Dendritic cells (DCs) not only exhibit the unique capacity to evoke primary immune responses, but may also acquire TLR-triggered cytotoxic activity. We and others have previously shown that TLR7/8- and TLR9-stimulated plasmacytoid DCs (pDCs) isolated from human peripheral blood express the effector molecule TRAIL. The exact mechanisms through which pDCs acquire and elicit their cytotoxic activity are still not clear. We now show that in the absence of costimulators, TRAIL induction on pDCs occurs with agonists to intracellular TLRs only and is accompanied by a phenotypic as well as functional maturation, as evidenced by a comparatively superior MLR stimulatory capacity. pDCs acquired TRAIL in an IFN-α/β–dependent fashion and, notably, TRAIL expression on pDCs could be induced by IFN-α stimulation alone. At a functional level, both TLR7/8- (imiquimod [IMQ]) and TLR9-stimulated (CpG2216) pDCs lysed Jurkat T cells in a TRAIL- and cell contact-dependent fashion. More importantly, IFN-α–activated pDCs acquired similar cytotoxic properties, independent of TLR stimulation and maturation. Both IMQ- and IFN-α–activated pDCs could also lyse certain melanoma cell lines in a TRAIL-dependent fashion. Interestingly, suboptimal doses of IMQ and IFN-α exhibited synergistic action, leading to optimal TRAIL expression and melanoma cell lysis by pDCs. Our data imply that tumor immunity in patients receiving adjuvant IMQ and/or IFN-α may involve the active participation of cytotoxic pDCs.

Dendritic cells (DCs) are characterized by their unique ability to evoke primary immune responses by presenting Ags to naive T cells after trafficking to the draining lymph nodes (1). Among the different Ag recognition strategies employed by DCs, TLRs are of particular importance as they detect unique molecular patterns conserved throughout entire classes of pathogens (2). TLRs 1, 2, 4, 5, and 6 are located at the extracellular surface membrane and specialize in the recognition of bacterial products. TLRs 3, 7, 8, and 9, in contrast, are situated intracellularly (i.e., in endosomal membranes), where they detect viral and bacterial nucleic acids that have gained access to the intracellular compartments (2). The TLR-mediated activation of DCs triggers an antimicrobial and inflammatory response via the downstream activation of NF-κB and other transcription factors, culminating in the production of proinflammatory cytokines and type I IFNs and the upregulation of costimulatory molecules (1, 3). Together, these factors initiate and direct the development of an adaptive immune response.

Human peripheral blood contains five nonoverlapping lineageHLA-DR+ subsets: hematopoietic CD34+ precursors; CD11c+ myeloid DCs (mDCs), which can be further subdivided according to their expression of blood DC Ag (BDCA)-1 (CD1c), BDCA-3 (CD141), and CD16; and, finally, BDCA-2+ (CD303) plasmacytoid DCs (pDCs) (4). PDCs exhibit TLR7, TLR9, and, perhaps, low levels of TLR1 and TLR6 (57). Their antiviral function is probably linked to their high expression of TLR7 and TLR9 (5, 6). After sensing viral nucleic acids, they produce large amounts of type I IFNs due to their constitutional expression of the transcription factor IFN regulatory factor-7 and so are essential in the induction of a robust antiviral immune response (3).

Recent evidence exists that TLR activation can endow pDCs with unique cytotoxic properties, challenging our current concept of DC biology. HIV, human T cell leukemia virus-1, and influenza virus, all natural ligands of TLR7, trigger a pDC innate immune response that not only involves the massive production of IFN-α, but also the upregulation of functionally active TRAIL (TRAIL/Apo-2L/TNFSF10) (810). TRAIL is also expressed by HIV-infected pDCs (8, 11, 12), and, notably, TRAIL+ pDCs may kill CD4+ T cells isolated from HIV-infected patients (12) and thus perhaps contribute to T cell depletion during HIV viremia. In addition to their role in viral immunity, several studies propose a direct role for cytotoxic pDCs in antitumor immunity. The synthetic TLR7/8 agonist imiquimod (IMQ) is a highly effective topical treatment for epithelial skin tumors such as basal cell carcinomas, viral acanthomas, and even melanomas (13). Regressing cancers were surrounded and heavily infiltrated by inflammatory-type DCs (14, 15), some of which expressed lytic molecules (15). In vitro studies further showed that IMQ induces the expression of perforin and granzyme B on mDCs and of TRAIL on pDCs (8, 15) and that TRAIL+ pDCs and granzyme B+ perforin+ mDCs can effectively lyse appropriate tumor targets (9, 15).

