To investigate the in vivo innate immune response to HBV, HBV infectious particles were intravenously injected to tree shrews, and total RNA was isolated from the liver at 0, 1, and 3 days post-infection (Figure 1A). RNA-seq analysis was performed using a next-generation sequencer. A heatmap of all genes suggested that expression of most genes was not altered (Figure 1B and Figures S1A,B in Supplementary Material); however, there were several genes whose expression was affected by HBV injection (Figure 1C), and the induced expression of MAP3K2, RNase L, MDA5, and CD69 was detected (Figure 1C and Figure S1C in Supplementary Material). The IFN gene expression was not detected by RNA-seq analysis because of its low expression level.

To detect IFN-β, -γ, and -λ expression in tree shrew tissues, we performed RT-qPCR. HBV infection did not increase the expression of IFN-β in the liver, spleen, and kidney (Figure 1D and Figure S1D in Supplementary Material). In contrast, IFN-γ expression was specifically increased in the liver at 1 and 3 days post-infection (Figure 1D and Figure S1D in Supplementary Material). Considering that the adaptive immune response cannot produce IFN-γ within 1 day of primary infection, this early IFN-γ expression suggests that group 1 innate lymphoid cells (ILCs) are responsible for the early IFN-γ expression. In addition, HBV intravenous injection increased hepatic DDX60 expression but not RIG-I, cGAS, or IFI16 (Figure 1D).

When hepatocyte cell lines, HepG2 and HuH-7, were transfected with a plasmid, pHBV, which carries 1.4× HBV genomic DNA and produces pregenomic RNA and all viral proteins, HBV RNA appeared at 3 h after transfection. However, the expression levels of DDX60 and IFN-γ did not increase until 24 h after transfection (Figure 2A and Figure S1E in Supplementary Material). As reported previously (13), HBV increased the RIG-I-dependent expression of IFN-λ in HuH-7 cells (Figure 2B). To further investigate the response of hepatic cells to HBV, we used primary hepatocytes and hepatic stellate cells. Infection of primary hepatocytes and hepatic stellate cells with HBV failed to increase IFN-γ and DDX60 expression (Figures 2C,D). These observations imply that hepatocytes and hepatic stellate cells cannot induce IFN-γ and DDX60 in response to HBV, and that non-parenchymal cells are required for hepatic IFN-γ and DDX60 expression.

Hepatitis B virus infects hepatocytes but not non-parenchymal cells. Therefore, it was expected that hepatocytes would release the molecule that activates non-parenchymal cells. It should be noted that HCV-infected hepatocytes release the exosomes containing viral RNA, thereby leading to the activation of dendritic cells via TLRs (9), so we investigated whether HBV RNA is present in EVs including exosomes in a similar manner to HCV. First, we isolated EVs from cell culture medium of hepatocytes. We confirmed that EVs contained the CD9 protein, which is a marker of exosomes (Figure 3A). Interestingly, HBV RNA, but not host GAPDH mRNA, was detected within EVs released by pHBV-transfected hepatocytes (Figure 3B). Second, we isolated exosomes from HepG2 cell culture medium using anti-CD81 antibody beads, because exosomes released from HepG2 contain CD81 (18). Using the anti-CD81 beads, we could concentrate exosomes, and CD81, CD9, and CD69, which are a marker of the exosomes, were detected in the exosomes by western blotting (Figure 3C). As expected, HBV RNA was also detected in the CD81⁺ exosomes (Figure 3D), and viral DNA was also detected in the CD81⁺ exosomes (Figure S2A in Supplementary Material). In addition, CD81⁺ exosomes released from HBV-infected HepG2-NTCP cells contained HBV RNA (Figure S2D in Supplementary Material).

