Hepatitis C virus infection suppresses hepatitis B virus replication via the RIG-I-like helicase pathway

Mechanisms of hepatitis B virus (HBV) reactivation after hepatitis C virus (HCV) elimination by direct-acting antiviral (DAA) treatment in HBV/HCV-co-infected patients remain unclear. We examined RIG-I-like helicase (RLH) pathway activation by HBV mono-infection, HCV mono-infection or HBV/HCV co-infection and interference between HBV and HCV in primary human hepatocytes. Interference between HBV and HCV and HBV reactivation after DAA treatment in humanized-liver mice were assessed. HCV infection activated RLH pathway, as evidenced by RIG-I, ISG15 and ISG56 expression induction; HBV caused only RIG-I induction in vitro. RLH activation was also found in HBV/HCV-co-infected cells, and HBV replication were suppressed in HBV/HCV-co-infected than in HBV-mono-infected cells. siRNA-mediated double knockdown of ISG15 and ISG56 increased HBV replication in HBV/HCV-co-infected cells. HCV infection activated RLH pathway and suppressed HBV replication in humanized-liver mice. Subsequent elimination of HCV by DAA administration downregulated RLH pathway and upregulated HBV replication in mice. RLH pathway was activated in livers of chronic hepatitis C patients compared to those of chronic hepatitis B or non-B, non-C patients. The RLH pathway activation was downregulated by HCV elimination. In conclusion, HCV infection activated RLH pathway and suppressed HBV replication in human hepatocytes. HCV elimination upregulated HBV replication, probably through RLH pathway downregulation.


HcV infection suppresses HBV replication and enhances the RLH system during HBV infection. The influence of HCV infection on HBV replication in PHHs was investigated. PHHs were incubated
with HBV with or without prior HCV infection ( Fig. 2A). Upon HBV infection, approximately two-thirds of the hepatocytes with or without prior HCV infection were positive for HBc. Approximately one-half of the HBV/ HCV-co-infected cells were positive for NS5A, and importantly, cells positive for both HBc and NS5A were identified, indicating that co-infection in cultured human hepatocytes truly occurred (Fig. 2B). mRNA expression levels of RIG-I, ISG15 and ISG56 in PHHs after HBV incubation with prior HCV infection were significantly higher than those in PHHs without prior HCV infection (Fig. 2C). HBV DNA and HBs antigen levels in the supernatant of PHHs (Fig. 2D) and pre-genomic RNA (pgRNA) levels in PHHs (Fig. 2E) after HBV incubation with prior HCV infection were significantly lower than those in PHHs without prior HCV infection. To examine the impact of ISG15 and ISG56 increase in HCV-infected PHHs on HBV replication, ISG15-and/or ISG56-knocked down PHHs were incubated with HBV with prior HCV infection. Although the expression levels of pgRNA in ISG15-knocked down PHHs or ISG56-knocked down PHHs did not differ from those in control PHHs, pgRNA levels in ISG15 and ISG56 double-knocked down PHHs were significantly higher than control PHHs or single-knocked down PHHs (Fig. 2F,G).

HcV infection inhibits HBV replication in chimeric mice infected with HBV and HcV. To develop
HBV/HCV-co-infected mice, three chimeric TK-NOG mice were infected with HCV and then infected with HBV. To develop HBV-infected mice, three chimeric mice were infected with HBV alone (Fig. 3A). The HBV/ HCV-co-infected mice showed significantly lower serum HBV DNA levels than the HBV-infected mice (Fig. 3B). The HBV/HCV-co-infected mice were treated with DAAs for 4 weeks. HCV RNA was not detected in any HBV/ HCV-co-infected mice from 2 weeks after the start of DAA therapy to sacrifice. The serum HBV DNA levels in the HBV/HCV-co-infected mice significantly increased after HCV elimination with DAAs (Fig. 3C). The pgRNA and cccDNA levels in the liver tissues of the HBV/HCV-co-infected mice were lower than those of the HBV-infected mice, and pgRNA and cccDNA levels increased after the elimination of HCV with DAAs (Fig. 3D). The mRNA levels of RIG-I, ISG15 and ISG56 in the hepatocytes of the HBV/HCV-co-infected mice before treatment were higher than those of mice infected with the HBV-infected mice, and the RIG-I, ISG15 and ISG56 mRNA levels decreased after HCV elimination with DAAs (Fig. 3E).
the RLH pathway was activated in HcV-infected livers of cHc patients. The mRNA expression levels of RIG-I, ISG15 and ISG56 in human liver biopsy samples from CHB, CHC, and non-B, non-C patients were measured by real-time reverse transcription PCR (RT-PCR) ( Table 1). In human liver biopsy samples from CHC patients, the mRNA expression levels of RIG-I, ISG15 and ISG56 were higher than those in liver biopsy samples from CHB patients and non-B, non-C patients (Fig. 4A). We further examined using pared liver biopsy samples from CHC patients before and after HCV elimination by treatment (Table 2). Liver biopsy samples before HCV elimination were obtained at the start of treatment with DAAs. Liver biopsy samples after HCV elimination were obtained 48 weeks after the end of treatment (EOT) with DAAs. RIG-I, ISG15 and ISG56 levels were decreased by HCV elimination (Fig. 4B).