In this study, we attempted to gain a better understanding of the events governing the acquisition of cytotoxic molecules by pDCs and of the molecular players involved in IMQ-induced pDC antitumor activity. We show in this study that IMQ endows pDCs with the ability to lyse tumor cells, a phenomenon that is strictly TRAIL- and IFN-α–dependent. It is of note that IFN-α–activated pDCs, even in the absence of IMQ, can kill tumor cells—in particular also melanoma cells—in a TRAIL-dependent fashion. This result suggests that the therapeutic effect of adjuvant IMQ and/or IFN-α in antitumor immunity may be, at least partly, mediated by killer DCs.

To isolate human peripheral blood pDCs, PBMCs were prepared by Ficoll-Paque density gradient centrifugation (Ficoll-PaquePLUS; GE Healthcare) from buffy coats purchased from the Austrian Red Cross Center (Vienna, Austria). T cells, B cells, NK cells, hematopoietic progenitor cells, monocytes, platelets, and erythrocytes were depleted by subsequent anti-CD3 (UCHT1), -CD11b (Bear1), -CD16 (3G8), -CD19 (J4.119), -CD34 (581), -CD41 (SZ22), -CD56 (C218), and -CD235a (11E4B-7-6; all from Beckman Coulter, used at 4 μg/ml each) immunolabeling and anti-mouse IgG1 immunomagnetic separation (MACS; Miltenyi Biotec). pDCs were selected from the remaining cell fraction using anti-CD304 (BDCA-4) microbeads (MACS Dendritic Cell Isolation Kit; Miltenyi Biotec). The purity of the isolated pDC population was generally 95–99% and contained <1% of other leukocyte populations tested for individually. CD3+ T cells from the same donor that served as negative controls in the cytotoxicity assays were prepared by anti-CD3 immunolabeling and anti-mouse IgG1 immunomagnetic selection. CD3+ T responder cells used in MLR and T cell activation experiments were isolated from PBMCs by negative selection using anti-CD11b, -CD11c (BU15), -CD16, -CD19, -CD34, -CD41, -CD56, -CD235a (Beckman Coulter), and -CD123 (9F5; BD Pharmingen; 4 μg/ml each) immunolabeling and anti-mouse IgG1 immunomagnetic depletion. The purity of CD3+ T cells was >95% as assessed by FACS analysis. Cell viability determined by trypan blue staining was >99% after isolation.

Isolated pDCs were resuspended in RPMI 1640 medium supplemented with 10% (v/v) heat-inactivated FCS, penicillin (100 IU/ml), and streptomycin (100 μg/ml; all from Invitrogen) at 5–10 × 105 cells/ml in 96-well plates and cultured overnight at 37°C in 5% CO2. The following TLR agonists were used for stimulation: Pam3CSK4 (TLR2/1, 0.1–1 μg/ml), FSL-1 (TLR2/6, 0.1–1 μg/ml), polyinosinic-polycytidylic acid (TLR3, 5 μg/ml), ultrapure Escherichia coli LPS (TLR4, 1 μg/ml), recombinant FLA-ST (flagellin; TLR5, 0.1 μg/ml), IMQ (TLR7/8, 5 μg/ml), CL-075 (TLR7/8, 2.5 μg/ml), ssPolyU/LyoVec (TLR8, 2.5 μg/ml) and CpG2216 (TLR9, 2.5 μM; all purchased from InvivoGen). Cell viability before and after culture was checked by trypan blue staining and was generally >90%. The effect of other maturation stimuli on lytic molecule induction was tested by stimulating freshly isolated pDCs with IL-3 (300 IU/ml; PeproTech), TNF-α (50 IU/ml; Strathmann), and CD40L (1 μg/ml; Enzo Life Sciences). To determine the role of type I IFNs in the induction of cytotoxic molecules, pDCs were preincubated with 10 μg/ml each neutralizing anti–IFN-α (MMHA-1) and anti–IFN-β (MMHB-12; PBL Biomedical Laboratories) Abs or an IgG1 isotype (MOPC-21; Sigma-Aldrich) 30 min before addition of TLR agonists, or directly stimulated with 50–50,000 IU/ml human rIFN-α2a (PBL Biomedical Laboratories) and then subjected to FACS analysis or used in cytotoxicity assays. pDCs that were used to lyse melanoma cells were activated with Roferon-A (rIFN-α2a; Roche; 25,000 IU/ml; low dose: 250 IU/ml) or IMQ (5 μg/ml; low dose: 0.05 μg/ml).

