High‐level expression of STING restricts susceptibility to HBV by mediating type III IFN induction

Abstract Hepatitis B virus (HBV) is a hepatotropic DNA virus causing hepatic diseases such as chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma. To study HBV, human hepatoma HepG2 cells are currently used as an HBV infectious cell culture model worldwide. HepG2 cells exhibit susceptibility to HBV by exogenously expressing sodium taurocholate cotransporting polypeptide (NTCP). We herein demonstrated that human immortalized hepatocyte NKNT‐3 cells exhibited susceptibility to HBV by exogenously expressing NTCP (NKNT‐3/NTCP cells). By comparing cyclic GMP‐AMP synthetase (cGAS)‐stimulator of interferon genes (STING) signaling pathway in several NKNT‐3/NTCP cell‐derived cell clones, we found that STING was highly expressed in cell clones exhibiting resistance but not susceptibility to HBV. High‐level expression of STING was implicated in HBV‐triggered induction of type III IFN and a pro‐inflammatory cytokine, IL‐6. In contrast, RNAi‐mediated knockdown of STING inhibited type III IFN induction and restored the levels of HBV total transcript in an HBV‐infected cell clone exhibiting resistance to HBV. These results suggest that STING regulates susceptibility to HBV by its expression levels. STING may thus be a novel target for anti‐HBV strategies.


| INTRODUCTION
Hepatitis B virus (HBV) is a hepatotropic virus classified into the Hepadnaviridae family. HBV infection causes chronic hepatitis, liver cirrhosis, and finally hepatocellular carcinoma (HCC). 1,2 The progression of hepatic diseases is tightly associated with the HBV-triggered host innate immune response and inflammatory response. To prevent the progression of hepatic diseases, it is important to suppress the HBV-triggered host innate immune response and inflammatory response.
The cytoplasmic DNA sensor, cyclic GMP-AMP synthetase (cGAS), is known to recognize viral DNA and cytoplasmic DNA as pathogen-associated molecular patterns (PAMPs). 3,4 After the recognition, cGAS produces cyclic GMP-AMP (cGAMP) and then uses cGAMP to activate a stimulator of interferon genes (STING). STING mediates activation of the transcription factor interferon regulatory factor 3 (IRF-3) and subsequently the induction of interferon (IFN)-β (type I IFN), 5 IFN-λ1, λ2, and λ3 (type III IFN). 6 Both type I and type III IFNs stimulate the induction of numerous IFN-stimulated genes (ISGs) such as ISG15 and ISG56 through the JAK-STAT signaling pathway. 7 On the other hand, STING also mediates the induction of pro-inflammatory cytokines such as IL-6 and IL-8 through the NF-κB signaling pathway. 8,9 As described here, both cGAS and STING are required for the innate immune response and inflammatory response. We previously reported that cGAS recognized HBV DNA and subsequently triggered an innate immune response in human hepatoma Li23 cells. 10 However, in that study, we could not examine the HBV-triggered inflammatory response, since Li23 cells were a human hepatoma cell line. To study HBV-triggered inflammatory responses, it will be necessary to establish an HBV infectious cell culture model from normal human hepatic cells rather than human hepatoma cells.
Sodium taurocholate cotransporting polypeptide (NTCP) is a functional receptor for HBV. 11 Human hepatoma HepG2 cells exhibit susceptibility to HBV by exogenously expressing NTCP. 11 HepG2/NTCP cells (HepG2 cells stably expressing exogenous NTCP) are currently used as an HBV infectious cell culture model for the study of HBV worldwide. However, we previously reported that HepG2 cells exhibited defective expression of endogenous cGAS. 10 This result suggests that HepG2/NTCP cells cannot be used for the study of endogenous cGAS-triggered innate immune response and inflammatory response. Our previous study also showed that cGAS was expressed in immortalized human hepatocyte NKNT-3 cells. 10 In the present study, we established NKNT-3 cells exhibiting susceptibility to HBV by the exogenous expression of NTCP. In addition, we obtained several NKNT-3/NTCPderived cell clones exhibiting susceptibility or resistance to HBV. Interestingly, STING was highly expressed in a cell clone exhibiting resistance to HBV. Here, we show that STING is an important host factor that regulates susceptibility to HBV by its expression levels. We also show that NKNT-3/NTCP cells are a novel HBV infectious cell culture model for the study of HBV-triggered innate immune responses and inflammatory responses.

