CD81/CD9 tetraspanins aid plasmacytoid dendritic cells in recognition of hepatitis C virus–infected cells and induction of interferon-alpha


  • Potential conflict of interest: Nothing to report.

  • This work was supported by the National Institutes of Health (grant no.: R37AA014372; to G.S.). The authors thank Drs. Charles M. Rice, Takaji Wakita, Christoph Seeger, and Kui Li for kindly providing reagents.

Address reprint requests to: Gyongyi Szabo, M.D., Ph.D., Department of Medicine, University of Massachusetts Medical School, 364 Plantation Street, Worcester, MA 01605. E-mail: fax: 508-856-4770


Recognition of hepatitis C virus (HCV)-infected hepatocyes and interferon (IFN) induction are critical in antiviral immune response. We hypothesized that cell-cell contact between plasmacytoid dendritic cells (pDCs) and HCV-infected cells was required for IFN-α induction through the involvement of cell-surface molecules. Coculture of human peripheral blood mononuclear cells (PBMCs) with genotype 1a full-length (FL) HCV genomic replicon cells or genotype 2a Japanese fulminant hepatitis type 1 (JFH-1) virus-infected hepatoma cells (JFH-1), and not with uninfected hepatoma cells (Huh7.5), induced IFN-α production. Depletion of pDCs from PBMCs attenuated IFN-α release, and purified pDCs produced high levels of IFN-α after coculture with FL replicons or JFH-1-infected cells. IFN-α induction by HCV-containing hepatoma cells required viral replication, direct cell-cell contact with pDCs, and receptor-mediated endocytosis. We determined that the tetraspanin proteins, CD81 and CD9, and not other HCV entry receptors, were required for IFN-α induction in pDCs by HCV-infected hepatoma cells. Disruption of cholesterol-rich membrane microdomains, the localization site of CD81, or inhibition of the CD81 downstream molecule, Rac GTPase, inhibited IFN-α production. IFN-α induction involved HCV RNA and Toll-like receptor (TLR) 7. IFN-α production by HCV-infected hepatoma cells was decreased in pDCs from HCV-infected patients, compared to healthy controls. We found that preexposure of healthy PBMCs to HCV viral particles attenuated IFN-α induction by HCV-infected hepatoma cells or TLR ligands, and this inhibitory effect could be prevented by an anti-HCV envelope glycoprotein 2–blocking antibody. Conclusion: Our novel data show that recognition of HCV-infected hepatoma cells by pDCs involves CD81- and CD9-associated membrane microdomains and induces potent IFN-α production. (HEPATOLOGY 2013;58:940–949)