For analysis of TRAIL and TRAIL-R expression, T cells were stimulated with plate-bound anti-CD3 (5 μg/ml) and soluble anti-CD28 (2 μg/ml, L293; BD Biosciences) for up to 5 d and then analyzed by flow cytometry. Fas ligand (FasL) expression on CD3+ T cells was determined by flow cytometry using PBMCs that had been stimulated with PMA (25 ng/ml) and ionomycin (2 μg/ml; both from Sigma-Aldrich) for 4 h.

Flow cytometric analysis of blood-derived pDCs before and after culture with TLR agonists or cytokines was performed on an FACSCalibur (BD Biosciences). The purity of isolated pDC populations was determined using lineage marker combinations of anti–HLA-DR (PerCP, L234; BD Biosciences) and anti-CD303/BDCA-2 (allophycocyanin, AC144; Miltenyi Biotec). The absence of other leukocyte subpopulations in freshly isolated DCs was checked by quadruple stainings with Abs against CD14 (FITC; MϕP9), CD56 (PE; mY31), CD19 (PerCP; SJ25C1), and CD3 (allophycocyanin, SK7; all from BD Biosciences). TRAIL expression was visualized using the PE-labeled Ab clones 75402 (R&D Systems) and RIK-2 (BioLegend), both yielding similar results. All FACS plots shown in this paper were generated using clone RIK-2. FasL expression was evaluated using a purified anti-CD178 mAb (NOK-1; BD Pharmingen) followed by Alexa Fluor 488 goat anti-mouse IgG (H+L) (Molecular Probes, Invitrogen) staining. Positive control stainings were performed using PMA/ionomycin-stimulated PBMCs and gating on CD3+ T cells. Intracellular cytotoxic molecules were visualized with anti-granzyme B PE (CLB-GB11; PeliCluster Sanquin), anti-granulysin PE (DH-2; BioLegend), anti-perforin FITC (δG9; BD Biosciences), and anti-lysozyme FITC (LZ-2; An Der Grub). Intracellular stainings were performed using a cell permeabilization kit (Fix&Perm; An Der Grub) after extracellular staining of pDCs with lineage markers. Positive control stainings for intracellular lytic molecules were performed using freshly isolated PBMCs and gating on NK cells (CD56+CD3) or monocytes (scatter characteristics). Costimulatory molecules were stained using anti-CD80 FITC (MAB104; Beckman Coulter), -CD40 FITC (5C3), -CD83 PE (HB15e), and -CD86 PE (2331; all from BD Biosciences). TRAIL-R expression on pDCs, T cells, and tumor target cells was determined using anti–TRAIL-R1 PE (69036), –TRAIL-R2 PE (71908), –TRAIL-R3 PE (90906), and –TRAIL-R4 PE (104918; all from R&D Systems). The activation status of T cells was checked using anti-CD69 FITC (FN50; BD Pharmingen). Type I IFNR expression was examined by anti-IFNAR2 PE (MMHAR-2; PBL Biomedical Laboratories). Corresponding isotype-matched controls were used for all stainings. After incubation with the respective Abs for 15 min at 4°C, cells were washed twice and then subjected to flow cytometric analysis. FACS plots depict mean fluorescence intensity (MFI) values of Ab stainings after subtraction of the MFI of the respective isotype.

The immunostimulatory capacity of killer molecule-expressing pDCs was assessed in allogeneic MLRs. Purified pDCs that had been stimulated with IMQ or left unstimulated overnight were incubated with 105 purified allogeneic CD3+ T cells in round-bottom 96-well plates. The effects of type I IFNs and TRAIL on pDC-induced T cell proliferation were investigated by adding 5 μg/ml each azide-free neutralizing anti-TRAIL (75411; R&D Systems), anti–IFN-α, and anti–IFN-β, or 5 μg/ml IgG1 isotype to the plate. T cell proliferation was measured by the uptake of [3H]thymidine (1 μCi/well; 6.7 Ci/mM [247.9 GBq/mM]; PerkinElmer), which was added to the culture on day 5. After 16 h of incubation, cells were harvested onto glass fiber filter paper with an automated harvester (Tomtec Harvester 96; Tomtec), and [3H]thymidine incorporation was measured by liquid scintillation spectroscopy (1205 Betaplate; Wallac). The T cell response is given as mean cpm ± SD for duplicate wells.

