcellular localization using confocal microscopy. Cells cultured in the presence of Tet (uninduced) are shown in the top row. DAPI (blue) was used to stain the nuclei.

In summary, blockade of TBK1 signaling, as well as inhibition of cGAS and STING expression, could restrain the antiviral response and allow sustained HBV gene expression in AML12HBV10 cells, which would otherwise be shut off due to the cellular response triggered by the viral DNA and transduced via the cGAS-STING-TBK1-IRF3 pathway. On the other hand, the same treatments also led to a decrease in viral gene expression at earlier time points.

The viral-DNA-induced antiviral effect results in core protein degradation through the proteasome pathway.

Although a weak effect on the viral RNAs (the 3.5-kb pgRNA and the 2.4-kb/2.1-kb envelope mRNAs) was observed, in terms of both the decline during the time course and the changes caused by blocking viral DNA synthesis or immune signaling, the effect on the viral core protein (and capsid assembly and RNA packaging) was stronger (Fig. 2D). Also, an effect of antiviral cytokines on the stability of HBV capsids in AML12HBV10 cells has been observed before (39), and HBV capsids can be degraded by the proteasome (51). Therefore, we were interested in testing whether the viral-DNA-induced host response could result in degradation of the viral core protein. To test this hypothesis, AML12HBV10 cells were maintained for 6 days in Tet-free medium to induce viral DNA synthesis, with or without ETV or BX795, to evaluate the effect of viral DNA or TBK1 signaling. At day 6, Tet was added back to stop viral gene expression at the RNA level. The cells were then harvested at 0 h, 6 h, and 12 h post-Tet add back, and the levels of viral core protein were detected by Western blotting (Fig. 5A). At 12 h post-Tet add back, the level of core protein declined dramatically (Fig. 5B). This decrease could be blocked by treatment with ETV or BX795, which were included in the medium during the 6-day induction period. Moreover, two different proteasome inhibitors, MG132 and epoxomicin, could also prevent the decrease of core protein when added at the same time as Tet add back (Fig. 5B). These results thus suggested that the viral-DNA-induced response could indeed trigger core protein degradation through TBK1 signaling and that the degradation was mediated by the proteasome.

FIG 5.

Degradation of HBV core protein through the proteasome pathway. (A) Schematic diagram of the experimental design. AML12HBV10 cells were cultured in Tet-free medium with DMSO (mock; 0.1%) (group 1), ETV (100 nM) (group 2), or BX795 (1 μM) (group 3) for 6 days, and then the medium was replaced with Tet-containing medium with DMSO (group 1a), ETV (group 2), or BX795 (group 3). In addition, two other groups of mock-treated cells were also treated with MG132 (1 μM) (group 1b) or epoxomicin (0.75 μM) (group 1c), respectively, at the same time as Tet add back. Cells were harvested at 0, 6, and 12 h post-Tet add back. (B) HBV core protein (HBc) and β-actin were detected by Western blotting following SDS-PAGE.

To identify putative effectors that mediated the observed antiviral effects, we attempted to measure the induction of antiviral cytokines that have been reported to be active against HBV (see the introduction). However, we failed to detect IFN-α, IFN-β, IFN-λ, TNF-α, and IL-6 in the culture supernatant of AML12HBV10 cells by ELISA when viral gene expression and replication were clearly being suppressed (Table 2). Also, efforts to detect cytokine induction by RT-PCR assays also failed to reveal any consistent induction of IFN-β or IL-6 at the RNA level by HBV DNA (data not shown).

TABLE 2.

Cytokine levels in AML12HBV10 cell supernatants

Cytokine(s)a	LOD (pg/ml)a	Conditionsb	Resultc
IFN-α2/4	7.5	1	At or below LOD
IFN-β	0.94	1, 2, 3	Below LOD
IFN-λ2	2.0	1, 2, 3	Below LOD
IL-6	58.8	1, 2, 3	At or below LOD
TNF	30.5	1, 2, 3	At or below LOD

^aAML12HBV10 cell supernatants cultured under different conditions were tested for the indicated mouse cytokines via ELISA using commercial kits, as described in Materials and Methods, with the indicated lower limits of detection (LOD).