cytochalasin D


cluster of differentiation


dendritic cell




envelope glycoproteins 1/2


full length


hepatitis C virus


cell-culture–derived HCV particles


human immunodeficiency virus




interferon-stimulated gene


Japanese fulminant hepatitis type 1


low-density lipoprotein


LDL receptor


mitochondrial antiviral-signaling protein




myeloid dendritic cell


nonstructural protein


pathogen associated molecular patterns


peripheral blood mononuclear cell


plasmacytoid dendritic cell

Poly I:C

polyinosinic/polycytidylic acid


RNA interference


standard deviation


small interfering RNA


scavenger receptor B1


single-stranded RNA


tetraspanin-enriched microdomain


Toll-like receptor


TIR-domain-containing adapter-inducing IFN-β

Hepatitis C virus (HCV), a positive-sense single-stranded RNA (ssRNA) virus, infects human hepatocytes.[1] The host immune defense determines successful HCV elimination or chronic infection that often leads to cirrhosis and hepatocellular carcinoma.[2] Type I interferons (IFNs) play a central role in anti-HCV immunity, and pegylated IFN-α plus ribavirin is used clinically for HCV eradication.[3] The process by which innate immune cells recognize HCV-infected hepatocytes is poorly understood. Although HCV infection activates IFN-β synthesis through the RIG-I and Toll-like receptor (TLR)3 pathways, this response is attenuated because of the serine protease activity of HCV nonstructural protein (NS)3/4A that cleaves the adaptor proteins, mitochondrial antiviral-signaling protein (MAVS) and TIR-domain containing adapter-inducing IFN-β (TRIF), respectively.[4, 5] Despite this, IFN-stimulated genes (ISGs) are expressed at high levels in the liver in chronic HCV infection, raising the possibility that other cell types and/or IFNs may participate in the induction of ISGs.[6] Indeed, HCV-infected livers are enriched with plasmacytoid dendritic cells (pDCs), a cell population with a high type I IFN production capacity.[7]

Dendritic cells (DCs) bridge innate and adaptive immunity and play a central role in antiviral defense. Human myeloid dendritic cells (mDCs) express TLR3, which sense double-stranded RNA and produce IFN-β, whereas pDCs are specialized to produce IFN-α in amounts larger than any other cell types.[8] Although TLR7 is highly expressed in pDCs and senses ssRNA within the endosomal compartment,[9] previous studies found that HCVcc (cell-culture–derived HCV particles) alone cannot or weakly induce mDC or pDC activation and HCVcc even inhibited pDC function upon TLR9 stimulation.[10] A recent report showed that pDCs produced type I IFNs through a TLR7-dependent pathway in response to Huh7.5 cells containing replicating HCV RNA.[13] However, it remains to be answered how pDCs recognize HCV hidden in infected hepatocytes to mount an IFN response.

HCV entry is a multistep process where interactions between HCV envelope glycoproteins 1/2 (E1/E2) and glycosaminoglycans contribute to binding of the virus particles to host cells. The low-density lipoprotein (LDL) receptor (LDL-R) has been proposed as a capture molecule. Upon this initial engagement, scavenger receptor B1 (SR-B1), a tetraspanin protein (CD81), together with the tight-junction proteins, claudin-1 and occludin, synergistically contribute to HCV entry.[14, 15] In this study, we evaluated the involvement of cell-surface molecules and found a novel mechanism for recognition of HCV in infected hepatoma cells by pDCs, which involves membrane microdomains formed by CD81- and CD9-associated tetraspanin webs and lipid rafts. We speculate that this novel pathway is critical for the activation of pDCs in the liver-tissue environment in HCV-infected patients.

Materials and Methods

Cells, Replicons, and Viruses

Huh7.5 cells and full-length (FL) genomic replicon cells were kindly provided by C. Rice, subgenomic replicons (BB7) by C. Seeger, and Japanese fulminant hepatitis type 1 (JFH-1)/HCV by T. Wakita. All cells were maintained in low-glucose Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 10 μg/mL of ciprofloxacin, and supplemented with nonessential amino acids. Additionally, both replicon cells were maintained in the presence of 500 μg/mL of G418. JFH-1 genomic RNA was in vitro transcribed from the linearized pUC-vJFH-1 plasmid as a template using an in vitro transcription kit (MEGAscript; Ambion Biotechnology, Austin, TX). Transfection of in vitro synthesized JFH-1 RNA constructs and production of JFH-1 virus stock were performed as previously described.[13]


Neutralizing SR-B1 and DC-SIGN antibodies were from R&D Systems (Minneapolis, MN), LDL-R antibody was from Novartis (Basel, Switzerland), and anti-E2 antibody (MBL-HCV1) was a generous gift from MassBiologics (Boston, MA).[16] Anti-CD81 (JS-81) was purchased from BD Biosciences (Franklin Lakes, NJ), and anti-CD9 and anti-CD63 were from BioLegend (San Diego, CA). See the Supporting Materials for information on other reagents.

Preparation of Human Peripheral Blood Mononuclear Cells and pDCs and Cocultures

Human peripheral blood mononuclear cells (PBMCs) and pDCs were isolated from peripheral blood from healthy adult human volunteers or HCV-infected patients after informed consent was obtained and according to procedures approved by the Committee for Protection of Human Subjects in Research at the University of Massachusetts Medical School (Worcester, MA). Descriptions of cell isolation, cocultures, RNA extraction, and analysis are described in the Supporting Materials.

