Human blood dendritic cell antigen 3 (BDCA3)+ dendritic cells are a potent producer of interferon‐λ in response to hepatitis C virus

The polymorphisms in the interleukin (IL)‐28B (interferon‐lambda [IFN]‐λ3) gene are strongly associated with the efficacy of hepatitis C virus (HCV) clearance. Dendritic cells (DCs) sense HCV and produce IFNs, thereby playing some cooperative roles with HCV‐infected hepatocytes in the induction of interferon‐stimulated genes (ISGs). Blood dendritic cell antigen 3 (BDCA3)+ DCs were discovered as a producer of IFN‐λ upon Toll‐like receptor 3 (TLR3) stimulation. We thus aimed to clarify the roles of BDCA3+ DCs in anti‐HCV innate immunity. Seventy healthy subjects and 20 patients with liver tumors were enrolled. BDCA3+ DCs, in comparison with plasmacytoid DCs and myeloid DCs, were stimulated with TLR agonists, cell‐cultured HCV (HCVcc), or Huh7.5.1 cells transfected with HCV/JFH‐1. BDCA3+ DCs were treated with anti‐CD81 antibody, inhibitors of endosome acidification, TIR‐domain‐containing adapter‐inducing interferon‐β (TRIF)‐specific inhibitor, or ultraviolet‐irradiated HCVcc. The amounts of IL‐29/IFN‐λ1, IL‐28A/IFN‐λ2, and IL‐28B were quantified by subtype‐specific enzyme‐linked immunosorbent assay (ELISA). The frequency of BDCA3+ DCs in peripheral blood mononuclear cell (PBMC) was extremely low but higher in the liver. BDCA3+ DCs recovered from PBMC or the liver released large amounts of IFN‐λs, when stimulated with HCVcc or HCV‐transfected Huh7.5.1. BDCA3+ DCs were able to induce ISGs in the coexisting JFH‐1‐positive Huh7.5.1 cells. The treatments of BDCA3+ DCs with anti‐CD81 antibody, cloroquine, or bafilomycin A1 reduced HCVcc‐induced IL‐28B release, whereas BDCA3+ DCs comparably produced IL‐28B upon replication‐defective HCVcc. The TRIF‐specific inhibitor reduced IL‐28B release from HCVcc‐stimulated BDCA3+ DCs. In response to HCVcc or JFH‐1‐Huh7.5.1, BDCA3+ DCs in healthy subjects with IL‐28B major (rs8099917, TT) released more IL‐28B than those with IL‐28B minor genotype (TG). Conclusion: Human BDCA3+ DCs, having a tendency to accumulate in the liver, recognize HCV in a CD81‐, endosome‐, and TRIF‐dependent manner and produce substantial amounts of IL‐28B/IFN‐λ3, the ability of which is superior in subjects with IL‐28B major genotype. (HEPATOLOGY 2013)

H epatitis C virus (HCV) infection is one of the most serious health problems in the world. More than 170 million people are chronically infected with HCV and are at high risk of developing liver cirrhosis and hepatocellular carcinoma. Genomewide association studies have successfully identified the genetic polymorphisms (single nucleotide polymorphisms, SNPs) upstream of the promoter region of the interleukin (IL)-28B / interferon-lambda 3 (IFN-k3) gene, which are strongly associated with the efficacy of pegylated interferon-a (PEG-IFN-a) and ribavirin therapy or spontaneous HCV clearance. [1][2][3][4] IFN-ks, or type III IFNs, comprise a family of highly homologous molecules consisting of IFN-k1 (IL-29), IFN-k2 (IL-28A), and IFN-k3 (IL-28B). In clear contrast to type I IFNs, they are released from relatively restricted types of cells, such as hepatocytes, intestinal epithelial cells, or dendritic cells (DCs). Also, the cells that express heterodimeric IFN-k receptors (IFN-kR1 and IL-10R2) are restricted to cells of epithelial origin, hepatocytes, or DCs. 5 Such limited profiles of cells expressing IFN-ks and their receptors define the biological uniqueness of IFN-ks. It has been shown that IFN-ks convey anti-HCV activity by inducing various interferon-stimulated genes (ISGs), 5 the profiles of which were overlapped but others were distinct from those induced by IFN-a/b. Some investigators showed that the expression of IL-28 in PBMC was higher in subjects with IL-28B major than those with minor; however, the levels of IL-28 transcripts in liver tissue were comparable regardless of IL-28B genotype. 2,6 At the primary exposure to hosts, HCV maintains high replicative levels in the infected liver, resulting in the induction of IFNs and ISGs. In a case of successful HCV eradication, it is postulated that IFN-a/b and IFN-k cooperatively induce antiviral ISGs in HCVinfected hepatocytes. It is of particular interest that, in primary human hepatocytes or chimpanzee liver, IFNks, but not type I IFNs, are primarily induced after HCV inoculation, the degree of which is closely correlated with the levels of ISGs. 7 These results suggest that hepatic IFN-k could be a principal driver of ISG induction in response to HCV infection. Nevertheless, the possibility remains that DCs, as a prominent IFN producer in the liver, play significant roles in inducing hepatic ISGs and thereby suppressing HCV replication.