^bThe treatment conditions that were compared were as follows: 1, Tet off versus Tet on; 2, ETV or BX795 treated versus mock treated for 10 days, all with Tet off; and 3, Tet off for 2 days (very low HBV expression) versus Tet off for 10 days (just after peak HBV expression and replication). All the cytokines were measured in a minimum of three separate experiments.

^cThe cytokine levels in the culture supernatants were always at or below the detection limit. No differences were observed between the different treatment conditions.

DISCUSSION

Our results implicate HBV DNA as a potential PAMP that can be recognized by the host DNA sensors (as PRRs) to trigger an innate host response that can regulate viral gene expression and replication. For reasons yet to be clarified, mature HBV NCs become hyperdestabilized in the cytoplasm of AML12HBV10 cells (41), exposing their interior DNA content, which was apparently recognized by the host cGAS-STING DNA-sensing mechanism (and probably other pathways, as well [see below]). A putative pathway of innate immunity activation (immune induction) during HBV replication can be outlined (Fig. 6). Thus, a role for the HBV DNA as a PAMP in triggering the response was strongly supported by the fact that inhibition of viral DNA synthesis by multiple approaches (without inhibiting viral RNA or protein expression) was sufficient to block the antiviral response. Moreover, a role for a major cytosolic DNA sensor, cGAS, and a major signal transducer involved in DNA sensing, STING, in immune induction was implied, since knockdown of these factors could abrogate (at least partially) the antiviral response. Similarly, a role for TBK1, which acts downstream of cGAS-STING activation, was indicated by the effects of TBK1 inhibition. Downstream activation of IRF3 could also be demonstrated by IRF3 nuclear localization. However, a role for additional DNA sensors and signal transducers is by no means excluded and requires further investigation. Indeed, the effects of TBK1 inhibition or siRNA knockdown of cGAS and STING on viral capsid protein or pgRNA packaging were generally not as strong as those achieved by inhibition of viral DNA synthesis, suggesting that the effects of viral DNA synthesis might be mediated by other cellular factors, in addition to the cGAS-STING-TBK1 pathway.

FIG 6.

Simplified model for CCC DNA synthesis and innate immune induction by HBV in AML12HBV10 cells. HBV NCs are destabilized in AML12HBV10 cells for reasons yet to be clarified. This leads to enhanced uncoating and delivery of RC DNA to the nucleus for CCC DNA synthesis. On the other hand, NC destabilization also leads to the exposure of RC DNA in the cytoplasm, which can be recognized by the cytosolic DNA sensor cGAS, leading to the activation of the STING-TBK1 pathway and IRF3 nuclear translocation. Innate antiviral factors (or proviral factors, at the early points of the time course, when viral levels are low) are then induced to modulate HBV gene expression (at the RNA and core protein levels). The inhibitors and manipulations used to dissect the antiviral response pathway are indicated where they are thought to act. Other factors/pathways are also likely involved but are not depicted for simplicity. See the text for details.

The identities of the antiviral factors (immune effectors) induced by HBV DNA remain to be elucidated. Previous studies suggested that type I IFN or IL-6 might be induced by HBV in an immortalized human hepatocyte line (HepaRG) or in Kupffer cells (26, 27), although neither the viral PAMP nor the cellular PRR was identified. Here, we failed to detect the induction of IFN-α, IFN-β, IFN-λ, TNF-α, and IL-6 when viral gene expression and replication were clearly being suppressed, suggesting that these cytokines were not induced or were induced only at very low levels in AML12HBV10 cells. These results are consistent with reports of a lack of strong type I IFN induction during natural HBV infection (19, 22). The putative antiviral effectors could be specific to hepatocytes, as cell-type-specific innate response to cytosolic DNA has been reported (52). In addition, both immune induction and effector mechanisms could be further complicated if HBV can indeed perturb innate immune signaling (25).

Consistent with previous studies showing that the host immune response can inhibit HBV at multiple stages of the viral replication cycle and that the effects may vary depending on the experimental systems (25, 32–34, 37, 39, 53), the antiviral response induced in AML12HBV10 cells also appeared to affect multiple steps of HBV gene expression and replication. A modest effect on viral RNA levels was observed. Moreover, core protein degradation, via the proteosome, was also stimulated. The effects on multiple stages of the HBV life cycle, together with the apparent lack of strong induction of the classical antiviral cytokines, suggest that multiple effector mechanisms might be induced, each at relatively low levels, which contributed cooperatively to the overall strong antiviral effects observed.