RNA Interference Knockdown

Two predesigned short interfering RNAs (siRNAs) for CD81 (S2722 and S2723) and one negative control siRNA (4390843) were used in this study. RNA interference (RNAi) knockdown was performed using the RNAiMAX system (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Two days after siRNA transfection, cells were collected for flow-cytometric analysis or coculture experiments.

Statistical Analysis

All data are from three or more independent experiments. Data are presented as mean ± standard deviation (SD) using the Student t test or Mann-Whitney's test according to data distribution. P < 0.05 was considered to be statistically significant.


Human PBMCs and Purified pDCs Produce IFN-α in Response to HCV-Infected Cells

Because immune cells can directly interact with hepatocytes in the liver, we evaluated whether coculture of human PBMCs with HCV-infected hepatoma cells could induce IFN-α production. PBMCs produced IFN-α in response to HCV FL replicon or JFH-1-infected Huh7.5 cells, whereas uninfected Huh7.5 cells or subgenomic HCV-replicons (BB7) failed to induce IFN-α production (Fig. 1A). There was no IFN-α production in HCV-infected Huh7.5 cells or in HCV-exposed PBMCs in the absence of hepatoma cells (data not shown). PBMC stimulation with the TLR9 ligand (CpG-A) was the positive control for IFN-α induction (Fig. 1A). pDCs produced large amounts of IFN-α when cocultured with HCV-infected hepatoma cells, whereas depletion of pDCs significantly reduced IFN-α production in PBMCs (Fig. 1B). Flow-cytometry analysis revealed an intracellular IFN-α increase in the pDC-gated populations in response to FL replicons, but not Huh7.5 cells or BB7 replicons (Fig. 1C). TLR7/8 or TLR9 ligands induced strong intracellular IFN-α expression in pDCs (Fig. 1C).

Figure 1.

IFN-α is induced in human PBMCs by HCV-infected cells. (A) Huh7.5, HCV replicon (BB7 or FL), or JFH-1-infected Huh7.5 cells were cocultured with human PBMCs; CpG-A (2 μM) was a positive control. (B) Huh7.5, HCV replicon, or JHF-1-infected cells were cocultured with pDCs or PBMCs depleted of pDC for 24 hours; IFN-α in the supernatants was measured by enzyme-linked immunosorbent assay. (C) Enriched pDCs were cocultured with Huh7.5 or HCV replicon cells or TLR agonists (CpG-A 2 μM, R837 2 μg/mL, R848 2 μg/mL, or garquimod 2 μg/mL), and GolgiStop (BD Biosciences, Franklin Lakes, NJ) was added after 6 hours. Ten hours later, intracellular IFN-α expression was determined by flow cytometry. n = 5, mean values ± SD (A and B); *P < 0.01 versus PBMC alone, PBMC+Huh7.5, or BB7 group in (A); *P < 0.01 in (B).

Cell-to-Cell Contact Between Human PBMC and HCV-Infected Cells Is Required for IFN-α Induction

Induction of IFN-α in pDCs by HCV-infected hepatoma cells may involve pathogen-associated molecular patterns (PAMPs), secreted mediators, and/or cell-cell interactions. We found that in contrast to live cells, lysates or staurosporine-treated apoptotic FL replicons could not elicit IFN-α induction, suggesting that live, intact cells, rather than their content, induced IFN-α in PBMCs (Fig. 2A,B). We determined a requirement for cell-to-cell contact for pDC activation by HCV-infected cells using transwell inserts separating PBMCs from HCV-infected cells and completely abolishing IFN-α production (Fig. 2C). Transwell separation did not affect CpG-induced IFN-α production (Fig. 2C).

Figure 2.

