ORIGINAL ARTICLE

Double-Stranded RNA-Exposed Human KeratinocytesPromote Th1 Responses by Inducing a Type-1 PolarizedPhenotype in Dendritic Cells: Role of Keratinocyte-DerivedTumor Necrosis Factor a,Type I Interferons, and Interleukin-18

M. Cristina Lebre,nw Jeanine C. Antons,nz Pawel Kalinski,n Joost H. N. Schuitemaker,nToni M. M. van Capel,n

Martien L. Kapsenberg,nw and Esther C. de JongnwnDepartment of Cell Biology and Histology, wDepartment of Dermatology, zDepartment of Pulmonology, Academic Medical Center,University of Amsterdam, Amsterdam,The Netherlands

Dendritic cells play a key role in establishing the classof immune response against invading pathogens. Uponengagement with double-stranded RNA, a majorbioactive constituent of many virus types, immaturedendritic cells develop into type 1 immunostimulatorydendritic cells that promote Th1 responses. Immaturedendritic cells reside in the epithelia and are in closecontact with keratinocytes. We studied to what extentdendritic cells can also adopt a type 1 immunostimula-tory dendritic cell phenotype indirectly, as a result ofthe interaction with keratinocytes responding to dou-ble-stranded RNA. In contrast to supernatants fromkeratinocytes activated by the combination of tumornecrosis factor a and interleukin-1b, supernatants fromkeratinocytes activated by synthetic double-strandedRNA, polyriboinosinic polyribocytidylic acid, com-prised tumor necrosis factor a and type I interferons,which induced maturation of human monocyte-derived immature dendritic cells. In addition, dendritic

cells matured in the presence of these supernatantsstrongly biased the development of Th1 cells from naiveTh cells. This bias was dependent on keratinocyte-derived interferon-a/b and interleukin-18, as neutraliza-tion of both interferon-a/b and interleukin-18 in thekeratinocyte culture supernatant reduced the develop-ment of interferon-c-producing Th cells. These ¢ndingssuggest that keratinocytes can contribute to the devel-opment of selective Th1/Th2 responses through the in-duction of maturation and functional polarization ofdendritic cells, indicating a novel role for keratinocytesas initiators and regulators of cutaneousT-cell-mediatedin£ammation. In addition, these results support theconcept that, in addition to direct interaction withpathogens, dendritic cells may also be activated andprimed by pathogen indirectly, via the e¡ect of residenttissue cells responding to pathogen. Key words: epithelialcells/innate immunity/skin/viral. J Invest Dermatol 120:990 ^997, 2003

The skin is the primary interface between the bodyand the environment and provides the ¢rst lineof defense against microbial and chemical agents.The most damaging consequence of the disruptionof skin is invasion by pathogenic microorganisms

(Robert and Kupper, 1999). Epidermal epithelial cells, keratino-cytes, represent the major constituent of the skin and participateactively in the skin immune system by producing, constitutivelyor upon stimulation, various soluble mediators, such as cytokines,chemokines, eicosanoids, and growth factors (Luger and Schwarz,1990; Fujisawa et al, 1997). For instance, several bacterial and viralcompounds induce in keratinocytes the release of the proin£am-

matory cytokine tumor necrosis factor a (TNF-a), which contri-butes to local in£ammatory reactions within the epidermis (K˛cket al, 1990; Ezepchuk et al, 1996).Speci¢c skin in£ammation, such as infections, contact

hypersensitivity, and skin graft rejection, is regulated by a com-plex and sequential mechanism involving dendritic cells (DC),keratinocytes, and T cells, and a variety of soluble mediatorsthey produce (Grabbe et al, 1991; 1992). Immature dendriticcells (iDC) are located in peripheral tissues (e.g., epidermalLangerhans cells and dermal DC) in close contact with keratino-cytes. DC are the only antigen-presenting cells that canprime naive Th cells and initiate immune responses. Afterexposure to antigens, activated DC migrate from peripheraltissues to T cell areas of the draining lymph nodes. Duringthis migration, iDC undergo maturation from cells thathave the capacity to take up antigen into potent immunostimula-tory e¡ector cells (Steinman, 1991; Banchereau and Steinman,1998).DC maturation can be induced in vitro by several factors in-

cluding the combination of the proin£ammatory cytokines inter-leukin-1b (IL-1b) and TNF-a, type I interferons (IFNs), variousmicrobial compounds, or CD40 ligand (CD40L) (Cella et al,

Reprint requests to: Esther C. de Jong, Academic Medical Center, Uni-versity of Amsterdam, Department of Cell Biology and Histology, POBox 22700, 1100 DE Amsterdam, The Netherlands; Email: E.C.deJong@AMC.UvA.NLAbbreviations: CD40L, CD40 ligand; DC, dendritic cells; iDC, imma-

ture dendritic cells; MF, maturation factors; poly I:C, polyriboinosinicpolyribocytidylic acid; SEB, Staphylococcus aureus enterotoxin B.

Manuscript received November 6, 2002; revised January 7, 2003;accepted for publication January 21, 2003

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

990

1996; 1999; Verhasselt et al, 1997; Luft et al, 1998; 2002; Sparwasseret al, 1998). During their maturation, DC may also gain the capa-city to polarize naiveTh cells intoTh1 or Th2 cells. This ability isdetermined in peripheral tissues by the type of invading patho-gen or their bioactive compounds, and is established either uponthe direct activation of iDC by pathogen (Paschen et al, 2000;Rescigno et al, 2000;Whelan et al, 2000), or upon the indirect ac-tivation by mediators of the local in£ammatory response of tissuecells in response to pathogen (Kalin� ski et al, 1999a; Vieira et al,2000).Double-stranded RNA (dsRNA) is an intermediate within

the replication cycle for RNA viruses and some DNA viruses,and is their major bioactive component (Colby and Duesberg,1969; Haines et al, 1991). Polyriboinosinic polyribocytidylic acid(poly I:C) is a synthetic dsRNA that is often used as a model ofviral infection.Considering the fact that iDC reside in the close vicinity of

keratinocytes, these cells may play an important role in the activa-tion and polarization into Th1 (DC1)- or Th2 (DC2)-promotinge¡ector DC. This is corroborated by the study of Pastore et al(1997) in which it was demonstrated that granulocyte macro-phage colony stimulating factor (GM-CSF) produced by kerati-nocytes enhanced the survival of DC. Therefore, we studiedwhether and how factors produced by keratinocytes, upon activa-tion by the proin£ammatory cytokines TNF-a/IL-1b or by a mi-mic for a virus infection, activate and/or polarize DC.

