Human Type 2 Myeloid Dendritic Cells Produce Interferon-λ and Amplify Interferon-α in Response to Hepatitis C Virus Infection

a

I

Hpt

fsd

lm

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER

GASTROENTEROLOGY 2013;144:414–425

Human Type 2 Myeloid Dendritic Cells Produce Interferon-� and AmplifyInterferon-� in Response to Hepatitis C Virus InfectionSHUYE ZHANG,1 KAREN KODYS,1 KUI LI,2 and GYONGYI SZABO1

1Department of Medicine, University of Massachusetts Medical School, Worcester, Massachusetts; and 2Department of Microbiology, Immunology and Biochemistry,
University of Tennessee Health Science Center, Memphis, Tennessee
t

tut

t

dostc

BACKGROUND & AIMS: The type III interferons (IFN-�s: interleukin [IL]-28a, IL-28b, and IL-29) have importantroles in hepatitis C virus (HCV) infection, but little isunderstood about what cells produce these cytokines orhow production is activated. We investigated whetherhuman immune cells recognize HCV-infected cells andrespond by producing IFN-�. METHODS: We culturedhealthy human peripheral blood mononuclear cells(PBMCs) with different populations of immune cells andJapanese fulminant hepatitis-1 (JFH-1) HCV-infectedHuh7.5 (cell culture– derived HCV particles [HCVcc]/Huh7.5) cells. RESULTS: Human PBMCs recognizedHCVcc/Huh7.5 cells and responded by producing IFN-�,IFN-�, and IFN-�. A rare subset of myeloid dendritic cells(mDCs), which are blood DC antigen (BDCA)� (alsocalled mDC2 cells), were the major source of IL-28 andIL-29 production in response to HCVcc/Huh7.5 cells.Plasmacytoid DCs produced IFN-�, whereas natural killer

nd natural killer T cells were the main source of IFN-�production in co-culture experiments. Of the endosomalToll-like receptors (TLRs)3, 7, 8, and 9, only TLR3 ordouble-stranded HCV RNA induced production of IL-28and IL-29 by mDC2s; endosomal maturation was re-quired. Production of IFN-� and IFN-� were linked—IFN-� increased production of IFN-� by plasmacytoidDCs and IFN-� significantly increased production ofFN-�. CONCLUSIONS: mDC2s are a major source of

IFN-� production by PBMCs in response to HCVcc/uh7.5 cells. mDC2s are activated through the TLR3

athway, indicating that human DCs efficiently can ini-iate an immune response against HCV infection. IFN-�

therefore has an important role in HCV infection.

Keywords: IL-28B SNP; NK Cells; NKT Cells; Viral ImmuneRegulation.

Hepatitis C virus (HCV) infection affects 120 millionpeople worldwide, typically resulting in chronic in-

ection. Early and high induction of interferon (IFN)-timulated genes (ISGs) in HCV infection is a marker ofecreased response to therapy,1 however, the coordinated

role of IFNs during HCV infection and the signalingpathways leading to their production are only partiallyunderstood. The IFNs are divided into type I (IFN-� and�), type II (IFN-�), and type III subtypes (IFN-�s: inter-eukin [IL]-28a, IL-28b, and IL-29) and have immuno-

odulatory and antiviral activities through their respec-

ive receptors that induce ISGs.2 All IFNs contribute tothe coordination of antiviral immunity in the control ofHCV. IFN-� has anti-HCV activity and is widely used toreat chronic HCV infection.3 IFN-� is produced by nat-ral killer (NK) cells or HCV-specific T cells and orches-rates anti-HCV immune responses.4 IFN-�s may play a

unique role in clearing HCV infection because severalsingle-nucleotide polymorphisms (SNPs) near the IL-28bregion predict the outcome of natural HCV infection orIFN-� therapy.5– 8 Despite using different receptors, IFN-�shares a common downstream Janus kinase-Signal trans-ducer and activator of transcription (JAK-STAT) pathwaywith type I IFNs and induces very similar antiviral im-mune responses.2 IFN-� is currently in clinical trials forreatment of HCV infection.9

Although recent reports have suggested that humanhepatocytes produce type I and III IFNs in response toHCV infection,10,11 it also is well recognized that HCV

isrupts IFN synthesis though NS3-4 –mediated cleavagef 2 crucial adaptor proteins, mitochondrial antiviral-ignaling protein and Toll/IL-1 resistance-domain-con-aining adapter-inducing interferon-�, in infected hepato-ytes.12 Thus, it is crucial for host immune cells to

recognize ongoing HCV infection and initiate an immuneresponse. Human dendritic cells (DCs) bridge the innateand adaptive immune systems and play an irreplaceablerole in immune defense against viral infections. Accordingto their distinct phenotype and functional characteristics,human peripheral DCs can be divided into 3 subsets:myeloid CD1c� DC (mDC1), myeloid CD141� DC ormyeloid blood DC antigen (BDCA)3� DC (mDC2), andplasmacytoid DC (pDC).13 The mDC1 (conventional DC)

Abbreviations used in this paper: BDCA, blood DC antigen; CpG,cytosine-phosphate-guanosine; CXCL, CXC-chemokine ligand; DC, den-dritic cell; dsRNA, double-strand RNA; HCV, hepatitis C virus; HCVcc, cellculture–derived hepatitis C virus particles; IFN, interferon; IL, interleu-kin; IP, interferon-inducible protein; ISG, interferon-stimulated gene;JFH-1, Japanese fulminant hepatitis-1; mDC, myeloid dendritic cell;mRNA, messenger RNA; MX1, myxovirus resistance 1; NK, naturalkiller; NKT, natural killer T; PBMC, peripheral blood mononuclear cell;pDC, plasmacytoid dendritic cell; PHH, primary human hepatocytes;PRR, pattern recognition receptor; RANTES, regulated on activation,normal T cell expressed and secreted; RLR, RIG-I like receptor; SNP,single-nucleotide polymorphism; ssRNA, single-strand RNA; TLR, Toll-like receptor; TRAIL, TNF-related apoptosis-inducing ligand.

© 2013 by the AGA Institute0016-5085/$36.00

http://dx.doi.org/10.1053/j.gastro.2012.10.034


i

pnwHI

rs[

aPsni

ld

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER

February 2013 HUMAN MDC2 RECOGNIZE HCV AND PRODUCE IFN–� 415

represents the largest mDC population in the blood thatproduces inflammatory cytokines and chemokines.14 ThemDC2, a minor subset, is the human homologue of themouse CD8� DC subset,15,16 which produces IL-12 andcross–presents antigen for CD8 class 1 restricted cytotoxic Tlymphocyte responses under Toll-like receptor (TLR)3 liga-tion.17 pDCs, natural type 1 IFN-producing cells, producehigh levels of IFN-� and have a central role in antiviralmmune responses.18

Because HCV cannot effectively infect immune cells,19

it is speculated that immune cells must recognize HCVvirions and/or HCV-infected cells. Previous studies havefound that cell culture– derived HCV particles (HCVcc)cannot or only weakly induce mDC or pDC activation.20,21

A recent report showed that pDCs can be triggered toproduce type I IFNs by Huh7.5 cells containing replicat-ing HCV RNA.22 We hypothesized that human immunecells can recognize HCV-infected hepatoma cells and pro-duce IFNs or inflammatory cytokines in response. We foundthat all 3 types of IFNs were produced in co-cultures ofhuman peripheral blood mononuclear cells (PBMCs) andHCV-infected hepatoma cells without significant inflamma-tory cytokine induction. We determined that induction ofthe different IFNs by HCV-infected hepatoma cells was cell-specific and involved different pattern recognition receptors(PRRs). We identified mDC2s as the main producer of IFN-�and confirmed IFN-� production by pDCs and IFN-� pro-duction by NK/natural killer T (NKT) cells. These findingsbroaden our view of anti-HCV immunity and reveal uniqueroles for different human dendritic cells in initiating im-mune responses in HCV infection.

Materials and MethodsCells, Replicons, and VirusesJapanese fulminant hepatitis-1 (JFH-1) genomic RNA

was derived from pUC-vJFH-1 plasmid using in vitro transcrip-tion. JFH-1 virions were produced by JFH-1-RNA–transfectedHuh7.5 cells. Detailed information and protocols are describedin the Supplementary Materials and Methods section.

Preparation of Human PBMCs, DC Subsets,and Co-culture ExperimentsHuman PBMCs were isolated from peripheral blood

from normal human volunteers and chronic HCV-infected pa-tients after informed consent was obtained according to proce-dures approved by the Committee for Protection of HumanSubjects in Research at the University of Massachusetts MedicalSchool. Detailed protocols for DC isolation and co-culture ex-periments are described in the Supplementary Materials andMethods section.