IFN-α induction requires cell-to-cell contact between pDCs and live HCV-infected cells. (A) Huh7.5, HCV replicon cells, or lysates from equivalent number of cells were cocultured with human PBMCs for 24 hours; CpG-A (2 μM) was a positive control. (B) Staurosporine was added to Huh7.5 or HCV replicon cells for 8 hours, cells were washed and cocultured with human PBMCs for 24 hours, and (C) transwell insert was introduced to cocultures, where indicated; n = 4, mean values ± SD (A-C); *P < 0.01 in (B). IFN-α in the supernatants was measured by enzyme-linked immunosorbent assay.

CD81 and CD9 Tetraspanins Are Involved in Recognition of HCV-Infected Hepatoma Cells by pDCs

Because only HCV FL or JFH-1-infected, and not subgenomic, replicons induced IFN-α in PBMCs requiring cell-to-cell contact, we hypothesized that cell-surface molecules were involved in the recognition of HCV structural proteins. HCV E1/E2 and selective host membrane receptors mediate viral attachment and entry.[14, 15] We found that neutralizing anti-SR-B1, -LDL-R, -DC-SIGN, and -HCV E2 antibodies did not prevent IFN-α induction in PBMCs by HCV-infected hepatoma cells (Fig. 3A,B). In contrast, the addition of an anti-CD81 antibody significantly inhibited IFN-α production in PBMCs in response to HCV-infected cells (Fig. 3B). IFN-α induction by TLR7/8 or TLR9 ligands was not affected by the anti-CD81-blocking antibody (Fig. 3B). Timing of anti-CD81 administration relative to coculture was critical in the inhibition of IFN-α production. The introduction of anti-CD81 up to 3 hours after the initial cell-cell contact prevented IFN-α induction, but the addition at later time points (>8 hours) failed to inhibit pDC activation (Fig. 3C). The inhibitory effect of anti-CD81 was dose dependent (Fig. 3D).

Figure 3.

Blocking CD81 inhibits pDC IFN-α induction by HCV replicons or JFH-1-infected cells. During coculture, (A) blocking anti-SR-B1, LDL-R, and DC-SIGN antibodies or (B) anti-HCV E2 antibody (5 μg/mL) or anti-CD81 antibody (5 μg/mL) were added. CpG-A (2 μM), R848 (2 μg/mL), CL097 (2 μg/mL), and garquimod (1 μg/mL) were included as controls. Results shown in (C and D) were from cocultures between isolated human pDC and FL cells. (C) Anti-CD81-blocking antibody was added at the indicated time points after cocultures; IFN-α was measured 24 hours later by enzyme-linked immunosorbent assay. (D) A different concentration of anti-CD81 antibody was used; n = 3, mean values ± SD.

Beyond serving as a receptor for HCV, CD81 has a role in cell communication through the formation of a tetraspanin web that allows an association with other protein partners.[17] We found that both human pDCs and HCV-infected hepatoma cells expressed high levels of surface CD81 (Supporting Fig. 1). Forty-eight hours after CD81/siRNA transfection, CD81 expression was decreased in hepatoma cells, compared to control siRNA, and HCV-NS3 expression was not affected (Supporting Fig. 2); however, CD81/siRNA knockdown in hepatoma cells did not inhibit IFN-α (Fig. 4A) or MxA induction (Fig. 4B) in PBMCs. Because siRNA knockdown could not fully prevent CD81 expression in hepatocytes, we utilized an anti-CD81-blocking antibody and found that pretreatment of hepatoma cells or pDCs with anti-CD81-blocking antibody could prevent IFN-α induction by HCV-infected hepatoma cells (Fig. 4C).

Figure 4.

Inhibition of CD81 on hepatoma cells does not affect IFN-α production by PBMCs. (A and B) Anti-CD81 siRNA was transfected into Huh7.5, HCV replicon, or JFH-1-infected cells 48 hours before their coculture with PBMCs for 12 hours; RNA was extracted from cells for ISG detection, and supernatants were collected at 24 hours for IFN-α measurement by enzyme-linked immunosorbent assay (ELISA). (C) pDC or FL cells were pretreated with anti-CD81 antibody separately for 1 hour, respectively, then washed and cocultured for 24 hours; IFN-α was measured by ELISA; n = 3, mean values ± SD. *P < 0.01 versus other groups in (C).