NK cells are the major group 1 ILCs in liver. In general, NK cells can produce IFN-γ (19), whereas Kupffer cell, which are hepatic macrophages, activates T and NK cells, resulting in IFN-γ production by T or NK cells (20). Hepatic NK cells isolated from mice failed to produce IFN-γ in the presence of HepG2-T23 cells, which stably express pHBV and produce HBV infectious particles (Figure 3E). Interestingly, hepatic NK cells produced IFN-γ in the presence of both HepG2-T23 and hepatic F4/80⁺ cells, which is a marker of macrophages (Figure 3E), suggesting the importance of macrophages. Therefore, we next, investigated whether EVs released from hepatocytes with pHBV affect macrophage function.

Extracellular vesicles were isolated from hepatocytes with or without pHBV. HBV transfection to hepatocytes barely affected the CD63 protein levels of EVs (Figures 3F,G). Interestingly, EVs released from hepatocytes with pHBV increased the mRNA expression of NKG2D ligands (MICA, ULBP1, ULBP2, and Rae1) in THP-1 macrophages or hepatic F4/80⁺ cells (Figures 3F,H). NKG2D ligands are known to elicit IFN-γ production from NK cells (21, 22). These data are correlated with upregulation of CD69, a NK cell activation marker, in tree shrew liver after HBV intravenous infection (Figure S1C in Supplementary Material). We confirmed that infection of HepG2-NTCP with HBV cells substantially increased the expression of ULBP1 and ULBP2 mRNA in co-cultured THP-1 macrophages, which was cultured using transwell co-culture system (Figure 3I).

To assess the role of exosomes in NKG2D ligand expression, we performed a depletion assay. The depletion of CD81⁺ exosomes attenuated the EVs-mediated ULBP1 and ULBP2 expression in THP-1 macrophages (Figure 4A), indicating the importance of CD81⁺ exosomes. Interestingly, knockdown of MyD88 reduced EVs-mediated ULBP1 and ULBP2 expression, and ULBP2 expression was also reduced by TICAM-1 or VISA knockdown (Figure 4B). These results indicate that the CD81⁺ exosomes stimulate the MyD88, TICAM-1, and VISA pathways. Knockdown of PYCARD moderately decreased the expression of ULBP2 (Figure 4B); however, the ULBP2 expression was induced by TLRs and RLRs stimulation even in the absence of PYCARD stimulation (see below).

To investigate whether exosomal nucleic acids are sufficient for the NKG2D ligand expression, exosomal RNA and DNA were extracted from exosomes released from HepG2 cells with pHBV, and THP-1 macrophages were stimulated with exosomal nucleic acids. The expression of ULBP2 in THP-1 macrophages was substantially increased by stimulation with exosomal RNA and DNA (Figure 5A). Viral RNA and DNA are well known to activate MyD88, TICAM-1, and VISA pathways, and thus, we investigated whether stimulation of TLRs and RLRs pathway mimics EVs that were released from HBV-infected cells. To stimulates the TLR3–TICAM-1 pathway, 50 μg of polyI:C was added to culture medium, and 1 μg of polyI:C was transfected to stimulate the RLRs–VISA pathway. For stimulation of TLR7 and TLR9, which trigger the signal via MyD88, 1 μg/ml of CL097 and 500 μM of ODN2216 were added to cell culture medium, respectively. The simultaneous stimulation of all four pathways with TLR3/RIG-I/MDA5, TLR7, and TLR9 ligands (polyI:C, CL097, and ODN2216) induced the expression of IL-12p35 as well as ULBP1 and ULBP2, whereas the stimulation of each separate pathway failed to increase the expression levels (Figure 5B and Figures S3A,B in Supplementary Material). IL12-p40 mRNA expression was increased by stimulation with both polyI:C (addition and transfection) and CL097, even in the absence of ODN2216 (Figure S3B in Supplementary Material). Given that EVs released by hepatocytes with pHBV did not increase IL-12p35 mRNA expression in THP-1 macrophages (Figure 3F), we hypothesized that EVs contain the agonists for each pathway as well as other molecules that regulate innate immune response.

