Potential conflict of interest: Nothing to report.
Supported by the National Natural Science Foundation of China (Grant 30900062), the National Basic Research Program of China (Grant 2012CB519005), the “Twelfth Five-Year” National Science and Technology Major Project of China (Grant 2012ZX10002007-001); the International Science & Technology Cooperation Program of China (Grant 2011DFA31030); and the German Research Foundation (SFB/Transregio TRR60).
Treatment with exogenous interferon (IFN)-α is not effective in the majority of patients with chronic hepatitis B virus (HBV) infection. Recent evidence suggests that HBV has evolved strategies to block the nuclear translocation of signal transducer and activator of transcription (STAT) 1 to limit IFN-α–induced cellular antiviral responses. However, it remains unclear whether STAT1 translocation is impaired in chronic hepatitis B patients and what mechanisms are involved. Here we report that the expression of HBV polymerase (Pol) in human hepatic cell lines inhibited induction of IFN-stimulated genes and resulted in a weakened antiviral activity of IFN-α. Ectopic expression of Pol suppressed IFN-α–induced STAT1 serine 727 phosphorylation and STAT1/2 nuclear accumulation, whereas STAT1 tyrosine 701 phosphorylation, and STAT1-STAT2 heterodimer formation were not affected. Further studies demonstrated that Pol interacted with the catalytic domain of protein kinase C-δ (PKC-δ) and perturbed PKC-δ phosphorylation and its association with STAT1, which resulted in the suppression of STAT1 Ser727 phosphorylation. Moreover, Pol was found to interfere with nuclear transportation of STAT1/2 by competitively binding to the region of importin-α5 required for STAT1/2 recruitment. Truncation analysis suggested that the terminal protein and RNase H domains of Pol were able to bind to PKC-δ and importin-α5, respectively, and were responsible for the inhibition of IFN-α signaling. More importantly, the inhibition of STAT1 and PKC-δ phosphorylation were confirmed in a hydrodynamic-based HBV mouse model, and the blockage of IFN-α–induced STAT1/2 nuclear translocation was observed in HBV-infected cells from liver biopsies of chronic HBV patients. Conclusions: These results demonstrate a role for Pol in HBV-mediated antagonization of IFN-α signaling and provide a possible molecular mechanism by which HBV resists the IFN therapy and maintains its persistence. (HEPATOLOGY 2013;)
Chronic hepatitis B (CHB) caused by hepatitis B virus (HBV) is a serious health problem worldwide. The mechanism by which chronic infection is established and maintained is unknown but is thought to be due, in part, to a suppressed host immune response. One key component of the host antiviral responses is the interferon (IFN) system. Viral infection of the host initiates the synthesis of type I IFNs, which consist predominantly of IFN-α and IFN-β (IFN-α/β). By binding to type I IFN receptors, IFN-α/β triggers the oligomerization and tyrosine phosphorylation of the two tyrosine kinases of the Janus family, Janus kinase 1 and tyrosine kinase 2, which in turn phosphorylate a single tyrosine residue of signal transducer and activator of transcription (STAT) 1 and 2. The activated STAT1/2 heterodimerize with interferon regulatory factor 9 (IRF9) to form the ISGF3 transcription factor complexes and then translocate into the nucleus, where they bind to the interferon-stimulated response element (ISRE) in the promoter of interferon-stimulated genes (ISGs) to initiate transcription of ISGs.1 IFN-α has been shown to inhibit HBV replication in a variety of systems. However, about 70% of CHB patients respond poorly to exogenous IFN-α treatment.2, 3
Increasing evidence suggests that HBV has developed strategies to counteract the type I IFN system, which may contribute to the ineffectiveness of IFN-α therapy.4, 5 We have shown that HBV polymerase (Pol) is able to inhibit IFN-α–induced MyD88 induction and nuclear translocation of STAT1.6 Consistently, it could be demonstrated in a chimeric mice model that HBV infection reduced IFN-α–mediated ISG production and STAT1 translocation.7 However, it is still unknown whether the translocation of STAT1 is impaired in chronic hepatitis B patients as well, and the molecular mechanism by which Pol interferes with the IFN responses remains unclear.