Reagents, Bioplex, Enzyme-LinkedImmunosorbent Assay, RNA Quantification,Flow Cytometry, HCV Double-Stranded RNAGeneration, IL-28B SNP Genotyping, andOther Methods and Statistical AnalysisSee the Supplementary Materials and Methods section

for more detail.

ResultsHCV-Infected Cells Induce Type I, II, and IIIIFNs in Human PBMCs

We hypothesized that HCV-infected hepatomacells represent danger signals and stimulate interferonsand inflammatory cytokines in human PBMCs. By usingco-cultures between normal human PBMCs and HCVcc/Huh7.5 or control Huh7.5 cells we found that all 3 typesof IFNs (IFN-�, IFN-�, and IFN-�s [IL-28 and IL-29]) were

roduced in the presence of HCVcc/Huh7.5 but not inoninfected or apoptotic Huh7.5 cells (Figure 1A). Thereas no IFN production by PBMCs, Huh7.5 cells, orCVcc/Huh7.5 cells alone (data not shown). IFN-�,

FN-�, IL-28, and IL-29 induction was highest in thepresence of HCVcc/Huh7.5 cells, with the highest percent-age of HCV infection (87% HCV infected) and the extentof IFN induction directly correlated with the percentageof HCV-infected Huh7.5 cells (Figure 1B). We found asignificant increase in the levels of interferon-induciblegene products, interferon-inducible protein (IP)-10(CXC-chemokine ligand [CXCL] 10), and TNF-relatedapoptosis-inducing ligand (TRAIL) but no induction ininflammatory cytokines (IL-1�, IL-6, IL-8, IL-10, IL-12,egulated on activation, normal T cell expressed andecreted [RANTES], and tumor necrosis factorTNF]-�) in co-cultures with HCVcc/Huh7.5 cells (Sup-

plementary Figure 1). We also confirmed the inductionof CXCL10 and no induction of IL-1B, IL-6, IL-8, IL-10, andTNF-� at messenger RNA (mRNA) levels (Supplement-

ry Figure 2). Addition of concentrated JFH-1 virions toBMCs resulted in no IFN-�, -�, or -� induction (data nothown), suggesting that HCV-infected hepatoma cells andot the isolated virus induced all 3 types of IFNs without

nducing inflammatory cytokines.

IL-28, IL-29, and Myxovirus Resistance 1Gene Are Induced Rapidly From Co-culturesBetween Human PBMCs and HCV-InfectedCells

Time-course experiments revealed that IL-28, IL-29, and ISG (myxovirus resistance 1 [MX1]) genes wereinduced rapidly in co-cultures between human PBMCsand HCVcc/Huh7.5 cells, but not with noninfectedHuh7.5 cells (Figure 1C). None of the type I or II IFNs(including IFNA1, IFNB, and IFNG) genes were inducedat early time point (4 hours), indicating that the earlytype III IFN response against HCV infection was inde-pendent of type I/II IFN production. We found in-creased protein levels of IL-28 and IL-29 at 8 hours afterco-culture, although IFN-� and IFN-� levels remainedow (Figure 1D). Furthermore, IP-10, another IFN-in-ucible gene, was detected at 4 hours, whereas IFN-�

was detectable only at 8 hours after co-culture (Figure1D), indicating the existence of IP-10 inducers that were
different from pDCs and IFN-�.

tttd

aw

W(c

cewaac

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER

416 ZHANG ET AL GASTROENTEROLOGY Vol. 144, No. 2

Type I, II, and III IFNs Are Produced byDifferent Immune Cell Subsets in Response toHCV-Infected CellsA recent report showed that pDCs produce

IFN-� in response to HCV-infected cells.22 To identifyhe cell types responsible for IFN-� and IFN-� produc-ion in the PBMC and HCVcc/Huh7.5 co-cultures, weested isolated pDCs, PBMCs depleted of pDCs, PBMCsepleted of CD56� cells, and CD56� cells (NK/NKT

cells). We found that isolated pDCs could recognizeHCV-infected cells and produce both IFN-� and IL-29,but very little IL-28 and not IFN-�. IFN-� release wasdiminished when HCVcc/Huh7.5 cells were co-cultured

Figure 1. All 3 types of IFNs (IFN-�, IFN-�, and IFN-�s [IL-28 and IL-ells. (A) Huh7.5 cells or HCVcc/Huh7.5 cells were co-cultured winzyme-linked immunosorbent assay (mean � SD; n � 6). (B) HCVccere co-cultured with human PBMCs for 24 hours. IFN levels were mnd every dot represents one data point. (C) Human PBMCs and hepnd 16 hours later. Total RNA was extracted from co-cultured cells, ahain reaction, using glyceraldehyde-3-phosphate dehydrogenase as

cells were co-cultured with PBMCs. Supernatants were collectedenzyme-linked immunosorbent assay (mean � SD; n � 3– 6).

with PBMCs depleted of pDC (PBMC-pDC), confirming

that pDCs were fully responsible for IFN-� production(Figure 2A). IFN-�, IL-28, and IL-29 still were produced

fter pDC depletion, although at lower levels comparedith PBMC co-cultures, indicating that IFN-�, IL-28,

and IL-29 production can occur in the absence of pDCs.NK/NKT cells are a major source of IFN-� production.

e found that PBMCs depleted of either NK/NKT cellsPBMC-CD56) or pDCs resulted in a significant de-rease in IFN-� production from co-cultures with

HCVcc/Huh7.5 cells. Interestingly, co-culture of puri-fied CD56� NK/NKT cells and HCVcc/Huh7.5 cellsresulted in no IFN-� production, suggesting that IFN-�was produced from NK/NKT cells after acquiring help

) are produced from co-cultures of human PBMCs and HCV-infectedBMCs for 24 hours. IFNs in the supernatants were measured byh7.5 cells with different infection percentages (NS3 expression level)ured by enzyme-linked immunosorbent assay. Mean � SD is shown,ma cells were collected before co-culture as 0 hour control, or 4, 8

FN or MxA (MX1) expression were examined by real-time polymeraseernal control (mean � SD; n � 3). (D) Huh7.5 cells or HCVcc/Huh7.5, 16, and 24 hours later and IFNs or IP-10 levels determined by

29]th P/Hueasato

nd Iint

4, 8

from other accessory cells in the PBMC population.

(mcle

2)

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER


mDC2s (BDCA3� Dendritic Cells) Are theMajor Immune Cell Subset That Initiates aType III IFN Response

Our data suggested a pDC-independent recogni-tion of HCV-infected cells and an early induction of IL-28and IL-29 in human PBMCs. To identify the immune cellsubset responsible for IL-28 and IL-29 production, iso-lated subsets of mDC1 (CD1c� dendritic cells), mono-cyte-derived dendritic cells, CD14� monocytes, CD16�monocytes, and platelet-associated cells (CD61� cells)were co-cultured with HCVcc/Huh7.5 cells. However,none of these cell populations induced IL-28 and IL-29 atmRNA levels (Figure 2B). Depletion of granulocytes

Figure 2. IFN-� is produced by pDC, IFN-� is produced by NK/NKT, awith human PBMCs, purified pDCs, PBMCs depleted of pDCs (PBMC(CD56�) cells. IFNs in the supernatants were measured by enzyme-linkB) HCVcc/Huh7.5 cells were co-cultured with PBMCs, purified CD1c

onocytes), CD16� monocytes, platelet-associated cells (CD61�), PBytes (PBMC-CD14), PBMCs depleted of pDC (PBMC-pDC), PBMCs d

ymphocytes (PBMC-[T, B, NK, pDC]). Co-cultured cells were preservedxtracted from the cells and IL-28 or IL-29 expression was examined by r

cells were co-cultured with PBMCs, PBMCs depleted of mDC2 cells (PBMwas examined by real-time polymerase chain reaction (mean � SD; n �

(PBMC-CD15), monocytes (PBMC-CD14), pDCs (PBMC-

pDC), or contaminated platelets (PBMC-CD61) from hu-man PBMCs by magnetic bead separation revealed thatthe remaining cells still retained the capacity to induceIL-28 and IL-29 expression when co-cultured with HCVcc/Huh7.5 cells (Figure 2B). Furthermore, after depletion ofall lymphocytes and pDCs PBMC-[T, B, NK, pDC] by acocktail of antibody-labeled beads (CD3�, CD19�,CD16/56�, and CD123�), the remaining cells pro-duced high levels of IL-28 and IL-29, indicating thepresence of an immune subset of IFN-�–producingcells.