DCs, as immune sentinels, sense specific genomic and/or structural components of pathogens with various pattern recognition receptors and eventually release IFNs and inflammatory cytokines. 8 In general, DCs migrate to the organ where inflammation or cellular apoptosis occurs and alter their function in order to alleviate or exacerbate the disease conditions. There-fore, the phenotypes and/or capacity of liver DCs are deemed to be influenced in the inflamed liver. In humans, the existence of phenotypically and functionally distinct DC subsets has been reported: myeloid DC (mDC) and plasmacytoid DC (pDC). 9 Myeloid DCs predominantly produce IL-12 or tumor necrosis factor alpha (TNF-a) following proinflammatory stimuli, while pDCs release considerable amounts of type I IFNs upon virus infection. 9 The other type of mDCs, mDC2 or BDCA3 þ (CD141) DCs, have been drawing much attention recently, since human BDCA3 þ DCs are reported to be a counterpart of murine CD8a þ DCs. 10 Of particular interest is the report that BDCA3 þ DCs have a potent capacity of releasing IFN-k in response to Toll-like receptor 3 (TLR3) agonist. 11 However, it is still largely unknown whether human BDCA3 þ DCs are able to respond to HCV.
Taking these reports into consideration, we hypothesized that human BDCA3 þ DCs, as a producer of IFN-ks, have crucial roles in anti-HCV innate immunity. We thus tried to clarify the potential of BDCA3 þ DCs in producing type III IFNs by using cell-cultured HCV (HCVcc) or hepatoma cells harboring HCV as stimuli. Our findings show that BDCA3 þ DCs are quite a unique DC subset, characterized by a potent and specialized ability to secrete IFN-ks in response to HCV. The ability of BDCA3 þ DCs to release IL-28B upon HCV is superior in subjects with IL-28B major (rs8099917, TT) to those with minor (TG or GG) genotype, suggesting that BDCA3 þ DCs are one of the key players in IFN-k-mediated innate immunity.

Patients and Methods
Subjects. This study enrolled 70 healthy volunteers (male/female: 61/9) (age: mean 6 standard deviation [SD], 37.3 6 7.8 years) and 20 patients who underwent surgical resection of liver tumors at Osaka University Hospital (Supporting Table 1). The study was approved by the Ethical Committee of Osaka University Graduate School of Medicine. Written informed consent was obtained from all of them. All healthy volunteers were negative for HCV, hepatitis B virus (HBV), and human immunodeficiency virus (HIV) and had no apparent history of liver, autoimmune, or malignant diseases.
Reagents. The specifications of all antibodies used for FACS or cell sorting TLR-specific synthetic agonists, pharmacological reagents, and inhibitory peptides are listed in the Supporting Materials.
Separation of DCs from PBMC or Intrahepatic Lymphocytes. We collected 400 mL of blood from each healthy volunteer and processed them for PBMCs. Noncancerous liver tissues were obtained from patients who underwent resection of liver tumors (Supporting Table 1). For the collection of intrahepatic lymphocytes (IHLs), liver tissues were washed thoroughly with phosphate-buffered saline to remove the peripheral blood adhering to the tissue and ground gently. After Lin-negative (CD3 À , CD14 À , CD19 À , and CD56 À ) cells were obtained by the MACS system, each DC subset with the defined phenotype was sorted separately under FACS Aria (BD). The purity was more than 98%, as assessed by FACS Canto II (BD). Sorted DCs were cultured at 2.5 Â 10 4 /well on 96-well culture plates.