The TRAIL-sensitive human leukemia T cell line Jurkat (obtained from American Type Culture Collection) and melanoma cell lines SKMel2 (American Type Culture Collection) and WM793 (Coriell Institute for Medical Research) were maintained in RPMI 1640 medium supplemented with 10% (v/v) heat-inactivated FCS, penicillin (100 IU/ml), and streptomycin (100 μg/ml). WM793 cells were grown in collagen-coated flasks (rat tail collagen type I; BD Biosciences), and both melanoma lines were used when cells were 70–80% confluent.

The ability of pDCs to kill tumor cells was assessed in a classic Europium-TDA release assay (DELFIA; PerkinElmer) according to the manufacturer’s protocol and as previously described (16). Briefly, target cells were labeled with the fluorescence-enhancing ligand BATDA, which is released into the supernatant after cytolysis. The supernatant is harvested and incubated with Europium solution to form a stable fluorescent chelate. Data were obtained using the 1234 DELFIA Luminometer (Wallac) and are expressed as the percentage of specific lysis calculated by the following formula: specific lysis (%) = (experimental release − spontaneous release)/(maximum release − spontaneous release) × 100. We used the cell lines Jurkat, SKMel2, and WM793 as targets, as well as unstimulated CD3+ T cells isolated from the same donor as the effector pDC population. Purified pDCs that had been stimulated with TLR agonists or IFN-α overnight or had been left unstimulated were incubated with 2–5 × 103 target cells as duplicates in 96-well plates at E:T ratios ranging from 20:1 to 2.5:1. Target cell lysis was measured after 2 h and, in the case of melanoma cells, after 4–6 h of incubation with effector pDCs. Inhibition experiments were performed by preincubating pDCs with neutralizing anti–IFN-α (10 μg/ml), anti–IFN-β (10 μg/ml), and anti-TRAIL (10 μg/ml) Abs or an isotype Ab for 30 min prior to addition of target cells. The maximum possible level of TRAIL-induced lysis of melanoma cells was evaluated after adding 100 ng/ml super killer TRAIL (skTRAIL; soluble human recombinant; Enzo Life Sciences) to melanoma target cells. To evaluate the contribution of soluble TRAIL (sTRAIL) present in the supernatant of pDC cultures to pDC killer activity, we performed cytotoxicity assays using supernatants of IMQ-activated pDCs (cultured at 5 × 105 cells/ml) that were added to 2 × 103 Jurkat target cells (v:v/1:1). The baseline cytotoxicity of supernatants from unstimulated pDCs was subtracted from that of TLR-activated pDCs to correct for influences of other secreted molecules and effects of factor consumption.

The concentration of sTRAIL in the supernatants of TLR-stimulated pDCs cultured at 5 × 105 cells/ml was determined by ELISA (eBioscience) according to the manufacturer’s instructions. The concentrations of IFN-α, TNF-α, and IFN-γ in the supernatants of TLR-stimulated pDCs and PMA/ionomycin-activated PBMCs, both cultured at 1 × 106 cells/ml overnight, were determined by FlowCytomix (eBioscience). Data analysis was performed using FlowCytomix Pro 2.4 software (eBioscience).

Data are expressed as mean ± SD or SEM. Statistical differences were calculated using the two-tailed Student t test, and p values <0.05 were considered statistically significant.

In a first series of experiments, we comparatively assessed the expression of different cytotoxic molecules on purified pDCs (Fig. 1A) that had been either left untreated or stimulated with agonists to TLR1–9. Fig. 1B shows representative histograms of lytic molecule expression and Fig. 1C the quantitative analysis of the overall percentage of pDCs expressing lytic molecules after stimulation with TLR agonists. In agreement with previous observations, we found that TLR7/8- and TLR9-stimulated pDCs, as opposed to their nonstimulated counterparts, express high levels of TRAIL (Fig. 1B). Kinetic analysis showed that TRAIL is detectable on the cell surface as early as 3 h after stimulation, peaks after 12 h, and is maintained for >24 h after stimulation (Fig. 1D). As pDCs do not express TLR8 (57), we did not find TRAIL expression using the selective TLR8 agonist ssPolyU as stimulus (Fig. 1C). Even though pDCs also express transcripts for the extracellular TLRs 1 and 6 that are involved in the recognition of bacterial products (6, 7), none of the extracellular TLR agonists used, including the respective TLR1 and TLR6 agonists Pam3CSK4 and FSL-1, caused lytic molecule expression (Fig. 1C). Notably, these agonists were functional in stimulating mDC maturation (data not shown).