CD81, a member of the tetraspanin protein family, associates with other tetraspannin members (especially CD9).[18] Human pDCs highly expressed CD81 and CD9, and not CD63, and JFH-1-infected hepatoma cells had high CD81 and low CD63 and CD9 expression (Fig. 5A). Both anti-CD9 and -CD81 antibodies significantly inhibited, whereas anti-CD63 had no effect on, IFN-α production by pDC in response to HCV-infected hepatoma cells (Fig. 5B and Supporting Fig. 4). Furthermore, CD9 and CD81 inhibition had additive negative effects on IFN-α production (Fig. 5C), indicating the importance of membrane structures formed by CD9 and CD81 in pDC activation.

Figure 5.

CD9 is expressed on pDCs, and CD9 blocking inhibits IFN-α induction by HCV-infected hepatoma cells. (A) PBMCs, purified pDCs, or JFH-1-infected Huh7.5 cells were stained with fluorescein isothiocyanate–conjugated anti-CD9, -CD63, and -CD9 or control antibodies and analyzed by flow cytometry. (B) PBMCs or enriched DCs were cocultured with JFH-1-infected Huh7.5 cells in the presence of anti-tetraspanin protein or control antibodies. Supernatants were collected 24 hours later for IFN-α measurement. n = 5, mean values ± SD. *P < 0.05 versus control groups in (B and C); #P < 0.05 versus anti-CD81 and -CD9 group in (C).

CD81 Molecule-Associated Lipid Rafts Expressed in pDCs Play a Role in Acquiring Ligands From HCV-Infected Cells Through Endocytosis

Tetraspanin-enriched microdomains (TEMs) are platforms for the interaction of membrane-associated proteins and their signal transduction.[19] CD81 and CD9 are heavily palmitoylated and directly bind to cholesterol.[20] Treatment of cells with methyl-beta-cyclodextrin (MβCD) depletes cholesterol, reduces CD81 oligomerization,[18] and disrupts lipid rafts, which are the platforms for CD81 interactions and TEM function. We found that disruption of lipid rafts in PBMCs with MβCD pretreatment significantly decreased IFN-α production in response to HCV FL replicons or JFH-1-infected cells, whereas pretreatment of hepatoma cells with MβCD had no effect (Fig. 6A). The reduction in IFN-α was not the result of cytotoxicity by MβCD, because disruption of lipid rafts did not inhibit IFN-α induction by TLR7/8 or TLR9 ligands (Fig. 6C). To dissect the inhibitory effect of anti-CD81 treatment in hepatoma cells and pDCs, CD81 knock-down hepatoma cells (CD81null cells) were cocultured with PBMCs in the presence of anti-CD81 or control antibody. The inhibitory effect of anti-CD81 was not influenced by CD81 expression on hepatoma cells, indicating that CD81 on pDCs played the role in the recognition of HCV-infected cells (Fig. 6B).

Figure 6.

CD81 protein, lipid rafts, and endocytosis pathways are required for IFN-α induction in PBMCs and hepatoma cocultures. (A) PBMCs or JFH-1-infected Huh7.5 cells were pretreated with MβCD (10 mM), respectively, and cocultured, as indicated. (B) Hepatoma cells were transfected with anti-CD81/siRNA or control siRNA for 48 hours, then cocultured with human PBMCs in the presence of anti-CD81 or control antibody. IFN-α was measured in the supernatants. (C-E) Human PBMCs were treated with MβCD (10 mM), CCD (10 μM), DMA (50 μM), or choloroquine (10 μM) for 30 minutes at 37°C, where indicated, and, after washing, were cocultured with HCV replicon or JFH-1-infected cells. CpG-A (5 μM), R848 (2 μg/mL), CL097 (2 μg/ml), and garquimod (1 μg/mL) were included as controls. Twenty-four hours later, IFN-α was measured in the supernatants. n = 3, mean values ± SD. *P < 0.05 versus other groups in (A), P < 0.01 versus MβCD pretreatment groups in (C) and versus chrloroquine pretreatment groups in (E).