In this study, we used cell and mouse models to gain a detailed understanding of how HBV and Pol interferes with IFN-α–induced STAT activation. Furthermore, liver biopsies of CHB patients were used to obtain more information on the blockage of IFN-α–induced STAT nuclear translocation by HBV.
The human hepatic cell lines Huh7 and HepG2 and 293T cells were obtained from the Cell Bank of the Chinese Academy of Sciences. The HepG2-derived HBV-producing stable cell lines HepG2.215 and HepAD38 were kindly provided by Yumei Wen. The cells were routinely cultured as described.8 Additionally, HepAD38 cells were treated with or without 1 μg/mL doxycycline (Dox) to regulate HBV pregenomic RNA transcription.9
Plasmid and Regents.
Plasmids used for transfection are listed in Supporting Table 1. All plasmids were prepared using Endo-Free Plasmid Kits (Omega). Human recombinant IFN-α (Calbiochem) was used at 500 U/mL unless specified otherwise. Rottlerin was obtained from Sigma. The small interfering RNAs (siRNAs) targeting Pol (Supporting Table 2) and an unrelated control siRNA were purchased from Ribobio (China).
Liver biopsies were collected from CHB patients in Shanghai Public Health Clinical Center with informed consent and the approval of the institutional ethics committee. The liver biopsies were obtained percutaneously with a Menghini needle. A part of the biopsy was used for routine histopathological diagnosis, and the remaining fresh tissue was incubated with 500 U/mL IFN-α for 0.5 hours at 37°C and then fixed in formaldehyde, embedded in paraffin, and sectioned for immunostaining. The clinical characteristics of the patients are shown in Supporting Table 3.
Co-immunoprecipitation, Glutathione S-transferase Pull-down, and Immunoblotting.
Co-IP and glutathione S-transferase (GST) pull-down were performed as described10 with minor modifications. Anti-Flag M2 agarose affinity beads (Sigma) were used to precipitate Flag-tagged proteins. Native polyacrylamide gel electrophoresis was performed8 to detect the STAT1/2 heterodimer. Immunoblotting was performed with the appropriate antibodies (Supporting Table 4) according to standard protocols. Results are representative of at least three experiments.
RNA Isolation and Real-Time Polymerase Chain Reaction.
Total RNA was extracted using the RNAsimple Total RNA Kit (TianGen) and reverse-transcribed using ReverTra Ace qPCR RT Kit (Toyobo). The complementary DNA samples were subjected to real-time polymerase chain reaction (PCR) using primers specific for the genes listed in Supporting Table 5. For comparisons, the expression of each gene was normalized to that of glyceraldehyde 3-phosphate dehydrogenase. Each PCR was performed in duplicate using Thunderbird SYBR qPCR Mix (Toyobo) in a StepOne Real-Time PCR System (ABI). The results are representative of three independent experiments.
Immunofluorescence was performed as described.8 Paraffin sections of liver biopsies were dewaxed, rehydrated, and microwaved before incubation with the primary antibodies.
Subcellular Fractionation Assay.
Cytoplasm and nuclear fractions were obtained as described.6 Extracts were analyzed by immunoblotting. Antibodies against β-tubulin and lamin A/C were used as cytoplasmic and nuclear markers, respectively.
The hydrodynamic-based HBV mouse model was as described by Huang et al.11 The methods are described in detail in the Supporting Information. All mouse experiments were conducted in accordance with the guide for the care and use of medical laboratory animals (Ministry of Health, China).
The data are presented as the mean ± SD. Statistical analyses were performed via Student t tests for comparison between two groups and one-way analysis of variance followed by Bonferroni tests for multiple comparisons using GraphPad Prism software. A value of P < 0.05 was considered statistically significant.
Inhibition of IFN-α–Induced Cellular Antiviral Responses by HBV Occurs Through Expression of Pol.