MATERIALS AND METHODS

Culture of human keratinocytes Primary cultures of normalhuman keratinocytes were prepared from neonatal foreskins or adult skinundergoing plastic surgery. The skin was incubated with thermolysin (500mg per ml; Sigma-Aldrich, St. Louis, MO) for 16 h at 41C. Epidermalsheets were removed from the dermis and single cell suspensionswere obtained by placing epidermal sheets in trypsin (0.025%; LifeTechnologies, Paisley, U.K.) for 5 min at 371C. After neutralizing with anequal volume of fetal bovine serum (FBS) (HyClone, Logan, UT) stratumcorneum debris was removed and then sieved through sterile nylon gauzeto obtain a single cell suspension. Isolated epidermal cells were seeded at adensity of 8^10�104 cells per cm2. At 70%^80% con£uence, keratinocyteswere detached with 0.025% trypsin, 2 mM ethylenediamine tetraacetic acidfor 5 min at 371C, and subcultured or frozen. Keratinocyte cultures weremaintained in Keratinocyte SFM (Gibco, Paisley, U.K.).

Stimulation of human keratinocytes Human keratinocytes wereplated at a concentration of 2.5�104 cells per ml (six-well plates, Costar,Cambridge, MA) and after 48 h the cells were pulsed for 2 h with polyI:C (200 mg per ml; Sigma-Aldrich). Keratinocytes were then washedextensively [¢ve times in 5 ml of phosphate-bu¡ered saline (PBS), i.e.,41�106 times diluted], fresh medium without hydrocortisone (5 ml)was added, and the cells were cultured for an additional 46 h.Keratinocytes were also pulsed with the proin£ammatory cytokinesrecombinant human (rhu) TNF-a (50 ng per ml, PBH, Hannover,Germany) and rhuIL-1b (100 ng per ml; PBH). The culture supernatantswere collected after the indicated time, centrifuged to remove cells, andstored at 41C before use or stored at ^201C for cytokine measurements.The concentrations of IL-8, TNF-a, IFN-a, and IL-18 in 46 hsupernatants were determined by enzyme-linked immunosorbent assay(ELISA) (see below). Keratinocytes were also stimulated for 4 h with thestimuli stated above; the cells were lyzed for examination of IFN-a andIFN-b gene expression by reverse transcription polymerase chain reaction(RT-PCR) (see below).

RNA isolation and cDNA synthesis Keratinocyte total RNA waspuri¢ed by using the NucleoSpins RNA II kit (Macherey-Nagel, Dˇren,Germany) according to the manufacturer’s instructions. ComplementaryDNA was generated using the ¢rst strand cDNA synthesis kit for RT-PCR (MBI Fermentas, St. Leon-Rot, Germany). To anneal the primer tothe RNA, 9 ml of total RNA, 1 ml oligo(dT)18, and 1 ml D(N)6 were added.This mix was then heated for 5 min at 941C.

PCR analysis The primer sequences were as follows: IFN-a2, sense, 50 -AGTCAAGCTGCTCTGTGGGC-30; antisense, 50GTGAGCTGGCATACGAATCA-30, de¢ning a 571 bp product; HPLC-treated IFN-b, sense, 50 -GATTCATCTAGCACTGGCTGG-30; antisense, 50 -CTTCAGGTAATGC

AGAATCC-30, de¢ning a 186 bp product. Both primers have anannealing temperature of 601C and 45 cycles. PCR products wereanalyzed on a 1% agarose gel containing ethidium bromide. A 100 bpDNA ladder standard (MBI Fermentas) was used as a size marker.

Generation of iDC from peripheral blood monocytes iDC weregenerated from monocytes in culture in Iscove’s modi¢ed Dulbecco’smedium (IMDM; Life Technologies) containing 1% FBS (HyClone),rhuGM-CSF (500 U per ml; Schering-Plough, Uden, The Netherlands),and rhuIL-4 (250 U per ml; PBH), as described previously (Sallusto andLanzavecchia, 1994; Kalin� ski et al, 1998).

E¡ect of keratinocyte supernatants on DC In order to study thee¡ect of keratinocyte-derived soluble mediators on DC, supernatants fromnonactivated or TNF-a/IL-1b- or poly I:C-activated keratinocytes wereadded to iDC at a concentration of 50% (vol/vol) of the culture. Ascontrols, iDC were exposed to keratinocyte culture medium orlipopolysaccharide plus TNF-aþ IL-1b [LPSþmaturation factors (MF)].After 48 h, DC were harvested and analyzed for their maturation statusby cell surface expression determined by £uorescence-activated cell sorter(FACS), their cytokine production upon CD40 ligation, their T cellstimulatory capacity, and their capacity to induce the development of Th1or Th2 cell responses in naive precursors. When indicated, the followingwere added to keratinocyte supernatants before addition to iDC:neutralizing antibodies against TNF-a (10 mg per ml, IgG1; DiacloneResearch, Besanc� on, France), or two neutralizing sheep antisera to humantype I IFN [Iivari (450,000 neutralizing U per ml anti-IFN-aþ 3000 U perml anti-IFN-b) and Kaalepi (30,000 U per ml anti-IFN-aþ 30,000 U perml anti-IFN-b)] (Mogensen et al, 1975), or IL-18 binding protein (IL-18 bp;10 mg per ml, Amgen, Thousand Oaks, CA). The following antibodies orserum were used as negative control Ig: sheep serum (Sigma-Aldrich) andmouse IgG1 (10 mg per ml, MOPC-21; Sigma-Aldrich).

Analysis of cell surface expression The expression of cell surfacemolecules associated with maturation of DC was determined usingmonoclonal antibodies against the following surface markers: CD1b(IgG1; Diaclone Research), CD86 (BD Pharmingen, San Diego, CA),CD80 (Pharmingen), CD83 (HB15a, IgG2b; Immunotech, Marseille,France). Fluorescein isothiocyanate (FITC) coupled goat F(ab0)2 antimouseIgG and IgM (Jackson ImmunoResearch Laboratories, West Grove, PA)was used as a secondary reagent. Samples were analyzed on a FACScan(Becton Dickinson).

IL-12p70 production by iDC iDC were washed and stimulated(2�104 cells in 200 ml) in 96-well £at-bottomed culture plates (Costar) inIMDM containing 10% FBS with either LPS (250 ng per ml; Sigma-Aldrich) and IFN-g (103 U per ml) or with irradiated (2500 Gy) CD40L-transfected J558 plasmacytoma cells (J558-CD40L, 2�104 cells per well;a gift of Dr. P. Lane, Birmingham, U.K.) in the absence or in the presenceof 50% (vol/vol) keratinocyte supernatants. Supernatants were harvestedafter 24 h, and the concentrations of IL-12p70 were measured by ELISA(see below).

IL-12p70 production by Dc matured in the presence of keratinocytesupernatants As IL-12 production by DC is dependent on theirmaturation status (Kalinski et al, 1999b), DC were matured by the additionof keratinocyte supernatants (50% vol/vol) in the presence of acombination of rhuIL-1b (10 ng per ml; PBH), rhuTNF-a (50 ng per ml;PBH) (MF), and LPS (250 ng per ml; Sigma-Aldrich) to iDC. As a controliDC were matured in the presence of IL-1b plus TNF-a plus LPS. On day8, DC were washed and 2�104 cells per well were stimulated in 96-well£at-bottomed culture plates (Costar) in IMDM containing 10% FBS in a¢nal volume of 200 ml with irradiated (2500 Gy) CD40L-transfected J558cells (2�104 cells per well), which were previously shown to induce IL-12p70 in an IFN-g-independent way (Cella et al, 1996). Supernatants wereharvested after 24 h, and the concentrations of IL-12p70 were measured byELISA.