Immunofluorescence Staining of Human Liver Tissue. Tissue specimens were obtained from surgical resections of noncancerous liver from the patients as described above. Briefly, the 5-mm sections were incubated with the following antibodies: mouse biotinylated antihuman BDCA3 antibody (Miltenyi-Biotec), and mouse antihuman CLEC9A antibody (Biolegend) and subsequently with secondary goat antirabbit Alexa Fluor488 or goat antimouse Alexa Fluor594 (Invitrogen, Molecular Probes) antibodies. Cell nuclei were counterstained with Dapi-Fluoromount-GTM (Southern Biotech, Birmingham, AL). The stained tissues were analyzed by fluorescence microscopy (Model BZ-9000; Keyence, Osaka, Japan).
Cells and Viruses. The in vitro transcribed RNA of the JFH-1 strain of HCV was introduced into FT3-7 cells 12 or Huh7.5.1 cells. The stocks of HCVcc were generated by concentration of the medium from JFH-1-infected FT3-7 cells. The virus titers were determined by focus forming assay. 13 The control medium was generated by concentration of the medium from HCVuninfected FT3-7 cells. Infectious JEVs were generated from the expression plasmid (pMWJEATG1) as reported. 14 HSV (KOS) was a generous gift from Dr. K. Ueda (Osaka University). Huh7.5.1 cells transduced with HCV JFH-1 strain was used for the coculture with DCs. The transcripts of ISGs in Huh7.5.1 were examined by reverse-transcription polymerase chain reaction (RT-PCR) methods using gene-specific primers and probes (Applied Biosystems, Foster City, CA).
Statistical Analysis. The differences between two groups were assessed by the Mann-Whitney nonparametric U test. Multiple comparisons between more than two groups were analyzed by the Kruskal-Wallis nonparametric test. Paired t tests were used to compare differences in paired samples. All the analyses were performed using GraphPad Prism software (San Diego, CA).

Results
Human BDCA3 1 DCs Are Phenotypically Distinct from pDCs and mDCs. We defined BDCA3 þ DCs as Lin À HLA-DR þ BDCA3 highþ cells ( Fig. 1A, left, middle), and pDCs and mDCs by the patterns of CD11c and CD123 expressions (Fig. 1A,  right). The level of CD86 on pDCs or mDCs is comparatively higher than those on BDCA3 þ DCs (Fig.  1B). The expression of CD81 is higher on BDCA3 þ DCs than on pDCs and mDCs (Fig. 1B, Supporting  Fig. S1). CLEC9A, a member of C-type lectin, is expressed specifically on BDCA3 þ DCs as reported elsewhere, 16 but not on pDCs and mDCs (Fig. 1B).
Liver BDCA3 1 DCs Are More Mature than the Counterparts in the Periphery. BDCA3 þ DCs in infiltrated hepatic lymphocytes (IHLs) are all positive for CLEC9A, but liver pDCs or mDCs are not (data not shown). The levels of CD40, CD80, CD83, and CD86 on liver BDCA3 þ DCs are higher than those on the peripheral counterparts, suggesting that BDCA3 þ DCs are more mature in the liver compared to those in the periphery (Fig. 1C).
In order to confirm that BDCA3 þ DCs are localized in the liver, we stained the cells with immunofluorescence antibodies (Abs) in noncancerous liver tissues. Liver BDCA3 þ DCs were defined as BDCA3 þ CLEC9A þ cells (Fig. 1D). Most of the cells were found near the vascular compartment or in sinusoid or the space of Disse of the liver tissue.  The percentages of BDCA3 þ DCs in the IHLs were significantly higher than those in PBMCs from relevant donors (Fig. 2C). Such relative abundance of BDCA3 þ DCs in the liver over that in the periphery was observed regardless of the etiology of the liver disease (Supporting Table 1).