We defined that pretreatment of human PBMCs with cytochalasin D (CCD), an inhibitor of endocytosis, prevented, whereas inhibition of macropinocytosis by dimethylamiloride (DMA) had no effect on, IFN-α induction in pDCs (Fig. 6D). These data suggested that pDCs utilize a CD81-dependent mechanism to acquire ligands from HCV-infected cells through endocytosis. We confirmed the requirement for endosomal acidification because chloroquine treatment of PBMCs prevented IFN-α induction by FL replicons and JFH-1-infected cells (Fig. 6E).

Rac GTPase Is Involved in pDC-IFNα Induction by HCV-Infected Hepatoma Cells

Recent studies suggest that CD81 regulates cellular movements through modulating downstream cytoskeleton regulators, such as Syk or the small GTPases, Rac and Rho.[21] Using specific inhibitors of Syk, Rac, and Rho, we found that only the Rac inhibitor (NSC-23766) prevented pDC IFN-α production, whereas inhibition of Syk (BAY61-3606) and Rho (CT04) had no effect (Fig. 7A). The Rac, but not Syk and Rho, inhibitor also prevented PBMC IFN-α induction by polyinosinic/polycytidylic acid (Poly I:C), R848, and CpG-A (Fig. 7A). The inhibitory effect of NSC-23766 on pDC activation was dose dependent. Low concentration of NSC-23766 pretreatment (20 and 60 μm) resulted in specific inhibition of PBMC IFN-α induction by HCV-infected hepatoma cells and showed a minimal effect on IFN-α induction by Poly I:C, R848, and CpG-A stimulation (Fig. 7B).

Figure 7.

Rac GTPase inhibition diminishes IFN-α induction in pDCs by HCV-infected hepatoma cells. (A) PBMCs were pretreated with Bay61-3606 (20 nM), NSC-23766 (200 μM), or CT04 (1 μg/mL) for 2 hours before coculture with JFH-1-infected Huh7.5 cells or stimulation with Poly I:C (20 μg/mL), R848 (2 μg/mL), and CpG-A (2 μM). (B) PBMCs were pretreated with different doses of NSC-23766 for 2 hours before coculture with JFH-1-infected Huh7.5 cells or stimulation with Poly I:C (20 μg/mL), R848 (2 μg/mL), and CpG-A (2 μM). IFN-α was determined by enzyme-linked immunosorbent assay.

IFN-α Production by PBMCs of HCV-Infected Patients in Response to HCV-Infected Hepatoma Cells Is Decreased

We and others have reported that pDCs from chronic HCV-infected patients have impaired IFN-α production in response to viral PAMPs.[29, 30] Here, we show that compared to healthy controls, PBMCs from treatment-naïve patients with chronic HCV infection had significantly decreased IFN-α production in response to HCV-infected hepatoma cells (Fig. 8A). Because previous studies found that HCV viral particles could inhibit CpG-induced IFN-α production in pDCs,[11, 12] we hypothesized that HCV viral particles might inhibit pDC activation. Pretreatment of healthy PBMCs with infectious HCV particles significantly decreased IFN-α production in response to HCV-infected hepatoma cells or CpG-A stimulation, compared to HCV-naïve PBMCs (Fig. 8B).

Figure 8.

Free HCV virus inhibits pDC IFN-α induction in human PBMCs and hepatoma cocultures. (A) PBMCs from healthy controls (NC-PBMC) or nontreated HCV-infected patients (HCV-PBMC) were cocultured with JFH-1-infected Huh7.5 cells. (B and C) Healthy PBMCs were pretreated with noninfectious (Huh7.5 Sup) or infectious supernatants (JFH-1 Sup) for 2 hours before coculture with JFH-1-infected Huh7.5 cells. Neutralizing anti-HCV E2 antibody (α-E2) was added during pretreatment (C). IFN-α was measured in supernatants; n = 3, mean values ± SD. *P < 0.05 versus other groups in (C).