| Cell culture
Human immortalized hepatocyte NKNT-3 cells, which were kindly provided by N. Kobayashi

| Establishment of an NKNT-3 cell line stably expressing exogenous NTCP and the derivation of its cell clones
NKNT-3 cells stably expressing exogenous NTCP (designated NKNT-3/NTCP cells) were established as previously described. 10 NKNT-3/NTCP-derived cell clones were isolated from their parental cells by the limited dilution method. We evaluated HBV susceptibility by HBV/NLuc assay 12 and, from the several tens of cell clones obtained, selected a cell clone exhibiting susceptibility or resistance to HBV. By repeating the cell cloning and selection process, we obtained cell clones exhibiting the different levels of susceptibility to HBV.

F I G U R E 1
The immortalized human hepatocyte cell line NKNT-3 exhibited susceptibility to HBV by expressing exogenous NTCP.

| Western blot analysis
Western blot analysis was performed as previously described. 13 Anti

| Flow cytometric analysis
Cell surface expression of exogenous NTCP was detected by a flow cytometer as previously reported. 14 Anti-Myc (PL14; Medical & Biological Laboratories), and FITC-conjugated goat anti-mouse antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) were used as primary and secondary antibody, respectively.

| RNA interference
The day after mock or HBV infection, we introduced small interfering RNAs (siRNAs) targeting STING or nontargeting siRNAs into NKNT-3/NTCP #28.3.25.13 cells as previously described. 17 At 4 days after the introduction of siRNAs, we isolated the total RNA or cell lysate, and subjected it to quantitative RT-PCR analysis or Western blot analysis, respectively.

| Generation of cells stably expressing exogenous STING
To construct pCX4bleo/HA-STING retroviral vector, we introduced STING (accession no. NM_198282) cDNA containing a full-length ORF into the pCX4bleo/HA retroviral vector as previously reported. 18 pCX4bleo/ HA-STING I200N, which causes the conformational disruption of STING, 19 was also constructed using PCR mutagenesis with primers containing base alterations. These vectors were introduced into NKNT-3/NTCP #28.3.8 cells by retroviral transfer and then the cells stably expressing exogenous STING or STING I200N were selected by Zeocin (Thermo Fisher Scientific, Carlsbad, CA, USA).

| Statistical analysis
Statistical analysis was performed to determine the significance of differences among groups by using Student's t test. P < 0.05 was considered statistically significant.

NKNT-3 cells exhibited susceptibility to HBV via their expression of exogenous NTCP
Since HepG2 cells were a human hepatoma cell line and exhibited defective expression of endogenous cGAS, 10 we tried to establish HBV infectious cell culture model from immortalized human hepatocyte NKNT-3 cells, which has been exhibited a nonneoplastic phenotype 20 and the