The exosomes deliver miR to other cells and regulate the innate immune response (11). Thus, we investigated the expression levels of 17 immunoregulatory miRs within EVs and exosomes (23). Interestingly, HBV increased the expression levels of miR-21 and miR-29a in EVs and CD81⁺ exosomes (Figure 5C and Figure S3C in Supplementary Material). Moreover, exosomal miR-21 and miR-29a levels were markedly increased by infection of HepG2-NTCP cells with HBV (Figure 5D). miR-21 downregulates IL-12p35 mRNA expression (24) and is induced by the HBx proteins (25). miR-29a is known to suppress IL-12p40 mRNA expression (26).

Interestingly, EVs released from hepatocytes reduced IL-12p40, but not IFN-β expression, in THP-1 macrophages that were stimulated with CL097 and polyI:C (addition and transfection) (Figure 5E). Moreover, EVs released from HepG2 with pHBV reduced stimulation-induced IL-12p35 expression in THP-1 macrophages compared to EVs from HepG2 without pHBV (Figure 5F). We confirmed that the expression of HBV in hepatocytes increased the levels of miR-21, miR-29a, and other immunoregulatory miRs in THP-1 macrophages via EVs (Figure 5G). In addition, miR-29a transfection to THP-1 macrophage decreased IL-12p40 expression (Figure S3D in Supplementary Material), and EVs released from hepatocytes that were transfected with miR-29a decreased IL-12p40 expression in THP-1 macrophages (Figure 5H). Overall, these results indicate that HBV-induced immunosuppressive miRs were transferred to macrophages from hepatocytes, thereby suppressing the IL-12 expression.

Hepatitis B virus intravenous infection increased the expression of a NKG2D ligand, ULBP3, in tree shrew liver but failed to increase IL-12p35 and IL-12p40 in vivo (Figure 1D). These data are nicely correlated with our in vitro studies.

IFN-γ itself has many functions, such as promotion of Th1 differentiation and macrophage activation. A previous study showed that IFN-γ destabilizes viral RNA (15); however, underlying mechanism is not fully elucidated. To investigate the physiological meaning of EVs-induced IFN-γ production in antiviral innate immune response, we sought to assess the role of IFN-γ in viral RNA degradation. To observe viral RNA degradation, HepG2 cells were transfected with pHBV, and viral RNA expression was attenuated by actinomycin D (ActD), and then viral RNA levels were determined by RT-qPCR. As previously reported, IFN-γ treatment promoted the degradation of viral RNA (Figure 6A). Interestingly, IFN-γ treatment increased the DDX60 mRNA and protein levels in hepatocytes (Figures 6B–E). It has been shown that DDX60 is involved in a viral RNA degradation pathway (6). The ectopic expression of DDX60 promoted the degradation of HBV RNA (Figure 6F). In contrast, DDX60 knockdown delayed the degradation of HBV RNA (Figures 6G,H). Interestingly, cytoplasmic viral RNA was degraded faster than nuclear viral RNA, and DDX60 knockdown delayed the degradation of cytoplasmic viral RNA but not nuclear viral RNA (Figures 6I,J). These data suggest that IFN-γ induces the expression of DDX60, which promotes viral RNA degradation.

Next, we assessed the role of DDX60 in viral replication. The ectopic expression of DDX60 decreased the HBV RNA levels in hepatocytes with pHBV (Figure 6K), and DDX60 knockdown substantially increased HBV RNA levels (Figure 6L). The HBsAg levels were also reduced in the culture medium by DDX60 ectopic expression (Figure 6M). IFN-γ-treatment reduced HBsAg, and the IFN-γ-mediated reduction of HBsAg was alleviated by DDX60 knockdown (Figure 6N), thereby suggesting the importance of DDX60 in the IFN-γ-antiviral activity against HBV.

Next, we further assessed the role of DDX60 during viral infection. IFN-γ treatment reduced the viral RNA levels in HBV-infected HePG2-NTCP cells (Figure S4A in Supplementary Material), and DDX60 knockdown increased the HBV RNA levels in cells infected with HBV for 6 days (Figure 6O). Collectively, these results indicate that DDX60-mediated HBV RNA degradation suppresses viral replication. Using a hydrodynamic injection mouse model, we confirmed that IFN-γ treatment could reduce the serum HBsAg level in vivo after HBV injection into mouse liver (Figure S4B in Supplementary Material). These results support a model that exosomes-mediated hepatic IFN-γ production plays a crucial role in antiviral innate immune response to HBV.