FIGURE 1.

Expression of cytotoxic molecules on human peripheral blood-derived pDCs. A, The purity of pDCs after isolation and purification was assessed by flow cytometry using anti–BDCA-2 and anti–HLA-DR stainings. BF, The expression of different cytotoxic molecules on freshly isolated and purified pDCs before and after stimulation with TLR agonists or cytokines was assessed by flow cytometry and by gating on BDCA-2+HLA-DR+ cells. B and C, Freshly isolated pDCs were incubated with the TLR agonists Pam3CSK4 (TLR2/1), FSL-1 (TLR2/6), polyinosinic-polycytidylic acid (Poly I:C, TLR3), LPS (TLR4), flagellin (TLR5), IMQ (TLR7/8), CL-075 (TLR7/8), ssPolyU (TLR8), and CpG2216 (TLR9). Cells cultured with IL-3 (300 IU/ml) or medium alone served as controls. After overnight culture, the expression of cytotoxic molecules was assessed using extracellular anti-TRAIL and intracellular anti-granzyme B stainings. Representative histograms of membrane TRAIL and intracellular granzyme B expression are shown in B. Data of quantitative analysis (mean ± SD) of pDCs expressing lytic molecules are shown in C (n = 4). D, The kinetics of TRAIL expression on IMQ-stimulated (5 μg/ml) pDCs were assessed over a period of 48 h using anti-TRAIL stainings. One representative example is shown (n = 3). E, The expression of other cytotoxic molecules by pDCs was assessed on freshly purified pDCs and after overnight stimulation with IMQ (5 μg/ml) using Abs against intracellular granulysin, perforin, and lysozyme as well as extracellular FasL. Stainings using freshly isolated PBMCs and gating on NK cells (CD56+CD3) or monocytes (scatter characteristics) and PMA/ionomycin-stimulated PBMCs and gating on T cells (CD3+) served as positive controls. Representative histograms are shown (n = 3). F, The expression of surface TRAIL on pDCs was assessed after stimulation with IMQ (5 μg/ml), IL-3 (300 IU/ml), TNF-α (50 IU/ml), and CD40L (1 μg/ml) using anti-TRAIL stainings. Data represent mean percentages of pDCs expressing lytic molecules ± SD (n = 3).

FIGURE 1.

Expression of cytotoxic molecules on human peripheral blood-derived pDCs. A, The purity of pDCs after isolation and purification was assessed by flow cytometry using anti–BDCA-2 and anti–HLA-DR stainings. BF, The expression of different cytotoxic molecules on freshly isolated and purified pDCs before and after stimulation with TLR agonists or cytokines was assessed by flow cytometry and by gating on BDCA-2+HLA-DR+ cells. B and C, Freshly isolated pDCs were incubated with the TLR agonists Pam3CSK4 (TLR2/1), FSL-1 (TLR2/6), polyinosinic-polycytidylic acid (Poly I:C, TLR3), LPS (TLR4), flagellin (TLR5), IMQ (TLR7/8), CL-075 (TLR7/8), ssPolyU (TLR8), and CpG2216 (TLR9). Cells cultured with IL-3 (300 IU/ml) or medium alone served as controls. After overnight culture, the expression of cytotoxic molecules was assessed using extracellular anti-TRAIL and intracellular anti-granzyme B stainings. Representative histograms of membrane TRAIL and intracellular granzyme B expression are shown in B. Data of quantitative analysis (mean ± SD) of pDCs expressing lytic molecules are shown in C (n = 4). D, The kinetics of TRAIL expression on IMQ-stimulated (5 μg/ml) pDCs were assessed over a period of 48 h using anti-TRAIL stainings. One representative example is shown (n = 3). E, The expression of other cytotoxic molecules by pDCs was assessed on freshly purified pDCs and after overnight stimulation with IMQ (5 μg/ml) using Abs against intracellular granulysin, perforin, and lysozyme as well as extracellular FasL. Stainings using freshly isolated PBMCs and gating on NK cells (CD56+CD3) or monocytes (scatter characteristics) and PMA/ionomycin-stimulated PBMCs and gating on T cells (CD3+) served as positive controls. Representative histograms are shown (n = 3). F, The expression of surface TRAIL on pDCs was assessed after stimulation with IMQ (5 μg/ml), IL-3 (300 IU/ml), TNF-α (50 IU/ml), and CD40L (1 μg/ml) using anti-TRAIL stainings. Data represent mean percentages of pDCs expressing lytic molecules ± SD (n = 3).