BDCA3 1 DCs Produce a Large Amount of IFN-ks upon Poly IC Stimulation. We compared DC subsets for their abilities to produce IL-29/IFN-k1, IL-28A/ IFN-k2, IL-28B/IFN-k3, IFN-b, and IFN-a in response to TLR agonists. Approximately 4.0 Â 10 4 of BDCA3 þ DCs were recoverable from 400 mL of donated blood from healthy volunteers. We fixed the number of DCs at 2.5 Â 10 4 cells/100 mL for comparison in the following experiments.
BDCA3 þ DCs have been reported to express mRNA for TLR1, 2, 3, 6, 8, and 10. 17 First, we quantified IL-28B/IFN-k3 as a representative for IFN-ks after stimulation of BDCA3 þ DCs with relevant TLR agonists. We confirmed that BDCA3 þ DCs released IL-28B robustly in response to TLR3 agonist/poly IC but not to other TLR agonists (Fig. S2). In contrast, pDCs produced IL-28B in response to TLR9 agonist/ CpG but much lesser to other agonists (Fig. S2). Next, we compared the capabilities of DCs inducing IFN-ks and IFN-b genes in response to relevant TLR agonists. BDCA3 þ DCs expressed extremely high levels of IL-29, IL-28A, and IL-28B transcripts compared to other DCs, whereas pDCs induced a higher level of IFN-b than other DCs (Fig. S3A).
Similar results were obtained with the protein levels of IFN-ks, IFN-b, and IFN-a released from DC subsets stimulated with TLR agonists. BDCA3 þ DCs produce significantly higher levels of IL-29, IL-28B, and IL-28A than the other DC subsets. In clear contrast, pDCs release a significantly larger amount of IFN-b and IFN-a than BDCA3 þ DCs or mDCs (Fig. 3A,  Fig. S3B). As for the relationship among the quantity of IFN-k subtypes from poly IC-stimulated BDCA3 þ DCs, the levels of IL-29/IFN-k1 and IL-28B/IFN-k3 were positively correlated (R 2 ¼ 0.76, P < 0.05), and those of IL-28A/IFN-k2 and IL-28B/IFN-k3 were positively correlated as well (R 2 ¼ 0.84, P < 0.0005), respectively (Fig. S3C). These results show that the transcription and translation machineries of IFN-ks may be overlapped among IFN-k subtypes in BDCA3 þ DCs upon poly IC stimulation.
Liver BDCA3 þ DCs sorted from IHLs possess the ability to produce IL-28B in response to poly IC (Fig.  3B), showing that they are comparably functional. In response to poly IC, BDCA3 þ DCs were capable of producing inflammatory cytokines as well, such as TNF-a, IL-6, and IL-12p70 (Fig. S4A). By using Huh7 cells harboring HCV subgenomic replicons (HCV-N, genotype 1b), we confirmed that the supernatants from poly IC-stimulated BDCA3 þ DCs suppressed HCV replication in an IL-28B concentrationdependent manner (Fig. S4B). Therefore, poly ICstimulated BDCA3 þ DCs are capable of producing biologically active substances suppressing HCV replication, some part of which may be mediated by IFN-ks.
BDCA3 1 DCs Produce IL-28B upon HCVcc or HCV/JFH-1-Transfected Huh7.5.1 Cells. We stimulated freshly isolated BDCA3 þ DCs, pDCs and mDCs with infectious viruses, such as HCVcc, Japanese encephalitis virus (JEV), and herpes simplex virus (HSV). In preliminary experiments, we confirmed that HCVcc stimulated BDCA3 þ DCs to release IL-28B in a dose-dependent manner (Fig. S5). BDCA3 þ DCs produced a large amount of IL-28B upon exposure to HCVcc and released a lower amount of IFN-a upon HCVcc or HSV (Fig. 4A). In contrast, pDCs produced a large amount of IFN-a in response to HCVcc and HSV and a much lower level of IL-28B upon HCVcc (Fig. S6). In mDCs, IL-28B and IFN-a were not detectable with any of these viruses (data not shown).