Recently, human mDC2s (also known as BDCA3�mDCs) were identified as the counterpart of murine

IFN-� is produced by mDC2. (A) HCVcc/Huh7.5 cells were co-culturedC), PBMCs depleted of NK/NKT (PBMC-CD56), or purified NK/NKT

immunosorbent assay after 24 hours (mean � SD; n � 3–6; *P � .05).mDC1, monocyte derived DCs (MoDC), purified monocytes (CD14�s depleted of granulocytes (PBMC-CD15), PBMCs depleted of mono-eted of platelet-associated cells (PBMC-CD61), or PBMCs depleted ofre co-culture as 0 hour control or collected after 4 hours. Total RNA wastime polymerase chain reaction (mean � SD; n � 2). (C) HCVcc/Huh7.5mDC2), or purified mDC2s. IL-28 and IL-29 expression from co-cultures.

nd-pD

ed�MCepl

befoeal-C-

CD8�� DCs15–17 and one report showed that both mu-

rradt

I

3

Pb

rnIr

HtpHp

bHpaFb(mr

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER


rine CD8�� DCs and human BDCA3� mDCs produceIFN-�s in response to poly I:C stimulation.23 Thus, wetested whether human mDC2 was responsible for IL-28and IL-29 induction by HCV-infected cells. We foundobust IL-28 and IL-29 induction from co-cultures of en-iched mDC2 with HCVcc/Huh7.5 cells, whereas IL-28nd IL-29 induction was diminished when PBMCs wereepleted of mDC2 (Figure 2C), indicating that mDC2 washe major cell population producing IFN-�s in response

to HCV-infected cells.

mDC2s Respond to HCV-Infected Cells or theTLR3-Ligand, Poly I:C, and Produce HighLevels of IL-29To evaluate the involvement of PRRs in IFN-� and

FN-� induction in pDCs and mDC2s, we tested TLR3, 7,8, or 9 ligands that mimic viral pathogen-associated mo-lecular patterns. We did not detect IFN-� release frommDC2s, whereas pDCs produced IFN-� after co-culturewith HCVcc/Huh7.5 cells or stimulation with TLR7, 8, or9 ligands (Figure 3A). We found high levels of IL-29 inco-cultures of HCVcc/Huh7.5 cells and mDC2s. In con-trast, pDC produced low levels of IL-29 in response toHCVcc/Huh7.5 cells (Figure 3B). We further determinedthat mDC2s produced high levels of IL-29 in response toTLR3 ligand (Poly I:C) and not to TLR7, 8, or 9 ligandstimulation (Figure 3B). IL-29 production in pDCs re-quired TLR9 (cytosine-phosphate-guanosine [CpG]-A)stimulation and TLR3, 7, or 8 ligands failed to induceIL-29 in isolated pDCs (Figure 3B). Consistent with thesefindings, we found that mDC2s had the highest level ofTLR3 mRNA expression among PBMCs, whereas pDCsspecifically expressed TLR7 and TLR9 mRNA (Figure 3C).

Because of their specific expression of either TLR3 orTLR7, we suspected that mDC2s and pDCs responded tosingle-stranded RNA (ssRNA) ligand or double-strandRNA (dsRNA) ligand in HCV-infected cells and producedIFN-� and IFN-�, respectively. To test this hypothesis, aseries of HCV-ssRNAs and dsRNAs with different lengthswere generated to stimulate human pDCs or mDC2s(Figure 3D). We found that mDC2 preferentially re-sponded to synthesized HCV-dsRNA rather than ssRNA,produced IL-28 and IL-29 and increased expression ofCD80 and CD86 (Figure 3E), whereas pDCs preferentiallyresponded to synthesized HCV-ssRNAs with IFN-� pro-duction and increased co-stimulatory molecule expression(CD80 and CD86) (Figure 3F and Supplementary Figure

). We also examined the involvement of PRRs in IFN-�induction by primary human hepatocytes (PHHs) in re-sponse to HCV infection. We found that TLR3 andTLR7/8 ligands failed to stimulate IFN-� production in

HHs, whereas JFH-1 virions or RIG-I ligand inducedoth IL-28 and IL-29 in PHHs (Supplementary Figure 4).

Further analysis of the pDC and mDC2 responses afterecognition of HCV-infected cells revealed unique immu-ologic features of mDC2. First, mDC2 showed rapid

L-29 production compared with pDCs, indicating rapid
ecognition and response in mDC2s after exposure to
CV-infected hepatoma cells (Figure 4A). Second, we de-ermined that, on average, the same number of mDC2sroduced 10 times more IL-29 than pDCs in response toCVcc/Huh7.5 cells (Figure 4B). We also confirmed thatDCs produced large amounts of IFN-� (Figure 4B). Next,

we showed that the IFN-� response by mDC2s includedoth IL-28 and IL-29 production in response to HCVcc/uh7.5 cells or Poly I:C stimulation. In contrast, pDCsroduced low levels of IL-29 and very little IL-28 whenctivated by HCVcc/Huh7.5 cells or CpG-A (Figure 4C).inally, we found that mDC2s and not pDCs producedoth IL-28 and IL-29 in response to JFH-1–infected PHHsFigure 4D). These observations identified mDC2s as the

ajor cellular source of IL-28 and IL-29 production inesponse to HCV-infected hepatocytes.

Recognition of HCV-Infected Cells by mDC2sRequires Cell–Cell Contact and Involves theEndocytosis Pathway

The liver tissue environment allows close contactbetween hepatocytes and immune cells. We found thatseparation of mDC2s and HCV-infected hepatoma cellswith a Transwell insert (BD Biosciences, Franklin Lakes,NJ) fully prevented IL-29 induction, indicating that cell–cell contact was necessary for mDC2 activation (Figure5A). We also found that concentrated HCV virions failedto induce IL-29 production in mDC2 (SupplementaryFigure 5). To further exclude the possibility that HCVvirions coming from infected cells cause mDC2 activa-tion, HCV replicon cell lines (subgenomic replicon [BB7]and full-length genomic replicon [FL]), which do notproduce infectious virions, were co-cultured with humanmDC2. Robust IL-29 induction was present at bothmRNA levels and protein levels in mDC2 co-cultures withreplicon cells (Figure 5B and Supplementary Figure 6). Wenext hypothesized that mDC2–IFN-� induction fromHCV-infected cells occurs through an endocytosis-medi-ated pathway. We determined that inhibitors of endocy-tosis (cytochalasin D and chlorpromazine) significantlydiminished IL-29 release from mDC2-hepatoma co-cul-tures, whereas inhibitors of macropinocytosis (dimethylamiloride) did not (Figure 5C). Finally, because TLR3signaling induced IFN-� production in mDC2, we testedwhether endosome function was required for mDC2 ac-tivation. We found that pretreatment of mDC2 with in-hibitors of endosome acidification (chloroquine and ba-filomycin) abolished the IL-29 production from mDC2and HCV-infected hepatoma co-cultures (Figure 5D), in-dicating the signaling transduction pathway involved en-dosome acidification. IL-29 production also was abolishedin endosomal-inhibitor–treated mDC2s stimulated withthe TLR3 ligand, poly I:C.

IFN-� and IFN-� Amplify Antiviral IFNInductionRecently, SNP studies revealed a close association

between the IL-28B gene and the outcome of HCV-in-
fected individuals, however, the molecular basis for this

a

THamHTT

2mbma5hiaPptstsiDIm

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER


finding is unknown.5– 8 We hypothesized that type I andtype III IFNs mutually amplify their antiviral effects dur-ing HCV infection. Because mDC2–IFN-� production pre-

Figure 3. mDC2s specifically expressLR3 and produce IFN-�s in response toCV-dsRNA or poly I:C stimulation. (And B) Human PBMCs, purified pDCs, orDC2s were co-cultured with HCVcc/uh7.5 cells, or stimulated with differentLR ligands (Invivogen, San Diego, CA):LR3-Poly I:C (10 �g/mL), TLR7-

Gardiquimod (1 �g/mL), TLR7/8-CL075 (2.5�g/mL), and TLR9-CpG-A (2 �mol/L) for4 hours. IFNs in the supernatants wereeasured by enzyme-linked immunosor-ent assay (mean � SD; n � 3). (C)RNA expression of TLRs (TLR3, 7, 8,

nd 9) and RLRs (DEAD box polypeptide8 [DDX58] and interferon induced withelicase C domain 1 [IFIH1]) was exam-

ned by real-time polymerase chain re-ction from different immune subsets inBMCs using glyceraldehyde-3-phos-hate dehydrogenase as internal con-

rol. Relative mRNA expressions arehown (mean � SD; n � 3). (D) Gel elec-rophoresis of in vitro synthesized HCV-sRNA and -dsRNA derived from the

ndicated HCV genome region. M, 1-kbNA ladder (New England Biolabs Inc,

pswich, MA). (E and F) Isolated pDCs orDC2s were stimulated by 10 �g/mL

HCV-ssRNA or -dsRNA mixed with Li-pofectamine 2000 (Invitrogen, Grand Is-land, NY) for 24 hours. IFN levels weremeasured by enzyme-linked immu-nosorbent assay (n � 3). (E) IFN-� pro-duction in pDCs, and (F) IL-28 and IL-29production in mDC2s, are shown.