Recently, it has been reported that an innate immune response could actually facilitate, rather than inhibit, HBV replication when viral levels are low (38). Some specific ISGs that facilitate viral replication have also been identified recently (54). In AML12HBV10 cells, early during induction when viral levels were low, blocking innate sensing or signaling (blocking HBV DNA synthesis, knockdown of cGAS and STING, or TBK1 inhibition) led to a decrease in HBV gene expression, which is consistent with the notion that HBV replication may indeed be stimulated by an innate immune response when viral levels are low. Future studies will be required to further elucidate the mechanism of this putative proviral effect of the innate host response.

Our observations linking HBV NC destabilization and innate immunity induction suggest that the key to viral evasion of host immunity versus triggering of the innate response may critically depend on the kinetics and timing of viral NC destabilization and disassembly (uncoating), which controls the exposure of its interior DNA content to host DNA sensors. Unfortunately, the mechanism of HBV NC uncoating remains poorly understood. Some reports suggest that mature NCs may be disassembled only when they reach the cytoplasmic side of the nuclear pore ready to release their DNA content into the nucleus for conversion to CCC DNA (55, 56). Thus, exposure of HBV DNA in the cytosol may be very limited or not occur at all during the initial stage of infection, which likely contributes to the stealth characteristics of HBV as an effective strategy of immune evasion. Furthermore, it remains to be determined if a viral-DNA-triggered innate response indeed occurs during a natural infection of normal hepatocytes in the human liver. However, we consider it plausible that as the infection progresses and spreads, the inherently unstable mature NCs (9) accumulate to high levels in the cytoplasm of an increasing number of hepatocytes, raising the chance, stochastically, of aberrant NC uncoating in the cytosol and thus exposure of RC DNA to the cytosolic DNA sensors. Furthermore, as we reported recently (41, 57), both host cell conditions and viral capsid mutations can disrupt the integrity of mature NCs and thus enhance the exposure of RC DNA. These findings raise the possibility that rare infected cells or HBc mutations, leading to enhanced RC DNA exposure, may arise as the infection proceeds, thus contributing to the induction of innate immunity by triggering the host DNA-sensing mechanism. Furthermore, the peculiar structure of the HBV RC DNA, a “broken” molecule, may render it particularly potent at triggering the cellular DNA sensors, given the connection between DNA damage and innate DNA sensing (58). Reports of at least some innate immune responses during natural or experimental HBV infections lend support to this suggestion (19, 59–61). By enhancing cytosolic viral DNA exposure that presumably occurs only very infrequently during natural HBV infection, AML12HBV10 cells may thus provide a valuable system to study this important but elusive aspect of HBV biology.

DNA sensing can also trigger an inflammatory response and cell death (11, 15, 62). These processes may be related to the pathogenesis of the inflammatory diseases associated with HBV-induced hepatitis, leading ultimately to cirrhosis and HCC. Indeed, HBV RT inhibitors (nucleoside analogues, such as ETV, used here) are highly effective in suppressing the necroinflammatory process in the HBV-infected liver, slowing/reversing cirrhosis and delaying/preventing HCC (48, 63, 64). This anti-inflammatory effect is observed despite the fact that RT inhibitors have no direct effect on viral gene expression but instead only block viral DNA synthesis. This clinical efficacy of RT inhibitors could be interpreted to mean that HBV DNA itself is a major trigger of inflammatory diseases, possibly via triggering of the host DNA-sensing mechanisms described here and others yet to be uncovered. In support of this conjecture, serum HBV DNA levels have been shown to be the most important predictor of HCC development (65), and suppression of HBV DNA levels by RT inhibitors reduces HCC in proportion to their DNA suppression potencies (64). Thus, enhanced cytoplasmic HBV DNA exposure may exacerbate, at least temporarily, the inflammatory disease. Nevertheless, our results here suggest the tantalizing possibility that manipulations of HBV NC uncoating and viral DNA exposure may have therapeutic potential by triggering the viral-DNA-induced innate immune response to eliminate viral persistence (i.e., induction of viral suicide).

Funding Statement

The Hepatitis B Foundation provided funding to Ju-Tao Guo through an appropriation from the Commonwealth of Pennsylvania.