Discussion
Although patients are sometimes co-infected with HBV and HCV, the interaction between HBV and HCV infection has been rarely examined due to the lack of a suitable co-infection system. There has been no report examining the interaction between HBV and HCV infection in normal hepatocytes. In the present study, we revealed that HBV and HCV could co-infect PHHs and addressed the interaction between HBV and HCV using these PHHs. This study is the first to examine the impact of HCV infection on HBV replication levels in normal hepatocytes in vitro and in vivo. www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ It has been reported that RIG-I and melanoma differentiation-associated gene-5 (MDA-5) in the cytoplasm and endosome are involved in the recognition of HCV RNA and general viral RNA 17 . RIG-I has also been reported to recognize HCV RNA in the early phase of infection and activate downstream host innate immunity 18 . In the present study, the mRNA levels of RIG-I, ISG15, and ISG56 increased during HCV infection in vitro and in mouse experiments. They are also increased in HCV-infected livers of CHC patients. Interferon-stimulated genes (ISGs) have broad antiviral activity. ISG15 is a critical ISG with antiviral activity against DNA and RNA viruses 19 . ISG56 limits HBV replication by regulating posttranscriptional steps 20 . In the present study, suppression of both ISG15 and ISG56, but not suppression of each alone, increased pgRNA levels, suggesting that ISG15 and ISG56 collaboratively suppress pgRNA levels. It is thought that in patients with HBV/HCV co-infection, HBV replication www.nature.com/scientificreports www.nature.com/scientificreports/ is strongly suppressed by HCV-induced upregulation of ISGs, including ISG15 and ISG56. Consequently, HCV elimination with DAAs decreases ISGs expression leading to enhancement of HBV replication.
HBV reactivation is caused by anticancer drugs or immunosuppressants 15 . Anticancer drugs in the clinic that cause HBV reactivation mainly fall into two classes: traditional cytotoxic chemotherapy drugs that mainly suppress T cells and biological agents related to B cells such as rituximab 21   www.nature.com/scientificreports www.nature.com/scientificreports/ not have cytotoxic effects on B cells or T cells. Although they also do not have immunosuppressive effects, DAAs do cause HBV reactivation. In addition, in the present study, treatment with DAAs resulted in HBV reactivation in immunodeficient TK-NOG mice, indicating that HBV reactivation proceeded without immune cells.
In conclusion, HCV infection suppressed HBV replication. At the same time, HCV enhanced the RLH pathway. In chimeric TK-NOG mice, elimination of HCV reversed HBV suppression. At the same time, elimination of HCV also reversed the enhancement of the RLH pathway. These data suggest that the enhancement of the RLH pathway induced by HCV infection during HBV/HCV co-infection contributes to inhibition of HBV replication. The reversal of RLH pathway enhancement by the elimination of HCV using DAAs contributes to HBV reactivation in HBV/HCV-co-infected patients after treatment with DAAs.

Materials and Methods
cell culture. PHHs were collected from chimeric mice with a chimeric liver rates of 50% or more using the two-step collagenase-pronase liver perfusion method in the same manner as previously reported 24 . All experiments began within 10 days of collecting PHHs from chimeric mice. The PHHs were seeded on 12-or 24-well collagen type I-coated microplates (AGC TECHNO GLASS CO., LTD., Shizuoka, Japan). Culture medium containing 2% dimethyl sulfoxide (DMSO) (NACALAI TESQUE, INC., Kyoto, Japan) was changed on the indicated days.
The culture supernatant of HepG2.2.15 cells was used as HBV inoculum in vitro. HepG2.2.15 cells have an integrated HBV genome, and HBV genotype D is present in the culture supernatant 25 . The culture supernatant of HepG2.2.15 cells was collected every 3 days, filtered through a 0.45-μm filter (Merck Millipore, Burlington, MA, USA), concentrated 200 times, and used as the inoculum. PHHs were incubated with culture medium containing the HBV inoculum and 4% polyethylene glycol 8000 (Promega, Fitchburg, WI, USA) for 24 hours. After incubation with the HBV inoculum, the PHHs were washed three times with phosphate-buffered salts (PBS) containing 2% DMSO, and the culture medium was changed.
The culture supernatant of Huh7 cells inoculated with JFH-1, which belongs to HCV genotype 2a 26 , was used as the HCV inoculum in vitro. Culture supernatant from Huh7 cells inoculated with JFH-1 was collected every 3 days, filtered through a 0.45-μm filter, and used as the inoculum. PHHs were incubated with culture medium containing the HCV inoculum for 72 hours, after which the PHHs were washed three times with PBS containing 2% DMSO, and the culture medium was changed.
Humanized-liver chimeric tK-noG mice. Humanized-liver chimeric TK-NOG mice were prepared as previously described 27 . The human hepatocyte chimeric rate is correlated with serum human albumin levels, and the estimated human hepatocyte chimeric rate can be calculated from serum human albumin levels 27  HBV or HcV inoculation of mice. CHB patient serum (genotype C, 8.5 log IU/ml) and CHC patient serum (genotype C, 6.8 log IU/ml) were used for mouse experiments under the approval of the Institutional Review Board for Clinical Research at Osaka University Hospital (12050, 13048). Written informed consent was obtained from these patients for experimental use of their sera. The CHB patient serum used in the experiment was diluted to 7.5 log IU/ml. One hundred μl of diluted CHB patient serum or CHC patient serum was injected intravenously into chimeric TK-NOG mice with a chimeric rate of 50% or more as previously described 28 . Mouse blood samples were collected from the external jugular vein. The protocols involving animal experiments were approved by the Animal Care and Use Committee of Osaka University Medical School.