It has been shown that HBV inhibits IFN-α-–mediated responses, and Pol may be responsible for the inhibition.5, 6 We first confirmed that HBV and Pol are able to interfere with IFN-α–induced ISRE-dependent gene expression and ISG induction in human hepatic cell lines (Supporting Result 1). To ensure that the inhibition of IFN-α–induced ISRE-dependent gene expression by Pol is not due to a nonspecific effect of overexpression, we used a viral replicon, in which viral replication is initiated under its own promoter after being transfected into cells. As shown in Fig. 1A, cells transfected with the HBV replicon resulted in an impaired ISRE activation, while the Pol-null-HBV construct-transfected cells exhibited a comparable level of ISRE-driven luciferase expression to that of control cells, implying that the Pol-mediated suppression of cellular response to IFN-α occurred at a physiologically relevant expression level of Pol. HepAD38 cell line that replicates HBV under Dox-off control (Supporting Fig. 2B,C) was employed to further substantiate the effect of Pol on IFN-α–stimulated cellular responses. The data showed that the expression of Pol (Dox-free) significantly reduced IFN-α–mediated ISRE activation (Supporting Fig. 2D) and protein kinase R production (Fig. 1B). Moreover, knockdown of Pol expression in HepG2.215 cells (Supporting Fig. 3) restored IFN-induced ISRE-dependent gene expression (Fig. 1C).
To assess the biological significance of the above observations, we compared cells expressing or not expressing Pol for their IFN sensitivity. As shown in Fig. 1D and Supporting Fig. 1C, the antiviral activity stimulated by IFN-α against vesicular stomatitis virus (VSV) was much lower in Pol-transfected cells and HepG2.215 cells than in control cells. Similar results were observed when HCV-Jc1-Gluc was used for viral challenge (Supporting Fig. 2F). Taken together, these results implicate a role for Pol in mediating the inhibitory effects of HBV on IFN-α–induced antiviral responses.
STAT1 Serine Phosphorylation But Not STAT1-STAT2 Heterodimerization Is Inhibited by Pol.
We next determined the effects of HBV and Pol on the expression of IFN-α signaling–related molecules. HepG2 and HepG2.215 cells treated with IFN-α for time points ranging from 30 minutes to 24 hours were analyzed for protein levels and phosphorylation (Fig. 2A). Although there was no significant difference in the basal levels of STAT1/2, IRF9, IFNAR1/2, Janus kinase 1, and tyrosine kinase 2 between the two cell lines, HepG2 cells showed robust up-regulation of STAT1, STAT2, and IRF9 upon IFN-α stimulation compared with HepG2.215 cells. Furthermore, the serine 727 phosphorylation of STAT1 was severely impaired in HBV-expressing cells, particularly in the later periods after IFN-α treatment. Increased induction of STATs and IRF9 was also observed after IFN-α treatment in Pol-expressing Huh7 cells but was much weaker than that observed in control cells (Fig. 2B). The levels of STAT1 Ser727 phosphorylation were clearly repressed by transfection of Huh7 cells with Pol; however, tyrosine phosphorylation of STAT1/2 was not affected.
To further investigate the effect of Pol on the tyrosine phosphorylation-induced STAT1-STAT2 heterodimerization, we performed co-immunoprecipitation (co-IP) experiments in Huh7 cells transfected with increasing amounts of Pol (Fig. 2C) and in Dox-regulated HepAD38 cells (Fig. 2D). The results showed that STAT1-STAT2 interaction in response to IFN-α was consistently observed in cells with or without Pol. Meanwhile, Flag-Pol was not detected in the immune complexes precipitated with anti-STAT1 or anti-STAT2 Abs, indicating no direct interaction between Pol and activated STAT1/2. Moreover, there was not much difference in IFN-α–induced heterodimer formation between cells expressing Pol and control cells (Fig. 2E,F), indicating that Pol does not affect the IFN-α–stimulated STAT1-STAT2 heterodimerization.
Pol Reduces STAT1 Ser727 Phosphorylation by Inhibition of PKC-δ.