ceded pDC–IFN-� production in co-cultures, we first ex- H

mined the possible effect of IFN-� on IFN-� production.We found that addition of recombinant IL-29 significantlyincreased IFN-� and IFN-� release in PBMCs plus HCVcc/

uh7.5 co-cultures (Figure 6A), whereas depletion of mDC2

Ci

tu

bmmF

tag

mmI

iud(

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER


from PBMCs (PBMC-mDC2) significantly reduced IFN-�production (Figure 6B). In contrast, depletion of mDC2failed to significantly affect PBMC IFN-� production after

pG-A stimulation, suggesting that the presence of mDC2n HCV infection promotes pDC function (Figure 6B). Next,

we tested whether IFN-� had a direct effect on pDC–IFN-�production. We found that addition of recombinant IL-29significantly enhanced pDC IFN-� production in responseo HCV-infected hepatoma cells or CpG-A stimulation (Fig-re 6C). There was no IFN-� induction by IFN-� alone

(Figure 6C). Furthermore, we determined that among all cellpopulations in our study, pDCs showed the highest expres-sion of IL-28RA, which is the specific receptor for IFN-�(Figure 6D). We also confirmed that human pDCs expressed

oth the functional and nonfunctional isoform of IL-28RARNA, whereas hepatoma cells and Huh7 and Huh7.5 cellsainly expressed the functional isoform (Supplementary

igure 7).Next, to examine the effect of IFN-� on IFN-� produc-

tion, we added recombinant IFN-� to human PBMCs or

o HCV-infected cells and observed no induction of IL-28nd IL-29 despite increases in MX1, an IFN-�–inducedene (Figure 6E). However, IFN-� treatment in human

PBMC co-cultures with HCVcc/Huh7.5 cells resulted insignificantly increased IL-28 and IL-29 induction at themRNA level (Figure 6E) and the protein level (Figure 6F).Finally, to test whether IFN-� directly could affect

DC2 function, we added recombinant IFN-� to hu-an mDC2s and found a significant increase in mDC2–

FN-� production in co-cultures or after Poly I:C stim-ulation (Figure 6F), indicating that IFN-� can augmentmmune responses against HCV infection through reg-lating type III IFN production by mDC2s. IFN-� aloneid not induce IL-28 and IL-29 production in mDC2s

Figure 6F).

Genotype of rs12979860, Chronic Hepatitis C,and IFN-� Production in PBMCsGiven the importance of IL-28 SNPs in clinical

Figure 4. Functional dichot-omy of pDCs and mDC2s in IFNproduction. (A) Purified pDC ormDC2 populations were co-cul-tured with HCVcc/Huh7.5 cells,and culture supernatants werecollected at different time pointsfor IFN-� and IL-29 measure-ment by enzyme-linked immu-nosorbent assay (mean � SD;n � 2). (B) PBMC-IFN produc-tion per cell was artificially set as1 unit, then the average IFN re-lease from pDCs or mDC2s wascalculated and compared (mean �SD; n � 3). (C) Purified pDCs ormDC2s were co-cultured withHCVcc/Huh7.5 cells, or stimu-lated with CpG-A or Poly I:C, re-spectively. Twenty-four hourslater, IL-28 or IL-29 in culture su-pernatants was measured byenzyme-linked immunosorbentassay (mean � SD; n � 3). (D)JFH-1–infected PHHs were col-lected 48 hours after infection forco-culture with human pDCs ormDC2s. Co-culture superna-tants were collected at differenttime points and IL-28 or IL-29levels were measured by en-zyme-linked immunosorbent as-say. Representative data fromone experiment is shown.

outcomes, we examined the effect of rs12979860 genotype

Hrgcotsnnwme(cpea

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER


on PBMC–IFN-� production in normal donors with C/Cand C/T or T/T genotypes. We found no significant dif-ference in IL-28 and IL-29 production in PBMCs fromT/T, C/T, or C/C groups in response to JFH-1–infectedhepatoma cells, poly I:C, or CpG-A stimulation (Figure7A). We found comparable frequencies of circulatingmDC1, mDC2, and pDC in chronic HCV-infected pa-tients and that their PBMCs showed no defect in IFN-�induction in response to JFH-1–infected hepatoma cellsor poly I:C stimulation (Figure 7B and Supplementary

Figure 5. HCV-induced IFN-�induction requires cell-to-cellcontact and endosomal acidifi-cation. (A) HCVcc/Huh7.5 cellswere co-cultured with humanPBMCs or purified mDC2s in anormal plate or a plate with aTranswell insert. IL-29 in the su-pernatants was measured byenzyme-linked immunosorbentassay after 24 hours (mean �SD; n � 3). (B) Huh7.5 cells and

CV replicon cells, subgenomiceplicon (BB7) and full-lengthenomic replicon (FL) cells, wereo-cultured with human PBMCsr purified mDC2s. IL-29 in cul-ure supernatants was mea-ured by enzyme-linked immu-osorbent assay (mean � SD;� 2). (C) HCVcc/Huh7.5 cellsere co-cultured with purifiedDC2 in the absence or pres-

nce of dimethyl amilorideDMA), cytochalasin D (CCD), orhlorpromazine (CLP). IL-29roduction was measured bynzyme-linked immunosorbentssay (mean � SD; n � 2). (D)

Purified human mDC2s were co-cultured with HCVcc/Huh7.5cells or stimulated with Poly I:Cin the presence of chloroquine orbafilomycin as indicated. IL-29production was measured byenzyme-linked immunosorbentassay after 24 hours (mean �SD; n � 2).

Figure 8).

Finally, we found significantly higher expression of DCmarkers, BDCA1 (cluster of differentiation 1c molecule[CD1C]), BDCA2 (C-type lectin domain family 4, memberC [CLEC4C]), BDCA3 (thrombomodulin [THBD]), andBDCA4 (neuropilin 1 [NRP1]) in chronic HCV-infectedlivers compared with noninfected livers (Figure 7C), indi-cating the recruitment of mDCs and pDCs into the liverduring HCV infection. We also confirmed that chronicHCV-infected livers had significantly higher ISG mRNAexpression, including ISG15, ISG56, MX1, and CXCL10
compared with controls (Supplementary Figure 9).

W

Ic

itcHIpmHrpii

Ih

icIfii

z

(P

mpcaw

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER


DiscussionHCV infection stimulates an early IFN response in

the liver and, paradoxically, high IFN-stimulated geneexpression correlates with poor viral clearance.1,4 How-ever, the exact mechanisms by which immune cells recog-nize HCV and/or HCV-infected hepatocytes only partiallyare understood. In this study, using co-cultures of humanPBMCs and HCV-infected hepatoma cells, we showed thatall 3 types of IFNs (IFN-�, IFN-�, and IFN-�) are induced.

e identified the mDC2 population as a source of IFN-�production in HCV infection. We further showed thatIFN induction is TLR- and cell-specific. IFN-� productionrequired TLR3 activation in mDC2s that responded toHCV-dsRNA in contrast to IFN-� production by pDCs

that required TLR7 activation by HCV-ssRNA. Our noveldata indicated that IFN-� induction in mDC2s was de-pendent on direct cell contact with the HCV-infectedhepatoma cells and both endocytosis and endosomal acid-ification were involved. Finally, we showed that IFN-� andIFN-� have mutual cross-regulatory effects and amplifyFN production in response to HCV-infected hepatomaells.

A recent report indicated that pDCs recognize HCV-nfected cells through cell– cell contact.22 We show herehat mDC2s recognize HCV-infected cells using a cell– cellontact– dependent mechanism. We found that onlyCV-infected cells and not HCV virions activated mDC2s.

n support of this, HCV replicon cells, which do notroduce HCV virions, also stimulated mDC2s. Further-ore, we showed that mDC2s capture their ligands fromCV-infected cells through an endocytosis-, and not mac-

opinocytosis-, mediated pathway. Interestingly, earlier re-orts indicated that this phenomenon also occurred dur-

ng other human viral infections, such as humanmmunodeficiency virus and influenza infection.24,25

Compared with type I IFNs, the cellular source andinduction pattern of IFN-�s, including IL-28a, IL-28b, andL-29, are undefined. Recently, it was reported that humanepatocytes produce IFN-�s in vitro after HCV infec-

tion.10 Here, we also found IFN-� induction by HCVinfection in primary human hepatocytes and our noveldata showed an interesting paradigm of IFN-� productionn human immune cells after recognition of HCV-infectedells in which mDC2s produce high levels of IL-28 andL-29 but pDCs only produce low levels of IL-29. Ourndings support the hypothesis that mDC2s are special-

zed IFN-� producers in response to HCV infection; this isin agreement with a recent study in which mouse CD8��

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™Figure 6. Cross-regulations exist between IFN induction in co-cultures.(A) HCVcc/Huh7.5 cells were co-cultured with human PBMCs in theabsence or presence of recombinant IL-29 (25 ng/mL). IFN-� and IFN-�levels were measured by enzyme-linked immunosorbent assay (mean �SD; n � 3). (B) Human PBMCs or PBMCs depleted of mDC2s (PBMC-mDC2) were co-cultured with HCVcc/Huh7.5 cells or stimulated withCpG-A. IFN-� production in the supernatants was measured by en-yme-linked immunosorbent assay (mean � SD; n � 3–4). (C) Purified

human pDCs were co-cultured with HCVcc/Huh7.5 cells or stimulatedwith CpG-A in the absence or presence of IL-29. IFN-� production wasmeasured by enzyme-linked immunosorbent assay (mean � SD; n � 3).D) Total RNA was extracted from different immune subsets in humanBMCs. IL-28RA expression was examined by real-time polymerase

chain reaction and normalized to glyceraldehyde-3-phosphate dehydro-genase (mean � SD; n � 3). (E) Human PBMCs, Huh7.5 cells, HCVcc/Huh7.5 cells, co-culture of PBMCs and Huh7.5 cells, or co-culture ofPBMCs and HCVcc/Huh7.5 cells were preserved before culture as0-hour control or collected after 4 hours culturing in the absence orpresence of IFN-�-2a (100 U/mL). IL-28, IL-29, and MX1 induction were

easured by real-time polymerase chain reaction; 1 representative ex-eriment of 3 is shown. (F) Human PBMCs or purified mDC2s wereo-cultured with HCVcc/Huh7.5 cells or stimulated with Poly I:C in thebsence or presence of IFN-�-2a. IL-29 production in the supernatantsas measured by enzyme-linked immunosorbent assay (mean � SD;

n � 3) (*P � .05). PBS, phosphate-buffered saline.

dd

bcOfi

pt

o(

easr

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER


DCs, the murine homologue of human mDC2, werefound to produce IL-28 and IL-29 during poly I:C injec-tion in vivo.23 Although IFN-�s are speculated to be in-

uced through similar signaling pathways as type I IFNsuring viral infection,26 we found that IFN-� and IFN-�

had distinct cellular sources in pDCs and mDC2s, respec-tively, indicating that the mechanism of controlling in-duction of type I IFNs and IFN-�s induction is cell- andstimulation-specific. In other studies, IFN-� inductionwas limited to a small fraction of pDCs in a similarsystem22 and at this time we cannot rule out the possi-

ility that within the mDC2 population only a subset ofells produce very large amounts of IL-28 and/or IL-29.ur results suggested that the IFN production profile is

Figure 7. Genotype of rs-12979860 does not correlatewith IL-28 and IL-29 induction inhuman PBMCs by HCV-infectedcells. (A) Human PBMCs wereco-cultured with JFH-1–infectedHuh7.5 cells or stimulated withPolyI:C (20 �g/mL) or CpG-A (2�mol/L) for 24 hours. Co-culturesupernatants were collectedand IL-28 or IL-29 levels weredetermined by enzyme-linkedimmunosorbent assay. Eachdot represents one individual. (B)Circulating mDC1, mDC2, andpDC frequencies from normalcontrol (NC) or naive HCV patients(HCV) were determined by flowcytometry. (C) BDCA1 (CD1C),BDCA2 (CLEC4C), BDCD3(THBD), and BDCA4 (NRP1)mRNA expression in human liverswere determined by real-timepolymerase chain reaction usingglyceraldehyde-3-phosphate de-hydrogenase as internal control.

nely tuned during immune response to HCV and de-

ends on differential use of PRRs or transcriptional fac-ors in different cell types.

The mammalian immune system mainly uses 2 groupsf PRRs to sense viral infection: RIG-I–like receptorsRLRs) and TLRs (TLR3, 7, 8, or 9).27 RLRs localize in the

cytosol and usually detect intracellular viruses. BecauseHCV does not infect immune cells effectively, it is specu-lated that immune cells recognize HCV infection throughTLR pathways. Indeed, human pDC can recognize HCVthrough a TLR7-mediated pathway.22 We present novelvidence that human mDC2s recognize HCV infectionnd that IFN-� induction is mediated by the dsRNA-ensing, TLR3-mediated pathway. Consistent with earliereports,28 mDC2s specifically expressed high levels of

TLR3 and very low levels of TLR7 and TLR9 mRNA in our

fm

Hc

t

(ieutmrtcmc

1

1

1

1

1

1

1

1

1

1

BA

SICA

ND

TRA

NSLA

TION

AL

LIVER


experiments; in addition, only the TLR3 ligand, poly I:C,or HCV-dsRNA, and not TLR7, 8, and 9 ligands or HCV-ssRNA, stimulated mDC2–IFN-� production. We alsoound that inhibitors of endosome acidification abolished

DC2–IFN-� production in co-cultures, further indicat-ing that a TLR and not an RLR pathway was responsiblefor recognition of HCV by mDC2.

SNP studies showed that IL-28b (IFN-�3) is associatedclosely with the recovery of natural HCV infection or afterIFN-� therapy5– 8; however, the molecular basis is unclear.Although we found no defects in IFN-� production indonors with unfavorable SNP genotype, our results pro-vide other key information for reconsidering the underly-ing mechanisms because we showed inter-regulatory rela-tionships between type I IFN and type III IFN. In ourstudy, IFN-� was released earlier than IFN-� in the PBMC-

CV–infected hepatoma co-cultures and IFN-� directlyould regulate pDC to increase their IFN-� production. In

contrast to earlier reports suggesting that minimal expres-sion or a spliced form of IL-28Ra, the specific receptor forIFN-�, on hematopoietic cells that results in their lowresponsiveness to IFN-�,29,30 we found that pDCs expressa higher level of IL-28Ra compared with other cell popu-lations in PBMCs, indicating that this responsiveness toIFN-� is unique and important for pDC. Conversely, wealso found that IFN-� directly could affect mDC2 functionand significantly increase IFN-� production. It is temptingo speculate that exogenous IFN-� would increase IFN-�

production by mDC2s during IFN-� therapy, and this couldprovide a potential explanation why SNP types of IL-28bpredict the outcome of IFN-� therapy.

The changes of cell number and function of mDC1sCD1c� mDCs) and pDCs in acute and chronic HCVnfection have been characterized in several studies, how-ver, the study of mDC2 was hindered by its previousnknown function and low frequency.31 Here, we foundhat the frequency and IFN-� production by circulating

DC2s from chronic HCV-infected patients was compa-able with normal controls. However, the expression ofhe mDC2 marker BDCA3 was increased significantly inhronic HCV-infected livers, indicating an enrichedDC2 population in the local liver environment in

hronic HCV infection. Importantly, because IFN-� in-duction in chronically HCV-infected hepatocytes could beinhibited as a result of mitochondrial antiviral signalingprotein and TIR-domain– containing adapter–inducinginterferon-� cleavage, we suspect that mDC2s and pDCsresiding in the liver might represent an alternative ornecessary resource of IFNs when hepatocyte-derived IFNproduction is disabled by HCV.

In conclusion, our novel data identify a unique DC pop-ulation, mDC2s, as an important source of IFN-� produc-tion in response to HCV-infected hepatoma cells. Theseobservations also indicate that innate immune cells effi-ciently can initiate immune responses against HCV infec-
tion.
Supplementary Material

Note: To access the supplementary materialaccompanying this article, visit the online version ofGastroenterology at www.gastrojournal.org, and at http://dx.doi.org/10.1053/j.gastro.2012.10.034.

References

1. Sarasin-Filipowicz M, Oakeley EJ, Duong FH, et al. Interferon sig-naling and treatment outcome in chronic hepatitis C. Proc NatlAcad Sci U S A 2008;105:7034–7039.

2. Kelly C, Klenerman P, Barnes E. Interferon lambdas: the nextcytokine storm. Gut 2011;60:1284–1293.

3. Pang KR, Wu JJ, Huang DB, et al. Biological and clinical basisfor molecular studies of interferons. Methods Mol Med 2005;116:1–23.

4. Rehermann B. Hepatitis C virus versus innate and adaptive im-mune responses: a tale of coevolution and coexistence. J ClinInvest 2009;119:1745–1754.

5. Ge D, Fellay J, Thompson AJ, et al. Genetic variation in IL28Bpredicts hepatitis C treatment-induced viral clearance. Nature2009;461:399–401.

6. Tanaka Y, Nishida N, Sugiyama M, et al. Genome-wide associationof IL28B with response to pegylated interferon-alpha and ribavirintherapy for chronic hepatitis C. Nat Genet 2009;41:1105–1109.

7. Suppiah V, Moldovan M, Ahlenstiel G, et al. IL28B is associatedwith response to chronic hepatitis C interferon-alpha and ribavirintherapy. Nat Genet 2009;41:1100–1104.