BDCA3 þ DCs produced significantly higher levels of IL-28B than the other DCs upon HCVcc stimulation (Fig. 4B). By contrast, HCVcc-stimulated pDCs released significantly larger amounts of IFN-b and IFN-a than the other subsets (Fig. 4B). Liver BDCA3 þ DCs were capable of producing IL-28B in response to HCVcc (Fig. 4C). These results show that, upon HCVcc stimulation, BDCA3 þ DCs produce more IFN-ks and pDCs release more IFN-b and IFNa than the other DC subsets, respectively. Taking a clinical impact of IL-28B genotypes on HCV eradication into consideration, we focused on IL-28B/IFN-k3 as a representative for IFN-ks in the following experiments.
In a coculture with JFH-1-infected Huh7.5.1 cells, BDCA3 þ DCs profoundly released IL-29, IL-28A, and IL-28B (Fig. 4D, the results of IL-29 and IL-28A, not shown), whereas BDCA3 þ DCs failed to respond to Huh7.5.1 cells lacking HCV/JFH-1, showing that IL-28B production from BDCA3 þ DCs is dependent on HCV genome (Fig. 4D). In the absence of BDCA3 þ DCs, IL-28B is undetectable in the supernatant from JFH-1-infected Huh7.5.1 cells, demonstrating that BDCA3 þ DCs, not HCV-replicating Huh7.5.1 cells, produce detectable amount of IL-28B (Fig. 4D). In the coculture, BDCA3 þ DCs comparably released IL-28B either in the presence or the absence of transwells, suggesting that cell-to-cell contact between DCs and Huh7.5.1 cells is dispensable for IL-28B response (Fig. 4E). In parallel with the quantity of IL-28B in the coculture, ISG15 was significantly induced only in JFH-1-infected Huh7.5.1 cells cocultured with BDCA3 þ DCs (Fig. 4F). A strong induction was observed with other ISGs in JFH-1-infected Huh7.5.1 in the presence of BDCA3 þ DCs, such as IFIT1, MxA, RSD2, IP-10, and USP18 (Fig. S7). The results clearly show that BDCA3 þ DCs are capable of producing large amounts of IFN-ks in response to cellular or cell-free HCV, thereby inducing various ISGs in bystander liver cells. were obtained from patients with non-B, non-C liver disease and that of case 17 was from an HCV-infected patient (Supporting Table 1). CD81 and Endosome Acidification Are Involved in IL-28B Production from HCV-Stimulated BDCA3 1 DCs, but HCV Replication Is Not Involved. It is not known whether HCV entry and subsequent replication in DCs is involved or not in IFN response. 18,19 To test this, BDCA3 þ DCs were inoculated with UVirradiated, replication-defective HCVcc. We confirmed that UV exposure under the current conditions is sufficient to negate HCVcc replication in Huh7.5.1 cells, as demonstrated by the lack of expression of NS5A after inoculation (data not shown). BDCA3 þ DCs produced comparable levels of IL-28B with UV-treated HCVcc, indicating that active HCV replication is not necessary for IL-28B production (Fig. 5A). We next examined whether or not the association of HCVcc with BDCA3 þ DCs by CD81 is required for IL-28B production. It has been reported that the E2 region of HCV structural protein is associated with CD81 on cells when HCV enters susceptible cells. 13,20 We confirmed that all DC subsets express CD81, the degree of which was most significant on BDCA3 þ DCs (Fig. 1B, Fig. S1). Masking of CD81 with Ab significantly impaired IL-28B production from HCVcc-stimulated BDCA3 þ DCs in a dose-dependent manner (Fig.  5A, Fig. S8), suggesting that HCV-E2 and CD81 interaction is involved in the induction. The treatment of poly IC-stimulated BDCA3 þ DCs with anti-CD81 Ab failed to suppress IL-28B production (Fig. 5B).
HCV enters the target cells, which is followed by fusion steps within acidic endosome compartments. Chloroquine and bafilomycin A1 are well-known and broadly used inhibitors of endosome TLRs, which are reported to be capable of blocking TLR3 response in human monocyte-derived DC. 21,22 In our study, the treatment of BDCA3 þ DCs with chloroquine, bafilo-mycin A1, or NH4Cl significantly suppressed their IL-28B production either in response to HCVcc or poly IC (Fig. 5A,B, NH4Cl, data not shown). These results suggest that the endosome acidification is involved in HCVcc-or poly IC-stimulated BDCA3 þ DCs to produce IL-28B. The similar results were obtained with HCVcc-stimulated pDCs for the production of IL-28B (Fig. S9). We validated that such concentration of chloroquine (10 mM) and bafilomycin A1 (25 nM) did not reduce the viability of BDCA3 þ DCs (Fig. S10).