IFN-α-induced phosphorylation of the serine residue at position 727 (Ser727) of STATs contributes critically to their transcriptional activity.12, 13 Although still controversial, PKC-δ, p38 and ERK have been reported to function as kinases that regulate Ser727 phosphorylation.14, 15 To elucidate the mechanism by which Pol interferes with STAT1 Ser727 phosphorylation, we examined the effect of Pol on IFN-α–induced phosphorylation of PKC-δ, p38 and ERK (Fig 3A). The results showed that Pol only inhibited PKC-δ but not p38 or ERK phosphorylation in IFN-α–stimulated Huh7 cells. Rottlerin, a selective inhibitor of PKC-δ, was used to verify the role of PKC-δ in Ser727 phosphorylation of STATs (Fig. 3B), and the data demonstrate that PKC-δ is specifically required for the Ser727 phosphorylation, but not for STAT tyrosine phosphorylation. IFN-α–stimulated PKC-δ phosphorylation was also found to be impaired in HepG2.215 cells compared with that in HepG2 cells (Fig 3C), but was restored by Pol siRNA transfection (Supporting Fig. 5A). In addition, we investigated whether Pol inhibits IFN-α signaling by regulating the level of STAT3, as it was reported to be a negative regulator of the type I IFN response.16 Little difference in the basal expression level and IFN-α–induced tyrosine phosphorylation of STAT3 was observed between the cells with or without Pol; however, PKC-δ–dependent Ser727 phosphorylation of STAT3 was inhibited by Pol in a dose-dependent manner (Fig. 3D). Furthermore, less STAT1 was coprecipitated with PKC-δ from lysates of Pol-expressing IFN-α–treated cells (Fig. 3E), and Pol was found to interact with the catalytic domain of PKC-δ (Fig. 3F and Supporting Fig. 5B). These results suggest that the perturbation of IFN-α–induced PKC-δ phosphorylation and association between PKC-δ and STAT1 may contribute to the suppression of STAT1 Ser727 phosphorylation by Pol.
Suppression of IFN-α–Induced STAT1/2 Nuclear Translocation by Pol.
The nuclear translocation of STAT1 and STAT2 is a key process in IFN-α transduction signaling. We and others have shown that STAT1 translocation is impaired in Pol-expressing and HBV-infected hepatocytes.6, 7 We further examined whether Pol interferes with IFN-α–induced STAT2 nuclear accumulation. In the absence of IFN-α, STAT2 was predominantly localized in the cytoplasm, whereas STAT1 was found in both the cytoplasm and nucleus; upon IFN-α stimulation, strong nuclear accumulation of both STAT1 and STAT2 was observed, but only in cells without Pol expression (Fig 4A). Furthermore, the protein levels of STAT1/2 in the cytoplasmic and nuclear fractions of Dox-treated or Dox-free HepAD38 cells were determined via immunoblotting. As shown in Fig. 4B, the accumulation of STAT1/2 in the nucleus following IFN-α treatment was significantly detained in Dox-free (Pol-expressing) cells. Impaired nuclear translocation of STAT1/2 was also observed in HepG2.215 cells compared with HepG2 cells (Supporting Fig. 6).
Pol Prevents IFN-α-Induced STAT1/2 Nuclear Translocation Through Competitive Binding of Importin-α5.
Importin-α5 (also known as karyopherin α1), a nuclear localization signal receptor, has been shown to specifically interact with the STAT1-STAT2 heterodimer and to be responsible for the nuclear transport of the complex.17, 18 We thus investigated whether Pol interferes with the interaction between importin-α5 and activated STATs. As shown in Fig. 5A, STAT1 and STAT2 were clearly detected in the importin-α5 immunoprecipitation complex when IFN-α was added, however, the STATs decreased in a dose-dependent manner, with increased expression of Pol. A similar inhibition was observed using the HepAD38 model (Supporting Fig. 5C). Moreover, impaired colocalization of STAT2 and importin-α5 was observed in HepG2.215 cells compared with that in HepG2 cells by immunofluorescence (Fig. 5B).
To investigate whether Pol directly interacts with importin-α5, GST pull-down assays were conducted, and Flag-Pol was pulled down by GST–importin-α5 (Fig. 5C). Furthermore, colocalization of Pol and importin-α5 in the cytoplasm was detected (Fig. 6D). The C-terminal arm repeats 8 and 9 of importin-α5 have been reported to form a unique binding site for activated STAT1-STAT2.19 Thus, we aimed to determine whether Pol binds to this region of importin-α5. Hemagglutinin-tagged full-length importin-α5 and several truncated constructs were transfected along with Flag-Pol, and the co-IP results showed that Pol was able to coprecipitate with all the truncations except importin-α5-1-406 (Fig. 5D), implying that Pol binds to importin-α5 through a region (407-504) that is also required for the importin-α5-STAT1/2 interaction. We conclude that Pol weakens the importin-α5-STAT1/2 interaction through competitive binding of importin-α5, providing an explanation for how Pol inhibits IFN-α–induced STAT1/2 nuclear translocation.