Because HCV E1/E2 proteins interact with CD81, we wondered whether the inhibitory effect of HCV pretreatment was the result of interference with CD81 function. We found that a neutralizing anti-E2 antibody that prevents binding of HCV to CD81 reversed the inhibitory effect of the infectious HCV-containing supernatants on PBMC IFN-α production in response to HCV-infected hepatoma cells (Fig. 8C).


Type I IFNs play a critical role in the host response that determines the outcome of HCV infection.[6] The mechanisms that define immune recognition of HCV-infected hepatocytes in the liver are still poorly understood. Here, we report that human pDCs interact with HCV-infected hepatoma cells and, in response, produce IFN-α. We present novel evidence for an intimate cross-talk between the pDC and infected hepatocma cell that requires direct cell-cell contact and endocytosis. We show, for the first time, that IFN-α induction in pDCs by HCV-infected cells depends on CD81/CD9 tetraspanins as well as the integrity of membrane lipid-rafts in pDCs. Our results demonstrate that the CD81-mediated pDC and hepatoma cell interaction and IFN-α induction require Rac-GTPase activity. Our novel data demonstrate that exposure of normal PBMCs to HCV virus can attenuate their IFN-α production in response to HCV-infected hepatoma cells or TLR ligands, suggesting a mechanism that may inhibit pDCs in vivo in a HCV-infected host. These data indicate that pDC can sample HCV-infected cells to recognize infection and mount an antiviral immune response with robust IFN-α production, but, likely, this process is attenuated in pDCs during acute and/or chronic HCV infection.

Plasmacytoid DCs produce large amounts of IFN-α as well as capture and cross-present antigens to trigger adaptive immune responses.[24] Studies showed that human immunodeficiency virus (HIV)-infected cells triggered more IFN-α production in pDCs than HIV particles.[25] Here, we present new evidence that pDCs sample live HCV-infected cells through cell-cell contact and endocytic pathways involving CD81- and CD9-associated cell-surface structures, but soluble HCV antigen, dying cells, or secreted HCV viral particles had no IFNα-inducing capacity. We found that HCV FL cells, but not subgenomic replicon, induced significant IFN-α production from pDCs, suggesting that HCV structural protein(s) played a role in this process. A recent report showed that HIV-1-transfected cells carrying defective envelope or fusion proteins, but not nucleocapsid protein, induced pDC activation.[26] This, together with our observations, may suggest that the requirement for viral envelope proteins is a common element in sensing infected cells harboring RNA virus, whereas the proper assembly of the virus is not necessary. We suspect that envelope proteins might affect innate sensing though modulating viral PAMP structure or transportation of PAMPs in the infected cells. However, investigation of the exact mechanisms of HCV envelope or other hepatocyte proteins in DC activation is beyond this study.

Considering the large size of hepatoma cells, it is unlikely that pDCs can take up an entire infected hepatoma cell during their close interaction, but rather capture viral components through the endocytosis pathway. Our results suggested that viral RNA causes pDC activation. Consistent with a previous report,[13] we confirmed that a TLR7-inhibitory oligodeoxynucleotide (IRS661) strongly inhibited pDC/IFN-α production, whereas the DNA-sensing receptor, the TLR9 pathway, remained unaffected (Supporting Fig. 3). In vitro synthesized HCV RNA segments or whole JFH-1 RNA triggered strong pDC/IFN-α induction, whereas extracted whole RNA from Huh7.5 cells failed to induce IFN-α production (Supporting Fig. 5). Finally, the magnitude of IFN-α induction in pDCs depended on the intracellular HCV RNA levels in infected cells, and eradication of HCV prevented pDC/IFN-α production (Supporting Fig. 6).