8. Thomas DL, Thio CL, Martin MP, et al. Genetic variation in IL28Band spontaneous clearance of hepatitis C virus. Nature 2009;461:798–801.

9. Muir AJ, Shiffman ML, Zaman A, et al. Phase 1b study of pegylatedinterferon lambda 1 with or without ribavirin in patients withchronic genotype 1 hepatitis C virus infection. Hepatology 2010;52:822–832.

0. Marukian S, Andrus L, Sheahan TP, et al. Hepatitis C virus inducesinterferon-lambda and interferon-stimulated genes in primary livercultures. Hepatology 2011;54:1913–1923.

1. Li K, Li NL, Wei D, et al. Activation of chemokine and inflammatorycytokine response in HCV-infected hepatocytes depends on TLR3sensing of HCV dsRNA intermediates. Hepatology 2011;55:666–675.

2. Lemon SM. Induction and evasion of innate antiviral responses byhepatitis C virus. J Biol Chem 2010;285:22741–22747.

3. Ziegler-Heitbrock L, Ancuta P, Crowe S, et al. Nomenclature ofmonocytes and dendritic cells in blood. Blood 2010;116:e74–e80.

4. Piccioli D, Tavarini S, Borgogni E, et al. Functional specialization ofhuman circulating CD16 and CD1c myeloid dendritic-cell subsets.Blood 2007;109:5371–5379.

5. Crozat K, Guiton R, Contreras V, et al. The XC chemokine receptor1 is a conserved selective marker of mammalian cells homolo-gous to mouse CD8alpha� dendritic cells. J Exp Med 2010;207:1283–1292.

6. Jongbloed SL, Kassianos AJ, McDonald KJ, et al. Human CD141�(BDCA-3)� dendritic cells (DCs) represent a unique myeloid DCsubset that cross-presents necrotic cell antigens. J Exp Med2010;207:1247–1260.

7. Poulin LF, Salio M, Griessinger E, et al. Characterization of humanDNGR-1� BDCA3� leukocytes as putative equivalents of mouseCD8alpha� dendritic cells. J Exp Med 2010;207:1261–1271.

8. Reizis B, Bunin A, Ghosh HS, et al. Plasmacytoid dendritic cells:recent progress and open questions. Annu Rev Immunol 2011;29:163–183.

9. Marukian S, Jones CT, Andrus L, et al. Cell culture-producedhepatitis C virus does not infect peripheral blood mononuclear
cells. Hepatology 2008;48:1843–1850.
http://www.gastrojournal.org



2

2

2


20. Shiina M, Rehermann B. Cell culture-produced hepatitis C virusimpairs plasmacytoid dendritic cell function. Hepatology 2008;47:385–395.

1. Gondois-Rey F, Dental C, Halfon P, et al. Hepatitis C virus is aweak inducer of interferon alpha in plasmacytoid dendritic cells incomparison with influenza and human herpesvirus type-1. PLoSOne 2009;4:e4319.

2. Takahashi K, Asabe S, Wieland S, et al. Plasmacytoid dendriticcells sense hepatitis C virus-infected cells, produce interferon,and inhibit infection. Proc Natl Acad Sci U S A 2010;107:7431–7436.

3. Lauterbach H, Bathke B, Gilles S, et al. Mouse CD8alpha� DCsand human BDCA3� DCs are major producers of IFN-lambda inresponse to poly IC. J Exp Med 2010;207:2703–2717.

24. Lepelley A, Louis S, Sourisseau M, et al. Innate sensing of HIV-infected cells. PLoS Pathog 2011;7:e1001284.

25. Lui G, Manches O, Angel J, et al. Plasmacytoid dendritic cellscapture and cross-present viral antigens from influenza-virus ex-posed cells. PLoS One 2009;4:e7111.

26. Iversen MB, Paludan SR. Mechanisms of type III interferon expres-sion. J Interferon Cytokine Res 2010;30:573–578.

27. Takeuchi O, Akira S. Innate immunity to virus infection. ImmunolRev 2009;227:75–86.

28. Yamazaki C, Miyamoto R, Hoshino K, et al. Conservation of achemokine system, XCR1 and its ligand, XCL1, between humanand mice. Biochem Biophys Res Commun 2010;397:756–761.

29. Sommereyns C, Paul S, Staeheli P, et al. IFN-lambda (IFN-lambda)is expressed in a tissue-dependent fashion and primarily acts on
epithelial cells in vivo. PLoS Pathog 2008;4:e1000017.
30. Witte K, Gruetz G, Volk HD, et al. Despite IFN-lambda receptorexpression, blood immune cells, but not keratinocytes or melano-cytes, have an impaired response to type III interferons: implica-tions for therapeutic applications of these cytokines. Genes Im-mun 2009;10:702–714.

31. Dolganiuc A, Szabo G. Dendritic cells in hepatitis C infection: canthey (help) win the battle? J Gastroenterol 2011;46:432–447.

Received January 19, 2012. Accepted October 15, 2012.

Reprint requestsAddress requests for reprints to: Gyongyi Szabo, MD, PhD,

Department of Medicine, University of Massachusetts Medical School,364 Plantation Street, Worcester, Massachusetts 01605. e-mail:[email protected]; fax: (508) 856-4770.

AcknowledgmentsThe authors thank Drs Charles M. Rice and Takaji Wakita for

kindly providing reagents.

Conflicts of interestThe authors disclose no conflicts.

FundingSupported by National Institutes of Health grants R37AA014372

(G.S.) and AI069285 (K.L.). Primary human hepatocyte cultures andhuman liver samples were provided by the National Institutes ofHealth–funded Liver Tissue Procurement and Cell Distribution System
(N01-DK-7-0004/HHSN26700700004C).
BA

SIC

AN

DTR

AN

SLA

TIO

NA

LLI

VER

mailto:[email protected]

cuiv

taBBMbiLI

courcbRuncdaFi1LaP

oBws(

c

p

425.e1 ZHANG ET AL GASTROENTEROLOGY Vol. 144, No. 2

Supplementary Materials and Methods

Cells, Replicons, and VirusesHuh7.5 cells and full-length genomic replicon

cells (FL) were kindly provided by Dr C. Rice. Sub-genomic replicon cells (BB7) were a generous gift of Dr C.Seeger. All cells were maintained in low-glucose Dulbec-co’s modified Eagle medium containing 10% fetal bovineserum, 10 �g/mL ciprofloxacin, and supplemented withnonessential amino acids. Replicon cells were maintainedin media containing 500 �g/mL G418. PHHs were ob-tained from Stephen Strom at the University of Pitts-burgh, through the National Institutes of Health–fundedLiver Tissue and Cell Distribution System, and main-tained in hepatocytes maintenance medium (HMM-hep-atocyte maintenance medium from Lonza, Allendale, NJ).JFH-1 genomic RNA was transcribed in vitro from thelinearized pUC–vJFH-1 plasmid using an in vitro tran-scription kit (MEGAscript; Ambion, Grand Island, NY).Transfection of synthesized JFH-1 RNA into Huh7.5 cellsand production of JFH-1 viruses were performed as previ-ously described.1 To produce concentrated JFH-1 virions,culture supernatants were collected from Huh7.5 cellshighly infected with JFH-1 HCVcc (90% infected cells ex-pressing NS3 examined by fluorescence-activated cell sorter[FACS]) and passed through 0.22-�m filters. Then JFH-1–ontaining supernatants were concentrated for 20–30 timessing Amicon Ultra-15 100K (Millipore, Billerica, MA). The

nfectivity of concentrated JFH-1 virions was confirmed andirus stocks were preserved at �80°C.

ReagentsODN2216 (CpG-A, TLR9 ligand), TLR3-ligand

Poly I:C, TLR7/8-ligands CL075 and Gardiquimod, en-dosome inhibitor chloroquine, and bafilomycin A1 werepurchased from Invivogen (San Diego, CA). Dimethylamiloride, cytochalasin D, and chlorpromazine were pur-chased from Sigma-Aldrich (St. Louis, MO). Recombi-nant IFN-�-2a and IL-29 were purchased from PBL In-erferonSource (Piscataway, NJ). Anti–lineage-1 FITC,nti-HLA DR PerCp, and anti-CD123 PE-Cy7 were fromiolegend (San Diego, CA). Anti-CD1c PE and anti-DCA3 APC were from Miltenyi Biotec (Cambridge,A). Anti-HCV NS3 (#1847) was from Virostat (West-

rook, ME). Anti-CD80 PE, anti-CD86 PE, and Transwellnserts were purchased from BD Biosciences (Franklinakes, CA). Alex flour 488 – conjugated goat-anti-mousegG (H�L) was from Invitrogen (Grand Island, NY).