The Terminal Protein and RNase H Domains of Pol Bind to PKC-δ and Importin-α5, Respectively, and Are Required for Inhibiting IFN-α–Induced STATs Activation.
The HBV Pol can be divided into four domains: terminal protein (TP), spacer region (SP), reverse transcriptase domain (RT), and RNase H (RH) domain.20 To determine the region(s) of Pol required for blocking IFN-α signaling, a series of N- and C-terminal truncations of Pol were constructed. All of the full-length and truncated Pol showed inhibitory effects on ISRE promoter activity compared with the control, although the degree of inhibition varied (Fig. 6A). We then examined IFN-α–induced Ser727 phosphorylation of STAT1 and nuclear accumulation of STAT1/2 in cells transfected with TP-SP domains or RT-RH domains of Pol. Interestingly, the TP-SP construct was sufficient to inhibit STAT1 Ser727 phosphorylation (Supporting Fig. 7A,B), whereas the RT-RH construct prevented nuclear accumulation of STAT1/2 (Supporting Fig. 7C,D). Moreover, RT-RH interfered with the IFN-α–induced importin-α5–STAT2 interaction, similar to the full-length Pol (Fig. 6B). RH, but not RT, coprecipitated with importin-α5 (Fig. 6C), and no colocalization was observed between Pol-ΔRH and importin-α5 (Fig. 6D). Besides, TP coprecipitated with PKC-δ (Supporting Fig. 7E) and had an inhibitory effect on IFN-α–induced STAT1 Ser727 phosphorylation compared with SP and Pol-ΔTP (Fig. 6E). In addition, IFN-α was found to be less effective to induce the antiviral status in HepG2-TP (stably expressing TP) than in control cells (Supporting Fig. 8). Additional truncated mutations of Pol were made that included both of the predicated domains responsible for inhibiting or not inhibiting IFN-α signaling. As expected, the TP-RH fusion protein exhibited a similar inhibitory effect compared with full-length Pol, whereas SP-RT was not inhibitory (Supporting Fig. 7F). These results indicate a role for TP and RH in the Pol-mediated modulation of IFN-α–induced STAT activation.
In Vivo Analysis of the Inhibition of IFN-α Signaling by HBV and Pol.
We further substantiated the effect of HBV and Pol on IFN-α–mediated response in a mouse model. C57BL/6 mice were hydrodynamically injected with plasmids expressing HBV, Pol, or the control vector. As shown in Fig. 7A, the transcription levels of Mx1, STAT1, MyD88, and TAP-1 (transporter associated with antigen presentation) were significantly increased in control mice after mouse interferon (mIFN)-α treatment; however, mIFN-α induced much less ISGs in the livers of HBV- and Pol-expressing mice. The liver samples of mice were also analyzed by immunoblotting. The data showed that mIFN-α–induced phosphorylation of STAT1-Ser727 and PKC-δ was significantly decreased in livers expressing HBV and Pol (Fig. 7B). Notably, the level of tyrosine-phosphorylated STAT1 was unaffected in vitro by HBV but was slightly lower in livers from mice injected with HBV and Pol compared with controls, implying some distinct in vivo mechanisms. Furthermore, mice injected with pAAV-HBV1.2 were treated with mIFN-α or mock-treated for 1 week. The levels of serum hepatitis B surface antigen and HBV DNA at 4 and 11 days posttransfection were determined and showed no significant difference between the two groups (Supporting Fig. 9F), indicating high resistance against the actions of IFN by HBV in our experimental settings.