Isolation of CD4þCD45RAþCD4RO^ naive Th cells Naive Thcells were isolated from peripheral blood leukocytes with the negativeselection human CD4þ/CD45RO^ column kit (R&D Systems,Minneapolis, MN). This method yielded highly puri¢ed (498%)CD4þCD45RAþCD45RO^ naive Th cells as assessed by £ow cytometry(data not shown).

Mixed lymphocyte reaction Mature DC were also tested for theirability to stimulate allogeneic naive Th cells in a mixed lymphocytereaction. Naive Th cells (2.5�104 cells per 200 ml) were cocultured in 96-well £at-bottomed culture plates with di¡erent concentrations of mature

DOUBLE-STRANDED RNA-ACTIVATED HUMAN KERATINOCYTES INDUCE DC1 991VOL. 120, NO. 6 JUNE 2003

DC. After 5 d, cell proliferation was assessed by the incorporation of[3H]thymidine (Radiochemical Center, Amersham, Little Chalfont, U.K.)after a pulse with 13 kBq per well during the last 16 h, as measured byliquid scintillation spectroscopy.

Induction of Th1 or Th2 cell responses by mature DC Naive Thcells (2�104 cells per 200 ml) were cocultured in 96-well £at-bottomedculture plates with heterologous DC (0.5�104 cells per 200 ml) maturedin the di¡erent conditions stated above and coated with the superantigenStaphylococcus aureus enterotoxin B (SEB, 100 pg per ml; Sigma-Aldrich).On day 5, IL-2 (10 U per ml; Cetus, Emeryville, CA) was added and thecultures were further expanded. After 14 d, resting memoryTh cells wereharvested, washed, and restimulated for 6 h with phorbol myristate acetate(PMA, 10 ng per ml; Sigma-Aldrich) and ionomycin (1 mg per ml; Sigma-Aldrich) in the presence of brefeldin A (10 mg per ml; Sigma) in order toanalyze Th cell cytokine production pro¢le on a single cell basis. The cellswere ¢xed in paraformaldehyde (2%; Sigma-Aldrich), permeabilized withsaponin (0.5%; Sigma-Aldrich), and labeled with FITC-coupled IFN-gmonoclonal antibody (Becton Dickinson) and phycoerythrin-coupled IL-4 monoclonal antibody (Becton Dickinson). The cells were evaluated byFACScan (Becton Dickinson).When indicated, DC were cocultured withnaiveTcells in the presence of antibodies against IL-12 (10 mg per ml, rabbitpolyclonal; a gift of Dr. P. van der Meide, U-CyTech, Utrecht, TheNetherlands).

Evaluation of cytokine production by ELISA Determination of IL-12p70 concentrations in culture supernatants was performed by speci¢csolid-phase sandwich ELISA as previously described (Kalinski et al, 1998).To determine IL-8, TNF-a, and IFN-a in culture supernatants, pairs ofspeci¢c monoclonal antibodies and recombinant cytokine standards wereobtained from BioSource International (Camarillo, CA). The detectionlimits of these ELISA are as follows: IL-8, 30 pg per ml; TNF-a, 20 pgper ml; IFN-a, 100 pg per ml; IL-12p70, 3 pg per ml. To determine IL-18in culture supernatants, an IL-18 ELISA kit was obtained from DiacloneResearch (detection limit 62.5 pg per ml).

Statistical analysis Data are expressed as mean 7SD. Data wereanalyzed for statistical signi¢cance with the GraphPad InStats software(version 3.00; GraphPad InStat, San Diego, CA) using ANOVA followedby Dunnett’s multiple comparisons test. A p-value o0.05 was consideredas the level of signi¢cance.

RESULTS

Poly I:C-activated, but not TNF-a/IL-1b-activated,keratinocytes induce DC maturation: role of TNF-a andtype I IFNs To evaluate whether keratinocyte-derived

products can contribute to DC maturation, iDC were exposedfor 48 h to supernatants (50% vol/vol) from keratinocytes thatwere pulsed for 2 h with poly I:C- or TNF-a/IL-1b andcultured for a further 46 h. The purity of the keratinocytes, asveri¢ed by the expression of the epithelial marker cytokeratineand the ¢broblast-speci¢c marker (ASO2), was more than 99%(data not shown). DC maturation was monitored by testing forthe induction of cell surface expression of CD83 and for theupregulation of expression of the costimulatory moleculesCD80 and CD86. In a comparative assay with iDC matured incontrol conditions with the combination of LPS and MF(consisting of TNF-a and IL-1b), we found that exposure ofiDC to supernatants from nonactivated or TNF-a/IL-1b-activated keratinocytes did not induce their maturation (Fig 1).This immature phenotype was similar to the phenotype ofcultured iDC, left untreated or treated with keratinocytemedium. In contrast, poly I:C-activated keratinocyte super-natants induced DC maturation in the majority of DC, as evidentfrom the induction of CD83 expression, accompanied byupregulation of the costimulatory molecules CD80 and CD86(Fig 1).Next, we questioned which factor(s) exclusively produced by

poly I:C-activated keratinocytes are responsible for the DCmaturation. As TNF-a (Uchi et al, 2000) and type I IFNs(Fujisawa et al, 1997) can be readily expressed by keratinocytesand these soluble factors can induce DC maturation (Verhasseltet al, 1997; Luft et al, 1998; 2002), the role of TNF-a and type IIFNs in keratinocyte-induced DC phenotypical changes wasevaluated by adding neutralizing antibodies, or their respectivecontrol Ig or serum, to keratinocyte supernatants before theiraddition to the DC cultures. Surprisingly, neutralization ofeither of these soluble factors strongly reduced the induction ofCD83, CD80, and CD86 expression by DC (Fig 1), suggestingthat both TNF-a and IFN-a/b produced by poly I:C-activatedkeratinocytes are involved in the induction of DC maturation.Addition of both antibodies against TNF-a and IFN-a/b tokeratinocyte supernatants did not result in further inhibition ofCD83, CD80, and CD86 expression by DC treated with polyI:C-activated keratinocyte supernatants.The role of TNF-a and type I IFNs from supernatants of poly

I:C-activated keratinocytes in DC maturation was furthercon¢rmed by studying the expression of these cytokines bydi¡erently stimulated keratinocytes. Therefore, keratinocyteswere pulsed for 2 h with either poly I:C or TNF-a/IL-1b and