Tree shrews (T. belangeri chinensis) were injected intravenously with 10⁷ copies of infectious HBV particles (C_JPNAT, accession number: AB246345.1). Tissues were extracted at days 0, 1, and 3. Total RNA was extracted with TRIzol (Invitrogen). C57BL/6 and BALB/c mice were purchased from Hokudo and SANKYO LABO SERVICE. All of the animal studies were conducted in strict accordance with the Guidelines for Animal Experimentation of the Japanese Associations for Laboratory Animal Science. The protocols were approved by the Animal Care and Use Committee of Hokkaido University, Japan (permit numbers: 09-0215, 13-0049, and 15-0017) and the ethics committee of the Tokyo Metropolitan Institute of Medical Science.

C57BL/6 mouse liver was dissociated into single-cell suspensions using a Liver Dissociation Kit (Miltenyi Biotec), according to the manufacturer’s instructions. Hepatic F4/80⁺ cells were isolated using anti-F4/80-biotin antibody and anti-biotin MicroBeads (Miltenyi Biotec). Liver NK cells were isolated using an NK cell isolation kit II (Miltenyi Biotec), according to the manufacturer’s instruction. THP-1 cells were purchased from JCRB cell bank and cultured in RPMI-1640 with 5% FCS. THP-1 cells were differentiated into macrophages with 60 ng/ml of PMA for 16 h. For stimulation of THP-1 macrophages, ODN2216 (500 μM), CL097 (1 μg/ml), and polyI:C (50 μg/ml) were added to cell culture, or polyI:C (50 μg/ml) were transfected into cells using 2 μl of lipofectamine 2000 (addition and transfection) in 24-well plate. For activating VISA pathway, 1 μg of polyI:C were transfected into cells in 24-well plate. HepG2-NTCP cells were cultured with D-MEM/F-12 + GlutaMax with 10 mM HEPES, 200 U/ml penicillin, 200 μg/ml streptomycin, 10% FCS, 50 μM hydrocortisone, and 5 μg/ml insulin. HepG2-T23 cells, which stably express HBV, were kindly donated by Chayama (Hiroshima University) (42). HepG2-NTCP cells, which stably express NTCP for the HBV receptor, were donated by Watashi (NIID) (43). Human primary hepatocytes and human primary hepatic stellate cells were purchased from ScienCell Research Laboratories. We used human and mouse IFN-γ (Cell Signaling Technology), actinomycin D (Life Technologies), anti-DDX60 antibody (SIGMA-Aldrich), anti-β-actin antibody (AC-15: SIGMA-Aldrich), and an HBsAg ELISA kit (XpressBio). ExoAB antibody kit, which includes anti-CD9, anti-CD63, and anti-CD81 antibodies, was purchased from System Biosciences.

HepG2-NTCP cells were treated with 10 ng/ml of IFN-γ for 8 h. Cells were washed with PBS, fixed with formaldehyde, and permeabilized with 0.2% of Triton-X 100 in PBS. Cells were subsequently blocked with 1% of BSA in PBS for 10 min and then incubated with anti-DDX60 rabbit polyclonal antibody (SIGMA-Aldrich) at 1/100 dilution. Cells were washed 4× with 1% of BSA in PBS and stained with Alexa Fluor 488 anti-rabbit antibodies. Cells were embedded with ProLong Gold Antifade Mountant with DAPI (Life Technologies).

Total RNA extracted from tissues of tree shrew was extracted using TRIZOL (Invitrogen), according to the manufacturer’s instruction. RNA-seq libraries were prepared using TruSeq RNA-seq kit (illumina). Sequencing libraries were sequenced on MiSeq using MiSeq sequencing reagent kit ver2 (illumina), according to the manufacturer’s instruction. The RNA-seq tags were then mapped to the reference transcripts of Tupaia chinensis. Normalization and transformation of expression values and hierarchal clustering analysis were carried out using CLC genomics workbench software. RNA-seq data have been deposited to DRA in DDBJ (accession number: DRA004164).