CD81 was expressed on both HCV-infected hepatoma cells and human pDCs. We found that siRNA knockdown of CD81 in hepatocytes failed to prevent pDC activation, whereas pretreatment of pDC with anti-CD81 antibody significantly inhibited IFN-α induction by HCV-infected hepatoma cells, suggesting a specific role of CD81 on pDCs, and not on HCV-infected cells, in the recognition process. CD81 is a member of the tetraspanin family.[19] Tetraspanin proteins interact with one another, membrane receptors, and signaling proteins to form functional structures in cell membranes (i.e., TEMs), which play important roles in migration, proliferation, differentiation, and infectious disease.[17] We determined that CD9, a CD81 partner protein, was expressed on pDCs and contributed to the interactions between pDCs and HCV-infected hepatoma cells, because anti-CD9 antibody prevented IFN-α induction. We found evidence for an additive cooperation between CD9 and CD81 in IFN-α induction by HCV-infected hepatoma cells. Cholesterol physically and functionally associates with TEMs[18, 20] and MβCD, which depletes cholesterol, disrupts CD81 oligomerization, and interferes with TEM function.[18] Similar to anti-CD81 antibody treatment, pretreatment of pDCs with MβCD prevented IFN-α induction by HCV-infected hepatoma cells, indicating that the intact function of the CD81-associated cell-membrane structures in pDCs was necessary for the recognition of HCV-infected cells.

As a result of its short cytosolic tail, CD81 cannot conduct signal transduction by itself and interacts with partner proteins; however the molecular mechanisms downstream of CD81 are yet to be defined.[15, 17] CD81 is important in cellular movement in human DCs through regulating small GTPases, such as Rho and Rac.[21] We hypothesized that CD81-mediated cell-cell interactions between pDCs and hepatoma cells involved Rho or Rac activities in pDCs, and found that inhibition of Rac activity prevented IFN-α production by pDCs. The effect of Rac inhibition was similar to CD81 blocking because they both specifically inhibited IFN-α induction in pDCs by HCV-infected hepatoma cells, and not by TLR agonists. Thus, we speculate that pDC and HCV-infected hepatoma interact through CD81, which signals by involving Rac activity in pDCs to induce IFN-α production.

Although impaired IFN-α induction by TLR-agonists in pDCs from HCV-infected patients have been demonstrated before,[27, 28] our novel data show impaired IFN-α production by pDCs of patients with chronic HCV infection in response to HCV-infected hepatoma cells. Previous studies revealed several possible explanations for the decreased IFN-α production capacity in pDCs of HCV-infected patients, including decreased number of circulating pDCs and functional impairment of the IFN-α-synthesis pathway.[29] In this study, we reproduced the decreased IFN-α production observed in patients by preincubating healthy pDCs with infectious HCV virions and found that anti-HCV E2 antibodies prevented this inhibition. It is tempting to speculate that HCV virion interaction between E2 and CD81 on pDCs may interfere with subsequent CD81-mediated cell-cell interaction, leading to attenuated IFN-α induction by HCV-infected cells. This finding also implicates that a large amount of released HCV virions or paticles during acute HCV infection might interfere with the innate sensing mechanisms of pDCs and lead to inefficient viral control as well as promote chronic infection. Consistent with this hypothesis, previous reports showed that HCV virions inhibited TLR9-, and not TLR7-induced, pDC activation[11] and that HCV E2 and CD81 interaction were associated with altered trafficking of DCs[32] and inhibition of natural killer cell activation.[33, 34]

In conclusion, we identified a novel mechanism for the recognition of HCV-infected hepatoma cells by pDCs involving membrane microdomains formed by CD81-associated tetraspanin webs and lipid rafts. We speculate that this novel pathway is critical in the activation of pDCs and the type I IFN response in the liver environment in HCV-infected patients. Our data also suggest an intimate contact between immune cells and infected hepatocytes, allowing “sampling” of the pathogenic signals to maintain host integrity.