The effect of HBV on the IFN-α–induced nuclear translocation of STAT1/2 was further tested in liver biopsies from CHB patients (Fig 8). Nuclear-localized STAT1/2 proteins were detected not only in hepatocytes but also in nonparenchymal resident liver cells. However, cells infected with HBV (using hepatitis B surface antigen as a positive marker for HBV infection) showed disrupted nuclear accumulation of STAT1/2. Taken together, these results strongly support the findings described in vitro.
IFN-α is used to treat patients with chronic HBV infection but has a poor response rate. The exact reason for the ineffectiveness has not been fully elucidated. Our previous studies suggest that HBV Pol inhibits IFN-α-induced nuclear translocation of STAT1 and consequently impairs the promoter activity of MyD88,6 which encodes a protein that inhibits HBV replication.21 Here we further demonstrated that Pol suppresses IFN-induced STAT1/2 nuclear translocation and STAT1 Ser727 phosphorylation via competitive binding to importin-α5 and inhibition of PKC-δ activation respectively (summarized in Supporting Fig. 10). More importantly, the inhibition of STAT1 and PKC-δ phosphorylation were confirmed in an HBV mouse model based on hydrodynamic injection, and the blockage of IFN-α–induced STAT1/2 nuclear translocation was for the first time observed in HBV-infected hepatocytes from liver biopsies of CHB patients. These results provide a better understanding of how HBV resists IFN-α treatment.
We showed that Pol had little effect on IFN-α–induced STAT1/2 tyrosine phosphorylation and subsequent heterodimerization; by contrast, the nuclear transportation of the heterodimer and the STAT1 Ser727 phosphorylation were inhibited in the presence of HBV and Pol. Considering that tyrosine phosphorylation–induced STAT1/2 complex formation is essential for nuclear translocation,1 whereas Ser727 phosphorylation is not required for interaction between STATs but is crucial for full transcriptional activation,12, 13 it seems that HBV has evolved smart strategies to inhibit the IFN-α signaling at two independent steps, either by blocking STAT1/2 translocation to the nucleus or by suppressing the transcriptional activity of both cytoplasmic and nuclear forms of STAT1. However, we cannot exclude the possibility that the lack of nuclear accumulation of STAT1 is related to the reduction of the transcriptional activity of STAT1. We also found that IFN-α induced rapid tyrosine phosphorylation of STAT1, whereas serine phosphorylation occurred much later (Fig. 2A,B) and was blocked by pretreatment of cells with a specific inhibitor of PKC-δ. These results support the concept that the process of IFN-α–induced Janus kinase–STAT signaling is highly sophisticated and regulated through various mechanisms. Indeed, several posttranscriptional modifications in addition to phosphorylation—including acetylation, O-glycosylation, and sumoylation—have been suggested to be involved in the regulation of STATs activity.22 It is thus necessary to further investigate the effect of Pol on these pathways so as to obtain a more accurate picture. Two reports showed that HBV suppresses IFN-α signaling by inhibiting STAT1 methylation,23, 24 however, the potential role for STAT1 methylation remains controversial.25, 26 Since the Pol-targeted importin-α5 is important in nuclear import of certain molecules and PKC-δ plays a fundamental role in growth regulation by targeting specific substrates27 and was recently reported to be involved in the IFN-α–mediated suppression of HBV enhancer II activation,28 it is also important to consider the possibility that Pol may cause disturbances in these processes.
By ectopic expression and knockdown experiments, we determined that Pol is responsible for the HBV-mediated inhibition of IFN-α signaling. In contrast to HBV structural proteins like core and HBs, Pol is believed to be produced at a much lower level during viral replication. A recent paper showed that polyinosinic:polycytidylic acid-induced IFN-α/β-dependent STAT3 phosphorylation was inhibited when the viral load was high.29 As our data showed that Pol inhibits polyinosinic:polycytidylic acid-induced IFN production8 and IFN-α–induced serine phosphorylation of STAT3 in a dose-dependent manner (Fig. 3D), we hypothesize that the physiological levels of Pol are correlated with the viral load and that HBV can only efficiently inhibit the IFN system when the viral replication level is high. This scenario is also supported by the clinical observation that patients with high HBV DNA levels are mostly nonresponders to IFN-α therapy.2, 3 In addition, TP and RH domains were found to exhibit similar inhibitory effects compared with full-length Pol (Fig. 6). Several studies have demonstrated the in vivo expression of viral proteins encoded by HBV spliced RNAs, which contain domains derived from the open reading frame of the Pol gene.30 The function of these proteins remains obscure. It is thus to consider the hypothesis that such splice variants may represent a source of the proteins containing TP or RH domains and contribute to the suppression of the host antiviral responses.