Figure1. Supernatants from poly I:C-activated, and not TNF-a/IL-1b-activated, keratinocytes induce DC maturation. iDC were cultured for48 h in the presence of 50% (vol/vol) supernatants from nonactivated or poly I:C- or TNF-a/IL-1b-activated keratinocytes and assayed for maturation bydetermining the expression of CD83, CD80, and CD86 by FACS analysis. As a control, DC were matured in the presence of a combination of LPS and thematuration factors TNF-a and IL-1b (MF). In contrast to no additions, keratinocyte medium, nonactivated or TNF-a/IL-1b-activated keratinocyte super-natants, exposure of iDC to poly I:C-activated keratinocyte supernatants enhanced the expression of CD83, CD80, and CD86 to levels comparable tocontrol DC. The role of TNF-a and IFN-a/b in DC maturation was tested by adding neutralizing antibodies to TNF-a (10 mg per ml) or IFN-a/b [twoneutralizing sheep antisera to human type I IFN were used: Iivari (450,000 neutralizing U per ml anti-IFN-aþ 3000 U per ml anti-IFN-b) and Kaalepi(30,000 U per ml anti-IFN-aþ 30,000 U per ml anti-IFN-b) (1567)], or their respective control Ig (CIg and sheep serum, respectively), to the supernatantsfrom poly I:C-activated keratinocytes, prior to the administration of these supernatants to the DC cultures. Addition of these antibodies to poly I:C-acti-vated keratinocyte supernatants blocked the induction of CD83, CD80, and CD86 expression by DC. DMFI represents the di¡erence between the variousstainings and the isotype control. Results are for one representative experiment out of three.

992 LEBRE ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

cultured for an additional 46 h, and the supernatants wereanalyzed by ELISA. Whereas poly I:C- and TNF-a/IL-1b-activated keratinocytes produced similar levels of IL-8 (Fig 2B),only keratinocytes activated by poly I:C produced TNF-a(Fig 2A). Type I IFNs were undetectable by standard ELISAin all groups (data not shown). RT-PCR analysis, however,readily revealed that only poly I:C-activated keratinocyteswere able to upregulate mRNA expression for IFN-a and IFN-b(Fig 2D).

DC matured in the presence of poly I:C- or TNF-a/IL-1b-activated keratinocyte supernatants exhibit similarcapacities to support naive Th cell proliferation Wesubsequently studied whether the exposure of maturing DC tokeratinocyte supernatants a¡ects their ability to stimulateallogeneic naive Th cells in mixed lymphocyte reactionconditions. As the capacity of DC to induce a particular Thdevelopment is dependent on their maturation status we addedLPSþMF to all DC cultures to synchronize their maturation.As shown in Fig 3, maturing DC exposed to nonactivated orpoly I:C- or TNF-a/IL-1b-activated keratinocyte supernatantswere as e¡ective in inducing Th cell proliferation as DCmatured with LPSþMF only. As expected, iDC were lesse¡ective in inducing Th cell proliferation. Although theproliferative response of iDC compared to all mature DCgroups was statistically signi¢cant, the comparison of theproliferative response between the mature DC groups with eachother did not reach the level of signi¢cance. These results indicatethat the exposure of DC to keratinocyte supernatants does nota¡ect their ability to stimulate naive Th cell proliferation.

DC matured in the presence of poly I:C-activatedkeratinocyte supernatants acquire the ability to induceTh1-like responses: role of type I IFNs and IL-18 To testwhether the presence of keratinocyte supernatants during

DC maturation primes these cells for the capacity to bias forthe development of Th1 or Th2, synchronized mature DCwere cocultured with naive Th cells in the presence ofsuperantigen, as previously described (Kalinski et al, 1998). After14 d, the percentage of cells producing IL-4 and/or IFN-g wasevaluated by stimulation of e¡ector cells with PMA/ionomycinand analysis of intracellular cytokine expression on a single cellbasis by FACS. Although the priming of naive Th cells with DCmatured in control conditions (LPSþMF) or DC cultured in theadditional presence of supernatants from nonactivated or TNF-a/IL-1b-activated keratinocytes resulted in the development of amixed population of Th1 and Th2 cells after restimulation, DCmatured in the presence of poly I:C-activated keratinocytesupernatants induced a strong bias toward the development ofTh1 cells (Fig 4).Type I IFNs and IL-18 have been implicated in the virus-

induced development of Th1-type immune responses (Sarenevaet al, 1998). Indeed, we found that poly I:C-activatedkeratinocytes not only produce type I IFNs (Fig 2D) butalso IL-18 (Fig 2C). In addition, neutralization of type I IFNs orIL-18 in the supernatants of poly I:C-activated keratinocytespartially inhibited Th1 polarization (Fig 5A). When thetwo soluble factors were neutralized, however, we observedno further decrease in the percentage of IFN-g-producingT cells to levels close to neutral conditions. These data indicatethat poly I:C-activated keratinocytes produce both type I IFNsand IL-18 that prime for DC inducing the development ofTh1 cells.

Figure 2. Poly I:C-activated, and not TNF-a/IL-1b-activated, kera-tinocytes produce signi¢cant amounts of TNF-a and IL-18, and ex-press IFN-a and IFN-b. Normal human keratinocytes were pulsed for2 h with either poly I:C or TNF-a/IL-1b, washed extensively (¢ve times in5 ml), and cultured for an additional 46 h in the presence of fresh medium.Supernatants were harvested and TNF-a (A), IL-8 (B), IL-18 (C), and IFN-a (data not shown) were determined by ELISA (see Materials and Methods).(D) shows the expression of IFN-a and IFN-b mRNA in 4 h stimulatedkeratinocytes. RNAwas extracted for RT-PCR analysis. Each ¢gure showsone representative out of three independently performed experiments.

Figure 3. DC matured in the presence of poly I:C- or TNF-a/IL-1b-activated keratinocyte supernatants exhibit similar capacities tosupport naive Th cell proliferation. Maturation of DC was induced byeither a combination of keratinocyte supernatants (50% vol/vol)and LPSþMF or LPSþMF alone. iDC or di¡erentially mature DCwere cultured at di¡erent numbers with 2.5�104 allogeneic naive(CD45RAþ )CD4þ Th cells. After 5 d, the cells were pulsed duringthe last 16 h with [3H]thymidine ([3H]-TdR). iDC (open triangles);DCþLPSþMF (closed triangles); DCþLPSþMFþ keratinocyte superna-tants, unstimulated (open circles), poly I:C (closed circles), TNF-a/IL-1b (opensquares). Results are expressed in counts per minute (cpm). Data were ana-lyzed for statistical signi¢cance using ANOVA followed by Dunnett’s mul-tiple comparisons test using iDC as control. npo0.05, nnpo0.01. Resultsare representative of three independent experiments.