BDCA3 1 DCs Produce IL-28B in Response to HCVcc by a TIR-Domain-Containing Adapter-Inducing Interferon-b (TRIF)-Dependent Mechanism. TRIF/TICAM-1, a TIR domain-containing adaptor, is known to be essential for the TLR3-mediated pathway. 23 In order to elucidate whether TLR3-dependent pathway is involved or not in IL-28B response of BDCA3 þ DCs, we added the cell-permeable TRIFspecific inhibitory peptide (Invivogen) or the control peptide to poly IC-or HCVcc-stimulated BDCA3 þ DCs. Of particular interest, the TRIF-specific inhibitor peptide, but not the control one, significantly suppressed IL-28B production from poly IC-or HCVccstimulated BDCA3 þ DCs (Fig. 6A,B). In clear contrast, the TRIF-specific inhibitor failed to suppress IL-28B from HCVcc-stimulated pDCs (Fig. 6C), suggesting that pDCs recognize HCVcc in an endosome-dependent but TRIF-independent pathway. These results show that BDCA3 þ DCs may recognize HCVcc by way of the TRIF-dependent pathway to produce IL-28B.
BDCA3 1 DCs in Subjects with IL-28B Major Genotype Produce More IL-28B in Response to HCV than Those with IL-28B Minor Type. In order to compare the ability of BDCA3 þ DCs to release IL-28B in healthy subjects between IL28B major (rs8099917, TT) In some experiments, UV-irradiated HCVcc was used at the same MOI, and BDCA3 þ DCs were treated with anti-CD81Ab (5 mg/mL), chloroquine (10 mM), or bafilomycin A1 (25 nM). The results are expressed as ratios of IL-28B quantity with or without the treatments. They are shown as mean 6 SEM from five experiments. *P < 0.05 by paired t test. C, control; CLQ, treatment with chloroquine; Baf, treatment with bafilomycin A1; UV, ultraviolet-irradiated HCVcc; n.d., not detected. and minor hetero (TG) genotypes, we stimulated BDCA3 þ DCs of the identical subjects with poly IC (25 mg/mL, 2.5 mg/mL, 0.25 mg/mL), HCVcc or JFH-1infected Huh 7.5.1, and subjected them to ELISA. The levels of IL-28B production by poly IC-stimulated BDCA3 þ DCs were comparable between subjects with IL-28B major and minor type (Fig. 7A). Similar results were obtained with the lesser concentrations of poly IC (Fig. S11). Of particular interest, in response to HCVcc or JFH-1 Huh7.5.1 cells, the levels of IL-28B from BDCA3 þ DCs were significantly higher in subjects with IL-28B major than those with minor type (Fig. 7B,C, S12).

Discussion
In this study we demonstrated that human BDCA3 þ DCs (1) are present at an extremely low frequency in PBMC but are accumulated in the liver; (2) are capable of producing IL-29/IFN-k1, IL-28A/IFN-k2, and IL-28B/IFN-k3 robustly in response to HCV; (3) recognize HCV by a CD81-, endosome acidification and TRIF-dependent mechanism; and (4) produce larger amounts of IFN-ks upon HCV stimulation in subjects with IL-28B major genotype (rs8099917, TT). These characteristics of BDCA3 þ DCs are quite unique in comparison with other DC repertoires in the settings of HCV infection.
At the steady state, the frequency of DCs in the periphery is relatively lower than that of the other immune cells. However, under disease conditions or physiological stress, activated DCs dynamically migrate to the site where they are required to be functional. However, it remains obscure whether functional BDCA3 þ DCs exist or not in the liver. We identified BDCA3 þ CLEC9A þ cells in the liver tissue (Fig. 1D). In a paired frequency analysis of BDCA3 þ DCs between in PBMCs and in IHLs, the cells are more abundant in the liver. The phenotypes of liver BDCA3 þ DCs were more mature than the PBMC counterparts. In support of our observations, a recent publication showed that CD141 þ (BDCA3 þ ) DCs are accumulated and more mature in the liver, the trend of which is more in HCV-infected liver. 24 We confirmed that liver BDCA3 þ DCs are functional, capable of releasing IFN-ks in response to poly IC or HCVcc.