FIGURE 2
The level of susceptibility to HBV in NKNT-3/NTCP #28.3.8 cells approximated that in HepG2/NTCP cells. A, Outline of cell cloning by the limited dilution method. NKNT-3/NTCP #28.3.8 cells were selected by three-round limited dilution. Red arrows with solid lines show the selection of a cell clone exhibiting higher susceptibility to HBV. B, Comparison of susceptibility to HBV among parent NKNT-3/NTCP cells and their derived cell clones by using HBV/NL assay. **P < 0.01, ***P < 0.001 versus HBV/NL-infected parent NKNT-3/NTCP cells. C, Flow cytometric analysis of the cell surface NTCP in their derived cell clones. Signals of the cell surface NTCP are shown in green. An isotype control was used as a negative control (violet area). D, Comparison of the amounts of HBV total transcript after HBV infection among parent NKNT-3/NTCP cells and their derived cell clones. The amount of HBV total transcript was measured after HBV infection by quantitative RT-PCR analysis. *P < 0.05 versus HBV-infected parent NKNT-3/NTCP cells. (E, F) Comparison of susceptibility to HBV between HepG2/NTCP cells and NKNT-3/NTCP #28.3.8 cells. Intracellular NLuc activity or the amounts of HBV total transcript were measured as described in Figure 1E,F. NS; not significant, **P < 0.01, ***P < 0.001 versus HBV/NL-or HBV-infected HepG2/NTCP cells, respectively. G, Comparison of susceptibility to HBV between HepG2/NTCP cells and NKNT-3/NTCP #28.3.8 cells by Northern blot analysis. Total RNA was isolated from HBV-infected cells at 13 d after HBV inoculation. 28S rRNA and 18S rRNA were included as a loading control. NKNT-3/NTCP #28.8.4 is another clone, which has been estimated to exhibit susceptibility to HBV by HBV/NL assay (data not shown) endogenous expression of cGAS. 10 HepG2 cells have been reported to exhibit susceptibility to HBV through their expression of exogenous NTCP. 11 Therefore, to establish NKNT-3 cells exhibiting susceptibility to HBV, we first prepared NKNT-3 cells stably expressing exogenous NTCP-myc (designated NKNT-3/NTCP cells; Figure 1A).
After the infection with HBV/NLuc or HBV, both level of NLuc activity and HBV total transcript were increased in NKNT-3/NTCP cells in a time-dependent manner, but not in NKNT-3/Control cells ( Figure 1C,D). We next compared the level of susceptibility to HBV in NKNT-3/NTCP cells with that in HepG2/NTCP cells. The levels of NLuc activity, HBV total transcript, and pregenomic RNA (pgRNA) in HBV/NLuc-or HBV-infected NKNT-3/NTCP cells were almost 10 times lower than those in HBV/NLuc-or HBVinfected HepG2/NTCP cells ( Figure 1E,F). We further examined whether or not the exogenous NTCP was functional in NKNT-3/NTCP cells. Cyclosporin A (CsA) was previously reported to inhibit HBV entry by targeting NTCP. 22 When administered before and during HBV inoculation, CsA inhibited the levels of HBV total transcript in HBVinfected NKNT-3/NTCP cells as well as in HBV-infected HepG2/NTCP cells ( Figure 1G). These results suggest that NKNT-3 cells exhibit susceptibility to HBV by exogenously expressing functional NTCP.

NKNT-3/NTCP #28.3.8 cells approximated that in HepG2/NTCP cells
Since susceptibility to HBV in NKNT-3/NTCP cells was lower than that in HepG2/NTCP cells ( Figure 1E,F), we next tried to select a subcloned cell line exhibiting higher susceptibility to HBV than NKNT-3/NTCP cells (Figure 2A). During three-round serial limited dilution, we obtained three distinct cell clones (#28, #28.3, and #28.3.8 cells, respectively; Figure 2A) that met this criterion ( Figure 2B). Exogenous NTCP was expressed on the cell surface in all three clones ( Figure 2C). Among them, the NKNT-3/NTCP #28.3.8 cells exhibited the highest levels of HBV total transcript after HBV infection ( Figure 2D). Therefore, we next compared the levels of susceptibility to HBV in NKNT-3/NTCP #28.3.8 cells with those in HepG2/NTCP cells. Upon the infection with HBV/NLuc or HBV, both levels of NLuc activity ( Figure  2E) and HBV total transcript ( Figure 2F) in NKNT-3/NTCP #28.3.8 cells approximated those in HepG2/NTCP cells. Consistent with these results, Northern blot analysis also showed that the levels of HBV pgRNA and 2.1/2.3 kb RNA in NKNT-3/NTCP #28.3.8 cells were roughly the same as those in HepG2/NTCP cells after HBV infection ( Figure 2G). These results suggest that NKNT-3/NTCP #28.3.8 cells are useful as an HBV infectious cell culture model in the manner of HepG2/NTCP cells.