DOUBLE-STRANDED RNA-ACTIVATED HUMAN KERATINOCYTES INDUCE DC1 993VOL. 120, NO. 6 JUNE 2003

Th1 polarization induced by poly I:C-activated keratinocytes is partially mediated by DC-derived IL-12 IL-12 is amajor Th1-driving cytokine, promoting cell-mediated immunity(O’Garra, 1998), and can be produced by DC and other antigen-presenting cells upon CD40 ligation by CD40L expressed byactivated T cells (Cella et al, 1996). An indication for a role of IL-12 in the enhanced Th1 polarization by poly I:C-activatedkeratinocytes was the ¢nding that the presence of suchsupernatants, and not supernatants from TNF-a/IL-1b-activatedkeratinocytes, enhanced the production of IL-12p70 in CD40L-stimulated iDC (Fig 6A). This enhanced production was blockedentirely by the additional presence of neutralizing antibodiesagainst IFN-a/b in the keratinocyte supernatants (Fig 6A),indicating that keratinocytes enhance IL-12p70 production byCD40L-activated iDC via the release of IFN-a/b.It remained to be established, however, to what extent

keratinocyte-derived supernatants were able to prime forenhanced IL-12 production, i.e., do DC matured in the presence

of keratinocyte-derived supernatants show enhanced IL-12 inresponse to CD40L? As the maturation status of DC mayreadily determine their ability to produce IL-12 (Kalin� ski et al,1999b), these experiments were performed with DC with asynchronized maturation status by adding LPS and MF to allthe iDC groups, which does not alter the ability to produce IL-12 in the mature DC (Kalinski et al, 1998). After 48 h, the matureDC were extensively washed and restimulated for 24 h withCD40L. In contrast to the above-mentioned experiments withiDC, the CD40L-induced IL-12p70 production by DC maturedin the presence of poly I:C-activated keratinocyte supernatantswas not signi¢cantly di¡erent from that of DC matured in thepresence of supernatants from nonactivated or TNF-a/IL-1b-activated keratinocytes (Fig 6B). In addition, the capacity for IL-12p70, IL-6, and TNF-a production by di¡erently matured DCupon CD40 activation was not altered, or only marginallyaltered, upon neutralization of type I IFNs or IL-18 in thekeratinocyte supernatants (Fig 6B).In order to analyze the contribution of the moderate levels of

T-cell-induced DC-derived IL-12 in the enhanced Th1polarization, neutralizing antibodies against IL-12 were added tococultures of DC and naive Th cells. As depicted in Fig 5(B),neutralization of IL-12 inhibited Th1 polarization in anycondition, including Th1 polarization induced by DC maturedin the presence of poly I:C-activated keratinocyte supernatants.This e¡ect was more evident when simultaneously type I IFNsand/or IL-18 were neutralized in the keratinocyte supernatants,indicating that both soluble factors produced by poly I:C-activated keratinocytes and IL-12p70 produced by mature DCare involved in the observed Th1 polarization.

DISCUSSION

In this study it is demonstrated for the ¢rst time that solublemediators from pathogen-activated keratinocytes can profoundlya¡ect DC function and consequently the class of the immune re-sponse by demonstrating that supernatants from poly I:C-acti-vated keratinocytes, but not from TNF-a/IL-1b-activatedkeratinocytes, induce iDC to mature into e¡ector DC, and thatthese DC appear to be type 1 DC (DC1) that bias the develop-ment of Th1 cells from naive T cells.

Keratinocytes and viral infections Human keratinocytes arethe target of several viruses such as herpes simplex virus(Mikloska et al, 1996; 1998), human papillomavirus (HPV) (Choet al, 2001), and varicella-zoster virus (Nahass et al, 2001). Inparticular, it was reported that infection of human keratinocyteswith HPV type 16 (Nees et al, 2001) or herpes simplex virus(Mikloska et al, 1998) induces the secretion several solublemediators. Moreover, it was suggested that herpes simplex virusdoes not directly infect DC but epithelial cells (Kumaraguru andRouse, 2002), indicating a crucial role for keratinocyte-derivedfactors in this viral infection.

Figure 5. Both IFN-a/b and IL-18 produced by poly I:C-activatedkeratinocytes and IL-12 produced by mature DC contribute to thedevelopment of Th1 cells. DCwere matured as described in Fig 3. After48 h, DC were thoroughly washed, loaded with SEB, and cocultured withnaiveTh cells. After 14 d, respondingTh cells were analyzed as described inFig 4. (A) Blockage of both IFN-a/b and/or IL-18 in poly I:C-activatedkeratinocyte supernatants reduced the frequency of Th cells expressingIFN-g. This e¡ect was more evident when neutralizing antibodies againstIL-12 were added to DC/Th naive cocultures (B). This ¢gure shows onerepresentative out of four independently performed experiments.

Figure 4. DC matured in the presence of supernatants from poly I:C-activated keratinocytes have the ability to induce Th1-like responses.Maturation of DC was induced as indicated in Fig 3. Di¡erentially matured DC were loaded with SEB and cocultured with naive Th cells. After 14 d,resting memoryTh cells were restimulated for 6 h with PMAþ ionomycin in the presence of brefeldin A. The expression of IL-4 and IFN-g was assessedby intracellular staining. Results are expressed as the percentage of cells in each population. Data are representative of six independent experiments.

994 LEBRE ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

DC maturation induced by keratinocyte-derived TNF-a andIFN-a Synthetic dsRNA (poly I:C) is often used as a model ofviral infection. It has been reported that poly I:C inducesmaturation of human DC (Cella et al, 1999; Verdijk et al,1999) via binding to Toll-like receptor 3 (Alexopoulou et al,2001). In addition, poly I:C-matured DC acquired the capacityto trigger naive T cells and drive polarized Th1 responses (Cellaet al, 1999), a phenomenon that is partially dependent on IL-12(de Jong et al, 2002). Our experiments extended theseobservations by demonstrating that poly I:C can also modulateDC function via the activation of keratinocytes. Indeed, in aseparate study we found that keratinocytes express various TLRtypes, including high levels of TLR3 (Lebre et al, submitted).Experiments with neutralizing antibodies indicated that polyI:C-activated keratinocytes induce DC maturation throughthe production of TNF-a and IFN-a/b, two well-knownDC-maturation-inducing factors (Verhasselt et al, 1997). Theimportance of TNF-a-dependent maturation of local DC inactivating adaptive immune response to virus infection issupported by a study showing that DC from TNF-de¢cientmice cannot mature in response to virus infection (Trevejo et al,2001).