BDCA3 þ DCs were able to produce large amounts of IFN-ks but much less IFN-b or IFN-a upon TLR3 stimulation. In contrast, in response to TLR9 agonist, Fig. 6. BDCA3 þ DCs produce IL-28B upon HCVcc stimulation in a TRIF-dependent mechanism. BDCA3 þ DCs or pDCs had been treated with 5 or 50 mM TRIF inhibitory peptide or control peptide for 2 hours. Subsequently, BDCA3 þ DCs were stimulated with Poly IC (25 lg/mL) or HCVcc (MOI ¼ 10), and pDCs were stimulated with HCVcc (MOI ¼ 10), respectively. IL-28B was quantified by ELISA. They are shown as mean 6 SEM from five experiments. *P < 0.05 by paired t test. C, TRIF control peptide; I, TRIF inhibitory peptide. Fig. 7. In response to HCVcc, BDCA3 þ DCs of healthy donors with IL-28B major genotype (rs8099917, TT) produced more IL-28B than those with minor type (TG). BDCA3 þ DCs of healthy donors with IL-28B TT (rs8099917) or TG genotype were cultured at 2.5 Â 10 4 cells with 25 mg/mL poly IC (A), with HCVcc at an MOI of 10 (B), or with JFH-1infected Huh 7.5.1 cells (C) for 24 hours. The supernatants were subjected to IL-28B ELISA. The same healthy donors were examined for distinct stimuli. The results are the mean 6 SEM from 15 donors with TT and 8 with TG, respectively. *P < 0.05 by Mann-Whitney U test.
pDCs released large amounts of IFN-b and IFN-a but much less IFN-ks. Such distinctive patterns of IFN response between BDCA3 þ DCs and pDCs are of particular interest. It has been reported that interferon regulatory factor (IRF)-3, IRF-7, or nuclear factor kappa B (NF-jB) are involved in IFN-b and IFN-k1, while IRF-7 and NF-jB are involved in IFN-a and IFN-k2/k3. 5 Presumably, the stimuli with TLR3/retinoic acid-inducible gene-I (RIG-I) (poly IC) or TLR9 agonist (CpG-DNA) in DCs are destined to activate these transcription factors, resulting in the induction of both types of IFN at comparable levels. However, the results of the present study did not agree with such overlapping transcription factors for IFN-ks, IFN-b, and IFN-a. Two possible explanations exist for different levels of IFN-ks and IFN-a production by BDCA3 þ DCs and pDCs. First, the transcription factors required for full activation of IFN genes may differ according to the difference of DC subsets. The second possibility is that since type III IFN genes have multiple exons, they are potentially regulated by posttranscriptional mechanisms. Thus, it is possible that such genetic and/or posttranscriptional regulation is distinctively executed between BDCA3 þ DCs and pDCs. Comprehensive analysis of gene profiles downstream of TLRs or RIG-I in BDCA3 þ DCs should offer some information on this important issue.
BDCA3 þ DCs were found to be more sensitive to HCVcc than JEV or HSV in IL-28B/IFN-k3 production. Such different strengths of IL-28B in BDCA3 þ DCs depending on the virus suggest that different receptors are involved in virus recognition. Again, the question arises of why BDCA3 þ DCs produce large amounts of IFN-ks compared to the amounts produced by pDCs in response to HCVcc. Considering that IRF-7 and NF-jB are involved in the transcription of the IL-28B gene, it is possible that BDCA3 þ DCs successfully activate both transcription factors upon HCVcc for maximizing IL-28B, whereas pDCs fail to do so. In support for this possibility, in pDCs it is reported that NF-jB is not properly activated upon HCVcc or hepatoma cell-derived HCV stimulations. 25 In the present study we demonstrated that HCV entry into BDCA3 þ DCs through CD81 and subsequent endosome acidification are critically involved in IL-28B responses. Involvement of TRIF-dependent pathways in IL-28B production was shown by the significant inhibition of IL-28B with TRIF inhibitor. Nevertheless, active HCV replication in the cells is not required. Based on our data, we considered that BDCA3 þ DCs recognize HCV genome mainly by an endosome and TRIF-dependent mechanism. Although the results with UV-irradiated HCVcc, anti-CD81 blocking Ab, and chloroquine were quite similar, the TRIF-specific inhibitor failed to suppress IL-28B from pDCs (Fig. 6, Fig. S9).