NK and F4/80⁺ cells were isolated from mouse liver. The 1 × 10⁶ NK cells and 1 × 10⁶ F4/80⁺ cells were co-cultured with 2 × 10⁵ HepG2 cells expressing HBV for 24 h. Culture supernatant was collected and analyzed for mouse IFN-γ by ELISA (GE Healthcare). Concentration of HBV surface antigen (HBsAg) in the serum and culture supernatant was quantified by ELISA, according to the manufacturer’s instruction (XpressBio).

Hydrodynamic injection was performed with TransIT-EE Hydrodynamic Delivery Solution (Takara), as described previously (44).

HepG2 or HuH-7 cells cultured in 60-mm dish were washed twice with serum-free medium and further incubated with serum-free medium for 24 h. Culture supernatant was recovered and centrifuged at 2000 rpm for 30 min to remove debris. Obtained supernatant was mixed with total exosome isolation solution and incubated at 4°C for 24 h, according to the manufacturer’s instruction (Thermo Fisher Scientific). Mixture was centrifuged at 1000 × g for 1 h, and then pellet was suspended with 0.5 ml of serum-free medium. The 0.1 ml of suspended exosome solution was added to 1 well of 24-well plate to activate THP-1 cells. CD81⁺ exosomes were isolated using cell culture total exosome isolation kit and CD81⁺ exosome isolation kit, according to the manufacturer’s instruction (Thermo Fisher Scientific).

The plasmid carrying the 1.4× HBV genome (HBV plasmid), in which HBV RNA is transcribed from the pCMV minimal promoter, was kindly gifted by Chayama. The DDX60 expression vector has been described elsewhere (6). HuH-7 cells were transfected with 0.2 μg of HBV plasmid using Lipofectamine 2000 reagent (Invitrogen). One to 3 days after transfection, the cells were treated with 10 μg/ml actinomycin D in order to inhibit HBV RNA transcription. Total RNA was extracted with the TRIZOL reagent (Invitrogen), according to the manufacturer’s instructions. Cytoplasmic and nuclear RNA were extracted with the Cytoplasmic and Nuclear RNA purification kit (NORGEN BIOTEK CORP.), according to the manufacturer’s instructions. Viral RNA levels were determined by qPCR.

Reverse transcription reactions of mRNA and miR were performed using the High-Capacity cDNA Reverse Transcription Kit (Life Technologies) and Mir-X miRNA First-strand Synthesis kit (Takara). Real-time PCR was performed with the SYBR Green Real-Time PCR Master Mix (Life Technologies) using the Step One Real-Time PCR System (Life Technologies). PCR primers used for the RT-qPCR procedures are described in the Supplementary Material.

To determine exosomal miR levels, HepG2-NTCP cells were seeded onto 24-well plate. The 40,000 genome copies of HBV isolated from HepG2-T23 cell culture medium were infected to HepG2-NTCP cells for 3 days. Cells were washed 3× with cell culture medium and were further incubated for 3 days. CD81⁺ exosomes were isolated from cell culture medium, and exosomal miR and HBV RNA levels were determined by RT-qPCR and normalized to U6 level.

To investigate the role of DDX60 in the suppression of viral replication, HBV particles isolated from the cell culture supernatants of HepG2-T23 cells were used to infect HepG2-NTCP cells at 40,000 genome copies per well in a 24-well plate.

To determine the ULBP1 and ULBP2 mRNA levels in THP-1 macrophages co-cultured with HBV-infected HepG2-NTCP cells, HepG2-NTCP cells were infected with 400,000 genome copies of HBV for 9 days and subsequently co-cultured with THP-1 macrophages using transwell co-culture system (pore size 3 μm) for 2 days. Total RNA of THP-1 macrophages were isolated with TRIZOL.

Primary human hepatocytes and human hepatic stellate cells were infected with 1,000,000 genome copies of HBV for 24 h in 24-well plates.