co-infection of chimeric tK-noG mice with HBV and HcV and elimination of HcV in HBV/
HcV-inoculated chimeric mice by treatment with DAAs. To develop HBV/HCV-co-infected mice, three humanized-liver chimeric TK-NOG mice were infected with HBV 4 weeks after HCV infection. www.nature.com/scientificreports www.nature.com/scientificreports/ To develop HBV-infected mice, three humanized-liver chimeric TK-NOG mice were infected with HBV. The HBV/HCV-co-infected mice were orally administered with DAAs (40 mg/kg asunaprevir +30 mg/kg daclatasvir +20 mg/kg beclabuvir) once a day for 4 weeks. Asunaprevir, daclatasvir, and beclabuvir were provided by Bristol-Myers Squibb (New York, NY, USA). Eight weeks after HBV infection, hepatectomy was performed in HBV-infected mice and HBV/HCV-co-infected mice to collect liver samples. Sixteen weeks after HBV inoculation, the chimeric HBV-infected mice and HBV/HCV-co-infected mice were sacrificed. Five humanized-liver chimeric mice without HBV or HCV infection were sacrificed as control mice.
Liver biopsy samples. Liver biopsy samples were obtained from CHB, CHC and non-B, non-C patients with approval by the Institutional Review Board for Clinical Research at Osaka University Hospital (15267). Liver biopsy samples were obtained from patients before and after HCV elimination by DAA therapy with approval by the Institutional Review Board for Clinical Research at Sendai Kousei Hospital . Informed consent was obtained in writing from these patients for experimental use of their liver biopsy samples. Residual liver biopsy samples were used for the experiment. Transfection of siRNA against ISG15 and ISG56. Two siRNAs against human ISG15(s18524) and ISG56 (s7151), and the negative controls (#4390843) were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Transfections were performed using Lipofectamine RNAiMAX (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. immunofluorescent staining. Cells were grown on chambered coverglasses (Matsunami Glass Ind., Ltd., Osaka, Japan), washed with PBS, and then fixed with ice-cold acetone-methanol. Cells were washed again (3× PBS) and blocked in PBS/0.2% BSA. HBc antigen monoclonal antibodies were generated by immunizing of BALB/c mice with recombinant HBc protein synthesized in a wheat cell-free protein production system, as previously described 29 . HBc antigen was probed by using a mouse monoclonal antibody against HBc antigen (clone 7B2, culture supernatant of the hybridoma) 30 diluted 100-fold in PBS (1 hour at room temperature). HCV NS5A was stained by using a rabbit monoclonal antibody against NS5A diluted 100-fold in PBS (1 hour at room temperature). As a secondary antibody, goat anti-mouse antibody labelled with Alexa Fluor 594 (Cell Signaling Technology, Danvers, MA, USA) or goat anti-rabbit antibody labelled with Alexa Fluor 488 (Cell Signaling Technology, Danvers, MA, USA) diluted 500-fold in PBS was used. Cell nuclei were stained using DAPI. Stained cells were analysed using a fluorescence microscope (Invitrogen EVOS FL Auto 2 Imaging System; Thermo Fisher Scientific, Waltham, MA, USA). Statistical analysis. The data are presented as the means ± standard error. Comparisons between two groups in the in vitro studies were performed by an unpaired two-sided t test. Comparisons between two groups in the in vivo studies were performed by an unpaired two-sided t test. For the in vitro, in vivo and liver biopsy samples from the patients of chronic liver disease, analysis of variance (ANOVA) was performed to detect an overall difference among multiple groups, followed by the Tukey-Kramer test. Comparisons between before and after HCV elimination in the paired liver biopsy samples were performed by Wilcoxon signed-rank test. A value of P < 0.05 was used to indicate statistical significance.