| DISCUSSION
Cytoplasmic DNA or RNA sensors trigger the innate immune responses and the inflammatory responses by recognizing viral PAMPs. We previously reported that one of the cytoplasmic DNA sensors, cGAS, recognized HBV DNA as viral PAMPs and subsequently induced the innate immune response through its adaptor protein, STING. 10 In the present study, we found that the immortalized human hepatocyte NKNT-3 cells exhibited HBV susceptibility by stably expressing the exogenous NTCP ( Figure 1C,D). Cells of one of the NKNT-3/NTCP cell-derived clones, NKNT-3/NTCP #28.3.25.13, highly expressed STING and exhibited resistance to HBV through STING-mediated type III IFN induction ( Figure 4C,E,F). Interestingly, STING was highly phosphorylated in p-dGdC-transfected NKNT-3/NTCP #28.3.25.13 cells, but not in the parent, #28, #28.3, or #28.3.8 cells ( Figure 4D). However, it is uncertain why the expression and phosphorylation levels of STING differed among the NKNT-3/NTCP cell-derived cell clones. In humans, several single nucleotide polymorphisms (SNPs) of STING have been discovered. 24 SNPs of STING have been shown to cause autoinflammatory diseases such as STING-associated vasculopathy with onset in infancy 25 and familial chilblain lupus. 26 These SNPs are implicated in the dysregulation of host innate immune responses and inflammatory responses through a loss-of-function mutation or a gain-of-function mutation of STING. Further analysis is needed to identify the gain-of-function mutation(s) in STING in NKNT-3/NTCP #28.3.25.13 cells.
In the present study, we showed that HBV infection induced type III IFN, but not IFN-β (type I IFN), through a STING-mediating signaling pathway in NKNT-3/NTCP #28.3.25.13 cells ( Figure 5F). Sato et al previously reported that a cytoplasmic RNA sensor, RIG-I, recognized HBV pgRNA and subsequently induced type III but not type I IFN through its adaptor protein, IPS-1, in human primary hepatocytes. 27 These results suggest that HBV suppresses the induction of type I IFN but not type III IFN. One of the HBV proteins, HBV polymerase, suppressed STINGmediated IFN-β induction by disrupting K63-linked ubiquitination of STING. 28 Another study also reported that HBx bound IPS-1 and suppressed the activation of IFN-β. 29 However, in these studies, it was unclear whether HBV suppressed the induction of type III IFN through these HBV proteins. Our results showed that HBV transiently induced ISG56 mRNA induction at 5 and 9 days, but not at 13 days, after HBV infection in NKNT-3/NTCP #28.3.25.13 cells ( Figure 3D). This result suggests that HBV possesses two opposite functions to simultaneously trigger or suppress the induction of type III IFN. Further analysis is needed to examine whether or not HBV suppresses the induction of type III IFN.
We also showed that HBV infection induced a proinflammatory cytokine, IL-6, through the noncanonical NF-κB signaling pathway in NKNT-3/NTCP #28.3.25.13 cells ( Figure 5F). STING also mediates host inflammatory responses by triggering its downstream NF-κB signaling pathway. 8,9 A STING-triggered host inflammatory response has been reported to be associated with hepatic diseases. 30,31 In nonalcoholic fatty liver disease, STING promotes hepatocyte injury by inducing inflammation. 30 In addition, STING mediates liver injury and fibrosis in mice administered CCl 4 (a chemical inducer of hepatocyte death). 31 Moreover, based on the results of several previous studies, STING is also thought to play an important role in tumor development. 32 Interestingly, STING may exert two opposite effects (tumor-suppressing and tumor-promoting effects) on tumor development under different situations. For example, in breast cancer, STING and its downstream signaling may suppress the tumor or the cancer metastasis. 33,34 In contrast, STING is also required for cell survival and regrowth in breast cancer. 35,36 However, the results of the present study do not clarify whether the HBV-triggered NF-κB signaling pathway causes liver diseases and tumor development. Further analysis will also be needed to examine how HBV causes liver diseases and finally HCC through a STING-mediated NF-κB signaling pathway.
In the present study, we established a novel HBV infectious cell culture model by using NKNT-3 cells. Since NKNT-3 cells exhibit a nonneoplastic phenotype, 20 our HBV infectious cell culture model is expected to be a useful tool for the study of hepatic carcinogenesis caused by HBV-triggered innate immune responses and inflammatory responses.
H. Dansako wrote the paper; and all authors reviewed the manuscript.