IFN-a- and IL-18-dependent Th1 response In a search forthe soluble factors driving the type 1 phenotype of DC maturedin response to the supernatants from poly I:C-activated DC, weidenti¢ed the cytokines type I IFNs and IL-18. Upregulation ofexpression of type I IFNs is one of the earliest cellular responsesupon contact with infectious agents, in particular viruses. Therapid induction of type I IFNs re£ects the crucial role that thesecytokines play in the inhibition of viral spread before thegeneration of a speci¢c immune response (De Maeyer andMaeyer-Guignard, 1998). In addition, because of this earlyexpression, the type I IFNs are ideal signaling molecules alertingthe speci¢c immune system to the presence of viral infection,which requires protective Th1 cell responses. Thus, type I IFNshave a potent antiviral function that is due not only to theirdirect e¡ects on infected cells (i.e., by inhibiting virusreplication), but also to their indirect e¡ects (i.e., by stimulationof iDC maturation resulting in Th1-biased responses) (Bogdan,

2000). Consistently, keratinocyte-derived type I IFNs belong tothe molecules linking innate and adaptive immunity.Importantly, the DC maturation observed is not due to

carryover of poly I:C present in keratinocyte supernatants as theaddition of supernatants derived from paraformaldehyde-¢xedkeratinocytes, after poly I:C activation, did not result inupregulation of CD83, CD80, and CD86 by DC (data notshown). Moreover, after exposure to poly I:C, keratinocyteswere washed ¢ve times in 5 ml of PBS, i.e.,41�106 timesdiluted.When the last wash (PBS) was added to iDC for 48 h,no induction of DC maturation was observed (data not shown).Altogether, these data suggest that keratinocyte-secreted factor(s),rather than poly I:C carryover, induce(s) DC maturation.IL-18 is widely expressed by both leukocytes and

nonleukocytes, and IL-18 expression in keratinocytes has beenshown before (Companjen et al, 2000a; 2000b). One of the mostwell known activities of IL-18 is the stimulation of IFN-gproduction during in£ammation, in particular the IL-12-inducedIFN-g production byTh1 cells (Ahn et al, 1997; Kohno et al, 1997).Our data are in line with an earlier report proposing thatkeratinocyte-derived IL-18 is involved in the cutaneous Th1-typeimmune response (Naik et al, 1999).

Role of DC-derived IL-12 We found that Th1 polarizationinduced by poly I:C-activated keratinocytes is partially mediatedby DC-derived IL-12. IL-12 is an important factor in theinduction of Th1 responses (O’Garra, 1998), although DC-derived IFN-a and IL-18 have also been implicated (Sarenevaet al, 1998). IL-12 is produced by both iDC and mature DC uponCD40 ligation (Cella et al, 1996; Koch et al, 1996; Hilkens et al,1997; Kalinski et al, 1997) and is regulated by cytokines andprostaglandin E2 (Vieira et al, 2000; Kalinski et al, 2000; 2001).Here, we have shown that type I IFNs from poly I:C-activatedkeratinocytes enhance CD40L-induced IL-12 production iniDC. Surprisingly, these keratinocytes did not prime forenhanced IL-12 production in CD40L-activated mature DC.This discrepancy is also found when DC are directly activatedby poly I:C, which induces high levels of IL-12 in iDC (Cellaet al, 1999; Verdijk et al, 1999) but not does not prime for highIL-12 in mature DC (de Jong et al, 2002). The enhanced IL-12

Figure 6. Poly I:C-activated keratinocyte supernatants enhanced IL-12 production by iDC upon CD40 ligation whereas it did not prime forhigh IL-12 production in mature DC. (A) iDC were stimulated with CD40L-transfected J558 cells (each 20,000 cells per well) in the absence or thepresence of poly I:C- or TNF-a/IL-1b-activated keratinocyte supernatants (50% vol/vol). In addition, supernatants of poly I:C-activated keratinocytes in-cubated with neutralizing antibodies against IFN-a/bwere added to iDC. Poly I:C-activated keratinocytes enhance bioactive IL-12 production by CD40L-stimulated iDC via the release of type I IFNs. (B) IL-12p70, IL-6, and TNF-a production by mature DC. Di¡erentially activated keratinocyte supernatantsdo not modulate the bioactive IL-12 production of mature DC. Maturation of DC was induced by either a combination of keratinocyte supernatants (50%vol/vol) and LPSþMF or LPSþMF alone. After 48 h DCwere thoroughly washed and stimulated with CD40L-transfected J558 cells. IL-12p70, IL-6, andTNF-a concentrations in 24 h supernatants were determined by ELISA. Results, expressed as mean 7SD of triplicate cultures, are from one representativeexperiment out of three with similar results. Data were analyzed for statistical signi¢cance using ANOVA followed by Dunnett’s multiple comparisons testusing DC matured in the presence of LPSþMFþ unstimulated keratinocyte supernatants as control. npo0.05, nnpo0.01.

DOUBLE-STRANDED RNA-ACTIVATED HUMAN KERATINOCYTES INDUCE DC1 995VOL. 120, NO. 6 JUNE 2003

production by iDC may be primarily important for the optimaland rapid activation of local e¡ector cells, e.g., primed Th cells,class-I-restricted cytotoxic T lymphocytes, or natural killer cells,to eliminate virus-infected keratinocytes. The lack of enhancedIL-12 production in DC primed by poly I:C-activatedkeratinocytes and the partial inhibition of the development ofTh1 cells indicate that additional DC-derived factors havecontributed to the development of these Th1 cells. As it has beensuggested that dsRNA and in£uenza virus can induce DCmaturation and confer the capacity to prime a Th1 responsethrough the production of both IL-12 and type I IFNs (Cellaet al, 1999) and that DC-derived IL-18 enhances the IL-12-induced production of IFN-g by Th1 cells (Ahn et al, 1997;Robinson et al, 1997), type I IFNs and IL-18 are likely candidates.Clearly, the direct activation of iDC by pathogen will contributeto the initiation and regulation of the speci¢c immune responseto this pathogen by committing the functional phenotype of theDC after their maturation.

DC polarizing capacity: role of microenviroment It hasbeen proposed that tissue-derived ‘‘danger’’ signals can also bevital in this respect (Matzinger, 1994). The nature of these signalsis dependent on the type of tissue and type of pathogen involved.The data suggest that viruses can commit DC indirectly tobecome DC1 by the activation of keratinocytes located in thevicinity of DC. It may be speculated that other stromal cells ofthe skin, e.g., ¢broblasts that are also capable of producing TNF-a or IFN-a in response to virus infection, may also be able toinitiate maturation and polarization of DC. Recently, Soumeliset al (2002) reported that thymic stromal lymphopoietin (TSLP)is elevated in human keratinocytes from atopic dermatitis patientsand that TSLP can induce the development of a Th2-cell-promoting DC phenotype DC2, characterized by the capacityto bias naive Th cells into Th2 cells. The authors thereforeproposed that enhanced levels of TSLP may contribute to theallergic in£ammation in atopic dermatitis. In contrast to ourstudy, however, this study does not show direct evidence thatTSLP present in keratinocyte culture supernatants is a factorcrucial in inducing the e¡ector DC2 phenotype. Our data andthe data from the study of Soumelis et al (2002) were obtainedby targeting monocyte-derived DC and peripheral bloodCD11cþ DC, respectively, and it was hypothesized that such DCcomprise precursors of epidermal DC (Ito et al, 1999). Although itremains to be established in full detail, Langerhans cells may alsobe responsive to keratinocyte-derived signals, as all DCpopulations assayed so far have shown a high degree of plasticityin response to environmental stimuli (Shortman and Liu, 2002).