In the coculture with JFH-transfected Huh7.5.1 cells, BDCA3 þ DCs presumably receive some signals for IL-28B production by way of cell-to-cell dependent and independent mechanisms. In the present study, most of the stimuli to BDCA3 þ DCs for IL-28B production may be the released HCVcc from Huh7.5.1 cells, judging from the inability of suppression with transwells. However, a contribution of contact-dependent mechanisms cannot be excluded in the coculture experiments. HCV genome is transmissible from infected hepatocytes to uninfected ones through tight junction molecules, such as claudin-1 and occuludin. Further investigation is needed to clarify whether such cell-to-cell transmission of viral genome is operated or not in BDCA3 þ DCs.
The relationship between IL-28B expression and the induction of ISGs has been drawing much research attention. In primary human hepatocytes, it is reported that HCV primarily induces IFN-k, instead of type-I IFNs, subsequently enhancing ISG expression. 7 Of particular interest is that the level of hepatic IFN-ks is closely correlated with the strength of ISG response. 26 These reports strongly suggest that hepatic IFN-ks are a crucial driver of ISG induction and subsequent HCV eradication. Besides, it is likely that BDCA3 þ DCs, as a bystander IFN-k producer in the liver, have a significant impact on hepatic ISG induction. In support of this possibility, we demonstrated in this study that BDCA3 þ DCs are capable of producing large amounts of IFN-ks in response to HCV, thereby inducing ISGs in the coexisting liver cells.
Controversial results have been reported regarding the relationship between IL28B genotypes and the levels of IL-28 expression. Nevertheless, in chronic hepatitis C patients with IL-28B major genotype, the IL-28 transcripts in PBMCs are reported to be higher than those with minor genotype. 2 In this study, by focusing on a prominent IFN-k producer (BDCA3 þ DCs) and using the assay specific for IL-28B, we showed that the subjects with IL-28B major genotype could respond to HCV by releasing more IL-28B. Of interest, such a superior capacity of BDCA3 þ DCs was observed only in response to HCV but not to poly IC. Since the pathways downstream of TLR3-TRIF leading to IL-28B in BDCA3 þ DCs should be the same, either HCV or poly IC stimulation, two plausible explanations exist for such a distinct IL-28B response. First, it is possible that distinct epigenetic regulation may be involved in IL-28B gene according to the IL-28B genotypes. Recently, in influenza virus infection, it is reported that micro-RNA29 and DNA methyltransferase are involved in the cyclooxygenase-2-mediated enhancement of IL-29/IFN-k1 production. 27 This report supports the possibility that similar epigenetic machineries could be operated as well in HCVinduced IFN-ks production. Second, it is plausible that the efficiency of the stimulation of TLR3-TRIF may be different between the IL-28B genotypes. Since HCV reaches endosome in BDCA3 þ DCs by way of the CD81-mediated entry and subsequent endocytosis pathways, the efficiencies of HCV handling and enzyme reactions in endosome may be influential in the subsequent TLR3-TRIF-dependent responses. Certain unknown factors regulating such process may be linked to the IL-28B genotypes. For a comprehensive understanding of the biological importance of IL-28B in HCV infection, such confounding factors, if they exist, need to be explored.
In conclusion, human BDCA3 þ DCs, having a tendency to accumulate in the liver, recognize HCV and produce large amounts of IFN-ks. An enhanced IL-28B/IFN-k3 response of BDCA3 þ DCs to HCV in subjects with IL-28B major genotype suggests that BDCA3 þ DCs are one of the key players in anti-HCV innate immunity. An exploration of the molecular mechanisms of potent and specialized capacity of BDCA3 þ DCs as IFN-k producer could provide useful information on the development of a natural adjuvant against HCV infection.