Determination of HCV Infection Percentagein Hepatoma CellsUninfected or infected Huh7.5 cells were collected

and fixed with 4% paraformaldehyde for 20 minutes atroom temperature. Next, fixed cells were permeabilizedwith 0.2% Tween-20/phosphate-buffered saline (contain-ing 0.5% bovine serum albumin), primarily stained with
monoclonal mouse–anti–HCV-NS3 antibodies for 1 hour t
at room temperature, and stained with Alex flour 488 –conjugated goat-anti-mouse antibodies. HCV protein ex-pression levels were determined by FACS analysis. Oneset of representative data is shown in SupplementaryFigure 10A, indicating that almost all cells were infectedat 48 hours after JFH-1 virion inoculation.

Preparation of Human PBMCs, DC Subsets,and Co-culture ExperimentsHuman PBMCs were isolated using Ficoll-Hypaque

density centrifugation. Untouched pDCs were selected fromPBMCs using a pDC isolation kit (Stemcell Technologies,Vancouver, BC, Canada) and mDC2s were enriched fromPBMCs using a BDCA-3 microbeads kit (Miltenyi Biotec)according to the manufacturer’s instructions. mDC1s wereisolated from PBMCs using a CD1c (BDCA-1)� dendriticell isolation kit (Miltenyi Biotec). In some cases, BDCA-3–r BDCA-4–positive selection beads (Miltenyi Biotec) weresed to deplete mDC2s or pDCs from human PBMCs,espectively. To generate monocyte-derived DCs, CD14�ells were positively isolated from PBMCs using magneticeads (CD14 microbeads; Miltenyi Biotec) and cultured inPMI 1640 medium supplemented with 100 ng/mL gran-locyte macrophage colony–stimulating factor and 10g/mL IL-4. Three days later, half of the medium washanged and fresh cytokines were added as we previouslyescribed.2 Monocyte-derived DCs were harvested on day 6nd the DC-phenotype (CD1a�CD14�) was confirmed byACS analysis. Isolated PBMCs, pDCs, mDC2s, or other

mmune cell populations were cultured in RPMI medium640 supplemented with 10% fetal bovine serum (Hyclone,ogan, UT), 100 U/mL penicillin, 100 mg/mL streptomycin,nd nonessential amino acids (Gibco, Grand Island, NY).BMCs (4 � 106), pDCs (5 � 104), or mDC2s (5 � 104) were

co-cultured or stimulated in 24-well plates with a finalvolume of 1 mL/well for experiments shown in Figures 3and 4, whereas 2 � 104 mDC2s per well were used in all

ther experiments. For co-culture experiments, Huh7.5,B7, FL, and HCVcc/Huh7.5 cells (2 days postinfectionith JFH-1 virions) were washed with phosphate-buffered

aline, trypsinized, and resuspended in RPMI medium 1640Gibco). Unless otherwise indicated, 2 � 105 hepatoma

cells/well (including Huh7.5, BB7, FL, and HCVcc/Huh7.5cells) were plated in 24-well plates in a 37°C/5% CO2 incu-bator before co-culture with human immune cells. At theindicated time points, supernatants were collected for en-zyme-linked immunosorbent assay (ELISA) and co-culturedcells were processed for RNA extraction.

Cytokine Detection by BioplexThe levels of 10 inflammatory cytokines in culture

supernatants (IL-1�, IL-6, IL-8, IL-10, IL-12, tumor ne-rosis factor-�, IP-10, TRAIL, RANTES, and macrophage

inflammatory protein [MIP]-1�) were measured by Bio-lex (Bio-Rad, Hercules, CA) according to the manufac-
urer’s instructions.

ut

ptQwruitetiewmcrcy

esT�cgac(mb

February 2013 HUMAN MDC2 RECOGNIZE HCV AND PRODUCE IFN–� 425.e2

ELISAHuman IFN levels in the cell-culture supernatant

were measured using commercially available ELISA kitsfollowing the manufacturer’s instructions. Human IFN-�,IL-28, and IL-29 ELISA kits were purchased from PBL In-terferonSource (NJ). According to the supplier, the IL-28ELISA kit barely cross-reacts with IL-29, whereas the IL-29kit cross-reacts with IL-28. The human IFN-� ELISA kit waspurchased from R&D Systems (Minneapolis, MN).

RNA Extraction and QuantificationAs we previously described,1 RNA was extracted

sing the RNeasy Mini kit (Qiagen, Valencia, CA). Quan-ification of HCV RNA and mRNA levels of IFNA1, IFNB,

IFNG, IL-28, IL-29, MX1, and other genes as indicated wereerformed by a 2-step real-time polymerase chain reac-ion (PCR). First, whole cellular RNA was extracted usingiagen RNA mini kits and complementary DNA (cDNA)as synthesized using reverse transcription. Then the

eal-time PCR was conducted on the Bio-Rad CFX systemsing the SYBR green (Bio-Rad) method. The PCR spec-

ficity was examined by melting curve, which indicateshe mispriming or primer-dimer artifacts in PCR. Glyc-raldehyde-3-phosphate dehydrogenase served as the in-ernal control for all experiments. When examining thenduction of IL-28, IL-29, ISGs, or other cytokine genexpressions in co-cultures, each co-culture was comparedith the corresponding co-culture 0 hour, which wasixed cells sitting at 4°C for 2 hours. The expression of

o-culture 0 hour was set as artificial unit 1, and then theelative values of each co-culture were calculated andompared. Primer sequences used for real-time PCR anal-sis are listed in Supplementary Table 1.

Plasmids and Generation of HCV ssRNAsPlasmids pBSII-core, pBSII-E1-p7, pBSII-HCV-

NS-3= untranslated region (UTR), and pCMV-NS5A, con-taining cDNA inserts of HCV-core, E1/E2/p7, a 6.6-kbinsert spanning partial NS2 through NS5B and the entire3=UTR sequences and NS5A sequences, respectively, havebeen described.3 The following methods for in vitro gen-ration of HCV ssRNAs and dsRNAs were used. HCVsRNAs were synthesized from linearized vectors using7 (for �strand RNAs of core, E1-P7, NS-3=UTR andstrand of NS5A RNA) and T3 (for �strand RNAs of

ore, E1-P7, NS-3=UTR and �strand of NS5A RNA) Me-aScript kits (Ambion). To form dsRNA duplexes, equalmounts of �ssRNAs and �ssRNAs (1 �g/�L) wereombined in a microcentrifuge tube in annealing buffer100 mmol/L NaCl, 20 mmol/L Tris, pH 8.0, and 1

mol/L EDTA) and incubated in 95°C water, followedy a slow cool-down to room temperature.

Determination of Isoforms of IL-28RAmRNAThe cellular RNA was extracted from isolated

pDCs, PBMCs, Huh7, and Huh7.5 cells. Then corre-

sponding cDNA was synthesized using a cDNA synthesiskit (Promega, Madison, WI) as subsequent PCR tem-plates. PCR was conducted using the following primers:the first pair, forward 5=-CTAAGCCCACCTGCTTCTTG-3=, reverse 5=-GGTTCAATGTAGGGCTGGAA-3=; or thesecond pair, forward 5=-CCTCAGTGCCAGAACCATCT-3=, reverse 5=-AAGCTGACGCCATCTTCTGT-3=. Gel elec-trophoresis was conducted to determine the alternateisoform of IL-28Ra mRNA according to the size of am-plified PCR products.

Flow CytometryFlow cytometric analysis was performed at the

Flow Cytometry Core Facility of the University of Mas-sachusetts Medical School campus using BD-LSR II (BDBiosciences), and the data were analyzed with FlowJosoftware (Tree star, Ashland, OR). The strategy of DCsubset-gating is shown in Supplementary Figure 10B.Specific cell surface markers, CD1c, BDCA3, and CD123,are used to indentify mDC1, mDC2, and pDC, respec-tively.

IL-28B SNP GenotypingGenomic DNA from blood donors was isolated

from peripheral blood using the QIAamp DNA mini kit(Qiagen). Then, the rs12979860 SNP (C/C, C/T, or T/T)was genotyped by sequencing of the 251-base pair regionafter PCR amplification using the following primers: for-ward, 5=-CCTCTGCACAGTCTGGGATT-3=; reverse, 5=-CACAATTCCCACCACGAGAC-3=.

Statistical AnalysisAll statistical analyses were performed on 3 or

more independent experiments. Data are presented as themean � SD, and differences between groups were com-pared using the Student t test or the Mann–Whitney testaccording to data distribution. A P value less than .05 wasconsidered statistically significant. In Figure 1 and Sup-plementary Figure 1, the correlation between cytokineinduction and the percentage of HCV infection in co-cultured infected Huh7.5 cells was analyzed using thelinear regression method by GraphPad Prism (La Jolla,CA) (r2 and P value are shown).

References

1. Chang S, Kodys K, Szabo G. Impaired expression and function ofToll-like receptor 7 in hepatitis C virus infection in human hepa-toma cells. Hepatology 2010;51:35–42.