Concluding remarks In summary, this study demonstratesthat pathogen-induced keratinocyte-derived factors can activateDC and modulate their function. These data support the notionthat iDC as sentinels of speci¢c immunity can respond topathogen-associated molecules not only directly via patternrecognition receptors (Janeway, 1992), but also indirectly, as aresult of their reactivity to factors produced by tissue cells inresponse to those pathogens (Matzinger, 1994).

This work was supported by Fundac� a� o para a CieŒ ncia e aTecnologia, Lisbon, Portugal(grant no. PRAXIS XXI/BD/13678/97 to M.C.L.).

REFERENCES

Ahn HJ, Maruo S,Tomura M, et al: A mechanism underlying synergy between IL-12and IFN-g-inducing factor in enhanced production of IFN-g. J Immunol159:2125^2131, 1997

Alexopoulou L, Holt AC, Medzhitov R, Flavell RA: Recognition of double-stranded RNA and activation of NF-kB by Toll-like receptor 3. Nature413:732^738, 2001

Banchereau J, Steinman RM: Dendritic cells and the control of immunity. Nature392:245^252, 1998

Bogdan C: The function of type I interferons in antimicrobial immunity. Curr OpinImmunol 12:419^424, 2000

Cella M, Scheidegger D, Palmer-Lehmann K, Lane P, Lanzavecchia A, Alber G:Ligation of CD40 on dendritic cells triggers production of high levels of inter-leukin-12 and enhances T cell stimulatory capacity: T^T help via APCactivation. J Exp Med 184:747^752, 1996

Cella M, Salio M, SakakibaraY, Langen H, Julkunen I, Lanzavecchia A: Maturation,activation, and protection of dendritic cells induced by double-stranded RNA.J Exp Med 189:821^829, 1999

ChoYS, Kang JW, Cho M, et al: Down-modulation of IL-18 expression by humanpapillomavirus type 16, E6 oncogene via binding to IL-18. FEBS Lett 501:139^145, 2001

Colby C, Duesberg PH: Double-stranded RNA in vaccinia virus infected cells. Nat-ure 222:940, 1969

Companjen AR, Prens E, Mee JB, Groves RW: Expression of IL-18 in humankeratinocytes. J Invest Dermatol 114:598^599, 2000a

Companjen AR,Vooys A, Debets R, Benner R, Prens EP: Human keratinocytes aremajor producers of IL-18: Predominant expression of the unprocessed form.Eur Cytokine Netw 11:383^390, 2000b

De Maeyer E, Maeyer-Guignard J: Type I interferons. Int Rev Immunol 17:53^73, 1998EzepchukYV, Leung DYM, Middleton MH, Bina P, Reiser R, Norris DA: Staphy-

lococcal toxins and protein A di¡erentially induce cytotoxicity and release oftumor necrosis factor-a from human keratinocytes. J Invest Dermatol 107:603^609, 1996

Fujisawa H, Kondo S,Wang B, Shivji GM, Sauder DN:The expression and modula-tion of IFN-a and IFN-b in human keratinocytes. J Interferon Cytok Res 17:721^725, 1997

Grabbe S, Bruvers S, Gallo RL, Knisely TL, Nazareno R, Granstein RD: Tumorantigen presentation by murine epidermal cells. J Immunol 146:3656^3661, 1991

Grabbe S, Bruvers S, Lindgren AM, Hosoi J,Tan KC, Granstein RD:Tumor antigenpresentation by epidermal antigen-presenting cells in the mouse: Modulationby granulocyte-macrophage colony-stimulating factor, tumor necrosis factor a,and ultraviolet radiation. J Leukoc Biol 52:209^217, 1992

Haines DS, Strauss KI, Gillespie DH: Cellular response to double-stranded RNA. JCell Biochem 46:9, 1991

Hilkens CMU, Kalinski P, de Boer M, Kapsenberg ML: Human dendritic cells re-quire exogenous interleukin-12-inducing factors to direct the development ofnaive T-helper cells towards the Th1 phenotype. Blood 90:1920, 1997

Ito T, Inaba M, Inaba K, et al: A CD1aþ /CD11cþ subset of human blood dendriticcells is a direct precursor of Langerhans cells. J Immunol 163:1409^1419, 1999

Janeway CA: The immune system evolved to discriminate infectious nonself fromnoninfectious self. ImmunolToday 13:11^16, 1992

de Jong EC,Vieira PL, Kalinski P, et al: Microbial compounds selectively induce Th1cell-promoting or Th2 cell-promoting dendritic cells in vitro with diverse Thcell-polarizing signals. J Immunol 168:1704^1709, 2002

Kalinski P, Hilkens CMU, Snijders A, Snijdewint FGM, Kapsenberg ML: IL-12-de-¢cient dendritic cells, generated in the presence of prostaglandin E2, promotetype 2 cytokine production in maturing human naive T helper cells. J Immunol159:28^35, 1997

Kalinski P, Schuitemaker JHN, Hilkens CMU, Kapsenberg ML: Prostaglandin E2

induces the ¢nal maturation of IL-12-de¢cient CD1aþCD83þ dendritic cells:The levels of IL-12 are determined during the ¢nal dendritic cell maturationand are resistant to further modulation. J Immunol 161:2804^2809, 1998

Kalinski P, Hilkens CMU,Wierenga EA, Kapsenberg ML: Type 1 and type 2 polar-ized DC: One more message for naive Th cells? ImmunolToday. 20:561^567, 1999a

Kalinski P, Schuitemaker JHN, Hilkens CMU,Wierenga EA, Kapsenberg ML: Finalmaturation of dendritic cells is associated with impaired responsiveness toIFN-g and to bacterial IL-12 inducers: Decreased ability of mature dendriticcells to produce IL-12 during the interaction with Th cells. J Immunol162:3231^3236, 1999b

Kalinski P, Smits HH, Schuitemaker JHN, et al: IL-4 is a mediator of IL-12p70induction by human Th2 cells: Reversal of polarized Th2 phenotype by den-dritic cells. J Immunol 165:1877^1881, 2000

Kalinski P,Vieira PL, Schuitemaker JHN, de Jong EC, Kapsenberg ML: Prostaglan-din E2 is a selective inducer of interleukin-12 p40 (IL-12p40) production andan inhibitor of bioactive IL-12p70 heterodimer. Blood 97:3466^3469, 2001

Koch F, Stanzl U, Jennewein P, et al: High level of IL-12 production by murine den-dritic cells: Upregulation via MHC class II and CD40 molecules and down-regulation by IL-4 and IL-10. J Exp Med 184:741, 1996

K˛ck A, Schwarz T, Kirnbauer R, Urbanski A, Perry P, Ansel JC, Luger TA: Humankeratinocytes are a source for tumor necrosis factor a: Evidence for synthesisand release upon stimulation with endotoxin or ultraviolet light. J Exp Med172:1609^1614, 1990

Kohno K, Kataoka J, Ohtsuki T, et al: IFN-g-inducing factor (IGIF) is a costimula-tory factor on the activation of Th1 but not Th2 cells and exerts its e¡ect in-dependently of IL-12. J Immunol 158:1541^1550, 1997