We used a hydrodynamic-based mouse model to substantiate the in vitro findings. Intriguingly, the basal levels of Mx1 were significantly higher in HBV-transfected livers compared with those in control mice, and ISG induction was strongly inhibited by HBV in the mouse liver (Fig. 7A). Although these results appear contrary to previous reports indicating that HBV is a “stealth” virus early in the infection31 and the findings obtained in vitro that HBV inhibits ISGs expression in hepatic cells by only two- to four-fold (Supporting Fig. 1), they are similar to the findings obtained in HBV-infected chimeric mice.7 It should be noted that the liver is an organ with predominant innate immunity and consists not only of hepatocytes (parenchymal cells) but also of nonparenchymal cells, including Kupffer cells and sinusoidal endothelial cells.32 Moreover, Pol and X have been shown to be able to counteract the pattern recognition receptor signaling in hepatocytes.8, 33, 34 Therefore, we speculated that nonparenchymal liver cells may contribute to intrahepatic ISG expression during HBV infection, though the mechanisms involved are yet to be determined. In addition, although HBsAg did not influence IFN signaling in vitro, we cannot neglect the role of it and hepatitis B e antigen in contributing to the IFN response defect in vivo, as they were reported to suppress Toll-like receptor–induced IFN-β and ISG induction in both the parenchymal and nonparenchymal liver cells.35 In addition, the data obtained from liver biopsies revealed that the nuclear translocation of STAT1/2 was impaired in HBV-positive cells, but was still intact in many neighboring cells without HBV infection. Therefore, comprehensive analysis of the interaction between HBV and the IFN system in both hepatocytes and nonparenchymal cells is necessary, and it will be interesting to compare the STAT activation pattern in different types of liver cells between the IFN responders and nonresponders to further explore factors affecting response to IFN-α therapy.
Viruses have evolved various strategies to circumvent the IFN response, thus allowing them to escape the host defenses.36 HCV, for example, impairs type I IFN response by blocking different levels of IFN-α signal pathway via its core, NS3 and NS5A proteins. For HBV, we propose a two-part mechanism by which Pol inhibits the IFN-α–stimulated antiviral responses. These findings, together with our previous finding that Pol can inhibit the type I IFN induction,8 suggest that Pol is a multifunctional IFN antagonist. This knowledge not only helps us understand the mechanisms of resistance of HBV-infected patients to IFN treatment, it also clarifies the role of Pol in HBV persistence. Once viral replication reaches high levels, Pol may exert its anti-IFN activities to ensure the survival of the virus.
Notably, the sensitivity to IFN-α differs between HBV and HCV. It was reported that IFN treatment resulted in a rapid reduction in HCV but a moderate reduction in HBV.37 HBV seems to have a stronger ability to interfere with the IFN antiviral actions compared with HCV. However, the inhibitory effect of HBV on the IFN-α–mediated ISG induction was found to be modest.37 HBV, as a hepatotropic DNA virus, may have low sensitivity to IFN-induced ISGs and counteract the IFN actions at different levels, including the IFN signal transduction and antiviral functions of ISG products. Nevertheless, future studies are required to fully understand HBV resistance to IFN-α and precisely define the mechanisms by which IFN-α inhibits HBV replication.
We thank Pei-Jer Chen (National Taiwan University College of Medicine) for his generous gift of the pAAV/HBV1.2 plasmid. We also thank Zekun Wang for preparing several of the truncations used in this study; Xiuhua Peng, Yinghui Liu, Shiyan Yu, Junyu Lin, Bisheng Shi, Wuhui Song, Fei Zhang, Dong Zeng, Yanling Feng, Wei Lu, Yanbing Wang, Huanping Ding, and Jiangxia Liu for technical assistance; and Jianhua Li for insightful criticism and suggestions.