2. Dolganiuc A, Kodys K, Kopasz A, et al. Hepatitis C virus core andnonstructural protein 3 proteins induce pro- and anti-inflammatorycytokines and inhibit dendritic cell differentiation. J Immunol2003;170:5615–5624.

3. Li K, Li NL, Wei D, et al. Activation of chemokine and inflammatorycytokine response in HCV-infected hepatocytes depends on TLR3sensing of HCV dsRNA intermediates. Hepatology 2011;55:666–
675.

HPf g/mL (data not shown).


Supplementary Figure 1. Interferon-inducible cytokines were secCVcc/Huh7.5 cells with different infection percentage (NS3-positive pBMCs. Cytokine production was measured by Bioplex. Each dot repre

actor-� in the co-culture supernatants was minimal and less than 20 p

Supplementary Figure 2. CXCL10, IL-29, and not inflammatory cHCV-infected cells. Human PBMCs were co-cultured with Huh7.5 cells0-hour control or collected at 8, 16, or 24 hours later. Gene induc

reted from co-cultures of human PBMCs and HCV-infected cells. (A–E)ercentage was measured by flow cytometry) were co-cultured with human

sents one data point. The induction of IL-1�, IL-10, IL-12, and tumor necrosis

ytokine genes were induced in co-cultures between human PBMCs andor HCVcc/Huh7.5 cells. Co-cultured cells were preserved before culture as

tion was measured by real-time PCR using glyceraldehyde-3-phosphate
dehydrogenase (GAPDH) as internal control (n � 3).

sc

hdpli

nonatants were measured after 24 hours by ELISA (mean � SD; n � 3).


Supplementary Figure 3. mDC2 and pDC respond to HCV-dsRNAand HCV-ssRNA, respectively, and increased expression of costimula-tory molecules. mDC2s and pDCs were collected after HCV-ssRNA orHCV-dsRNA stimulation. Expression of CD80 and CD86 on DCs weredetermined by flow cytometry. Poly I:C and CpG-A were used as pos-
itive control for mDC2 and pDC stimulation, respectively.
n � 2).

Supplementary Figure 4. PHHs produce IL-28 and IL-29 in re-ponse to JFH-1 HCVcc infection in vitro. PHHs were infected withoncentrated JFH-1 virions or stimulated with PolyI:C (20 �g/mL), PolyI:

C/Lyovec (1 �g/mL), or R848 (2 �g/mL). PHHs were maintained inepatocyte maintenance medium (PMM; Lonza), and half of the me-ium was changed every 24 hours during the experiment. Culture su-ernatants were collected at different time points, and IL-28 or IL-29

evels were determined by ELISA. Representative data from one exper-
ment is shown.
Supplementary Figure 5. Concentrated JFH-1 HCV virions alone donot induce IL-29 production. (A) Human PBMCs were co-cultured withconcentrated JFH-1 virions (HCVcc), and IL-29 induction was com-pared with control PBMCs after 4 hours by real-time PCR (mean � SD;

� 3). (B) PBMCs or mDC2 were co-cultured with HCVcc/Huh7.5 cellsr stimulated with concentrated HCVcc. IL-29 levels in culture super-

Supplementary Figure 6. The IL-29 gene is induced in co-cultures ofhuman mDC2 and HCV replicon cells. Human PBMCs were co-cul-tured with HCV subgenomic or full-length replicon cells (BB7 and FL)and co-cultured cells were collected at the beginning and 4 hours later.IL-29 mRNA levels were quantified using real-time-PCR (mean � SD;

Supplementary Figure 7. pDCs express the functional isoform ofIL-28RA mRNA. Total cellular RNA was extracted from PBMC, pDC,Huh7, and Huh7.5 cells, and cDNA was synthesized. DNA segmentsspanning the spliced gap were amplified by PCR using cDNA as tem-plate to determine the alternatively spliced form of IL-28RA mRNA in
different cell types.

tcwt

egt


Supplementary Figure 8. PBMCs from chronic HCV-infected indi-viduals produced a comparable amount of IL-29 in response to HCV-infected hepatoma cells with normal controls. IL-29 production by hu-man PBMCs in response to JFH-1/Huh7.5 or Poly I:C (20 �g/mL).PBMC cells (4 � 106) from normal control (NC) and HCV-infected pa-ients (HCV) were co-cultured with 2 � 105 JFH-1–infected Huh7.5ells. Co-culture supernatants were collected after 24 hours, and IL-29as measured by ELISA. No significant changes were observed be-

ween the NC and HCV groups (P � .05).

Supplementary Figure 9. ISG mRNA expression was increased sig-nificantly in HCV-infected livers. ISG15, ISG56, MX1, and CXCL10 genexpression in livers were determined by a 2-step real-time PCR usinglyceraldehyde-3-phosphate dehydrogenase (GAPDH) as internal con-
rol.

dHps


Supplementary Figure 10. (A) Examination of HCV infection in JFH-1 infected Huh7.5 cells. Uninfected or infected Huh7.5 cells were collected atifferent time points after JFH-1 inoculation and intracellularly stained with anti-HCV-NS3 antibody. The HCV-NS3 expression representing theCV-infection percentage was determined by flowcytometry. (B) Multicolor FACS-analysis strategy of human DC subsets in PBMCs. Total DCopulation was first gated out using Lin1 and HLA-DR, then mDC1, mDC2, and pDC were examined using CD1c, BDCA3, and CD123 as their
pecific markers as indicated.


Supplementary Table 1. Primer Sequences Used for Real-Time PCR Analysis

Gene symbol Forward primer (5=-3=) Reverse primer (5=-3=)

IL-28 TAAGAGGGCCAAAGATGCCTT CTGGTCCAAGACATCCCCCIL-29 GCCCCCAAAAAGGAGTCCG AGGTTCCCATCGGCCACATAIFN-G TCGGTAACTGACTTGAATGTCCA TCCTTTTTCGCTTCCCTGTTTTIFN-B ATGACCAACAAGTGTCTCCTCC GCTCATGGAAAGAGCTGTAGTGIFN-A1 GCCTCGCCCTTTGCTTTACT CTGTGGGTCTCAGGGAGATCAGAPDH GAAGGTGAAGGTCGGAGTC GAAGATGGTGATGGGATTTCMX1 ACAGGACCATCGGAATCTTG CCCTTCTTCAGGTGGAACACIL-28RA AAGACCCTATTTCCAGTCACTCC GAACGTGTAGATGGTTCTGGCTLR3 CAAACACAAGCATTCGGAATCTG AAGAAAGTTGTATTGCTGGTGGTTLR7 AGAACCATGTGATCGTGGACT TGTTCGTGGGAATACCTCCAGTLR8 CCAGAACGGAAATCCCGGTAT GGGTATTTGGGGTAACTGGTTGTLR9 CTGCCACATGACCATCGAG GGACAGGGATATGAGGGATTTGGDDX58 TGCGAATCAGATCCCAGTGTA TGCCTGTAACTCTATACCCATGTIFIH1 TCACAAGTTGATGGTCCTCAAGT CCTTCTCCAGATTTGGCTGAACCXCL10 AGGAACCTCCAGTCTCAGCA CAAAATTGGCTTGCAGGAATIL-1B ATGATGGCTTATTACAGTGGCAA GTCGGAGATTCGTAGCTGGAIL-6 ACTCACCTCTTCAGAACGAATTG CCATCTTTGGAAGGTTCAGGTTGIL-8 ACTGAGAGTGATTGAGAGTGGAC AACCCTCTGCACCCAGTTTTCIL-10 ATGCCCCAAGCTGAGAACCAAGACCCA TCTCAAGGGGCTGGGTCAGCTATCCCATNF-A ATCTTCTCGAACCCCGAGTGA CGGTTCAGCCACTGGAGCTISG15 GACAAATGCGACGAACCTCT CGGCCCTTGTTATTCCTCAISG56 GGGCAGACTGGCAGAAG CTATAGCGGAAGGGATTTGACD1C CCAGAATACAACATGGGTGCC TCTGTGACGCCTTCATACTGACLEC4C ACCTGCGTCATGGAAGGAAAG TCCAAGATTGCATCCCAGTAGATHBD ATTTCCTTGCTACTGAACGGC TGGTGTTGTTGTCTCCCGTAANRP1 ACGTGGAAGTCTTCGATGGAG CACCATGTGTTTCGTAGTCAGA

CD1C, cluster of differentiation 1c molecule; CLEC4C, C-type lectin domain family 4, member C; GAPDH, glyceraldehyde-3-phosphate dehydro-
genase; NRP1, neuropilin 1; THBD, thrombomodulin.

Documents

Human Type 2 Myeloid Dendritic Cells Produce Interferon-λ and Amplify Interferon-α in Response to Hepatitis C Virus Infection