Kumaraguru U, Rouse BT: The IL-12 response to herpes simplex virus is mainly aparacrine response of reactive in£ammatory cells. J Leukoc Biol 72:564^570,2002

Lebre MC, van Baarsen L, van Capel TMM, Kapsenberg ML, de Jong EC: The roleof toll-like receptors (TLRs) in chemokine and cytokine production by hu-

996 LEBRE ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

man keratinocytes in response to di¡erent pathogen-associated molecular pat-terns (PAMPs). Manuscript, submitted

Luft T, Pang KC, Thomas E, Hertzog P, Hart DN, Trapani J, Cebon J: Type I IFNsenhance the terminal di¡erentiation of dendritic cells. J Immunol 161:1947^1953,1998

Luft T, Luetjens P, Hochrein H, et al: IFN-a enhances CD40 ligand-mediated activa-tion of immature monocyte-derived dendritic cells. Int Immunol 14:367^380,2002

Luger TA, Schwarz T: Evidence for an epidermal cytokine network. J Invest Dermatol95:100S^104S, 1990

Matzinger P: Tolerance, danger, and the extended family. Ann Rev Immunol 12:991^1045, 1994

Mikloska Z, Kesson AM, Penfold ME, Cunningham AL: Herpes simplex virus pro-tein targets for CD4 and CD8 lymphocyte cytotoxicity in cultured epidermalkeratinocytes treated with interferon-g. J Infect Dis 173:7^17, 1996

Mikloska Z, Danis VA, Adams S, Lloyd AR, Adrian DL, Cunningham AL: In vivoproduction of cytokines and beta (C^C) chemokines in human recurrentherpes simplex lesions ^ do herpes simplex virus-infected keratinocytes contri-bute to their production? J Infect Dis. 177:827^838, 1998

Mogensen KE, Pyhala L, Cantell K: Raising antibodies to human leukocyte inter-feron. Acta Pathol Microbiol Scand 83:443^450, 1975

Nahass GT, Penneys NS, Leonardi CL: Interface dermatitis in acute varicella-zostervirus infection: Demonstration of varicella-zoster DNA in keratinocytes by insitu polymerase chain reaction. Br J Dermatol 135:150^151, 2001

Naik SM, Cannon G, Burbach GJ, et al: Human keratinocytes constitutively expressinterleukin-18 and secrete biologically active interleukin-18 after treatmentwith pro-in£ammatory mediators and dinitrochlorobenzene. J Invest Dermatol113:766^772, 1999

Nees M, Geoghegan JM, Hyman T, Frank S, Miller L,Woodworth CD: Papilloma-virus type 16 oncogenes downregulate expression of interferon-responsivegenes and upregulate proliferation-associated and NF-kB-responsive genes incervical keratinocytes. J Virol 75:4283^4296, 2001

O’Garra A: Cytokines induce the development of functionally heterogeneous Thelper cell subsets. Immunity 8:275^283, 1998

Paschen A, Dittmar KEJ, Grenningloh R, et al: Human dendritic cells infected byListeria monocytogenes: Induction of maturation, requirements for phagolysoso-mal escape and antigen presentation capacity. Eur J Immunol 30:3447^3456, 2000

Pastore S, Fanales-Belasio E, Albanesi C, Chinni LM, Giannetti A, Girolomoni G:Granulocyte macrophage colony-stimulating factor is overproduced by kerati-nocytes in atopic dermatitis. Implications for sustained dendritic cell activationin the skin. J Clin Invest 99:3009^3017, 1997

Rescigno M, Granucci F, Ricciardi-Castagnoli P: Molecular events of bacterial-in-duced maturation of dendritic cells. J Clin Immunol 20:161^166, 2000

Robert C, Kupper TS: In£ammatory skin diseases,T cells, and immune surveillance.N Engl J Med 341:1817^1828, 1999

Robinson D, Shibuya K, Mui A, et al: IGIF does not drive Th1 development butsynergizes with IL-12 for interferon-g production and activates IRAK andNF-kB. Immunity 7:571^581, 1997

Sallusto F, Lanzavecchia A: E⁄cient presentation of soluble antigen by culturedhuman dendritic cells is maintained by granulocyte/macrophage colony-stimu-lating factor plus interleukin 4 and downregulated by tumor necrosis factor a.J Exp Med 179:1109^1118, 1994

SarenevaT, Matikainen S, Kurimoto M, Julkunen I: In£uenza Avirus-induced IFN-a/b and IL-18 synergistically enhance IFN-g gene expression in humanT cells.J Immunol 160:6032^6038, 1998

Shortman K, Liu Y-J: Mouse and human dendritic cells. Nature Rev Immunol 2:151^161, 2002

Soumelis V, Reche PA, Kanzler H, et al: Human epithelial cells trigger dendritic cellmediated allergic in£ammation by producing TSLP. Nat Immunol 3:673^680,2002

Sparwasser T, Koch ES,Vabulas RM, Heeg K, Lipford GB, Ellwart JW,Wagner H:Bacterial DNA and immunostimulatory CpG oligonucleotides trigger ma-turation and activation of murine dendritic cells. Eur J Immunol 28:2045^2054,1998

Steinman RM: The dendritic cell system and its role in immunogenicity. Ann RevImmunol 9:271^296, 1991

Trevejo JM, Marino MW, Philpott N, Josien R, Richards EC, Elkon KB, Falck-Pedersen E: TNF-a-dependent maturation of local dendritic cells is criticalfor activating the adaptive immune response to virus infection. Proc Natl AcadSci USA 98:12162^12167, 2001

Uchi H, Terao H, Koga T, Furue M: Cytokines and chemokines in the epidermis.J Dermatol Sci 24:S29^S38, 2000

Verdijk RM, Mutis T, Esendam B, Kamp J, Melief CJM, Brand A, Goulmy E:Polyriboinosinic polyribocytidylic acid (poly (I.C) induces stable maturationof functionally active human dendritic cells. J Immunol 163:57^61, 1999

Verhasselt V, Buelens C,Willems F, De Groote D, Hae¡ner-Cavaillon N, GoldmanM: Bacterial lipopolysaccharide stimulates the production of cytokines and theexpression of costimulatory molecules by human peripheral blood dendriticcells: Evidence for a soluble CD14-dependent pathway. J Immunol 158:2919^2925, 1997

Vieira PL, de Jong EC,Wierenga EA, Kapsenberg ML, Kalinski P: Development ofTh1-inducing capacity in myeloid dendritic cells requires environmental in-struction. J Immunol 164:4507^4512, 2000

Whelan M, Harnett MM, Houston KM, Patel V, Harnett W, Rigley KP: A ¢larialnematode-secreted product signals dendritic cells to acquire a phenotype thatdrives development of Th2 cells. J Immunol 164:6453^6460, 2000

DOUBLE-STRANDED RNA-ACTIVATED HUMAN KERATINOCYTES INDUCE DC1 997VOL. 120, NO. 6 JUNE 2003