Close modal

We next asked whether TLR7–9 agonists induce also lytic molecules other than TRAIL in freshly isolated pDCs. Positive controls were generated using NK cells (perforin, granulysin), monocytes (lysozyme), and T cells (FasL). Data obtained show that unstimulated, and, to a lesser extent, TLR7/8- and TLR9-stimulated pDCs express granzyme B (Fig. 1B, 1C). Granulysin, perforin, or FasL were consistently absent (Fig. 1E). In contrast to what has been reported previously (17), we were unable to detect lysozyme expression (Fig. 1E), even in the CD2high pDC subset (data not shown).

DC activation and maturation, and therefore possibly also TRAIL expression, can be induced by a multitude of different stimuli. To investigate whether TRAIL induction is exclusively linked to TLR activation, we treated freshly isolated pDCs with different cytokines and then evaluated TRAIL expression. Interestingly, IL-3 not only caused DC maturation, measured by the upregulation of the costimulatory molecule CD83 (data not shown), but also induced TRAIL on up to 30% of cells after 24 h (Fig. 1F). The levels of IL-3–induced TRAIL, however, were much lower when compared with those measured after IMQ stimulation. Notably, pDCs treated with TNF-α and CD40L exhibited signs of maturation (upregulation of CD83, data not shown) but did not express TRAIL (Fig. 1F). These findings imply that TRAIL induction is not a general attribute of DC activation or maturation, but is confined to distinct pDC stimuli. Taken together, our results demonstrate that, with the exception of intracellular granzyme B, human blood-derived pDCs do not react with Abs against the major lytic molecules identified so far. They can, however, be stimulated to express TRAIL, the induction of which is exclusively linked to the occupancy of cytoplasmic rather than extracellular membrane-bound TLRs.

The capacity of DCs to induce a productive response in naive T cells is critically linked to the expression of certain costimulatory molecules (1). We asked whether TLR-mediated killer molecule expression on pDCs interferes with them acquiring such moieties. Results obtained show that, with the exception of low-level CD86 expression (∼40%), freshly isolated pDCs are essentially devoid of activating costimulatory molecules (Fig. 2A). After TLR7/8 (IMQ, CL-075) and TLR9 (CpG2216) stimulation, the expression of killer molecules was accompanied by a marked upregulation of CD40, CD80, CD83, and CD86 as well as HLA-DR in the vast majority of pDCs, as indicated by the increase in MFI (Fig. 2A) and the overall percentage of positive cells (Fig. 2B). We next examined the capacity of cytotoxic molecule-expressing pDCs to induce T cell proliferation in an allogeneic MLR. Overall, pDCs exhibited a low stimulatory capacity compared with mDCs (M.L. Kalb, unpublished observations). Notably, IMQ-stimulated pDCs were more effective stimulators compared with the nonactivated pDC population, despite their cytotoxic molecule expression (Fig. 2C). These results were somewhat surprising, as one may have expected that IFN-α produced by pDCs would exert an antiproliferative effect and that killer molecule-expressing pDCs would lyse activated T cells (12). Indeed, we found that anti–CD3-/CD28-activated, but not freshly isolated CD3+ T cells express the proapoptotic receptor TRAIL-R1 (DR4/TNFRSF10A) (Fig. 2D). Fig. 2C shows, however, that neutralizing Abs against TRAIL and/or IFN-α do not alter the extent of pDC-induced T cell proliferation. These data demonstrate that TLR7/8 agonists, in particular IMQ, not only induce lytic molecules in pDCs but also enhance the immunostimulatory capacity of these cells.

FIGURE 2.

TRAIL+ pDCs express high levels of costimulatory molecules and exhibit allostimulatory capacity. A and B, The expression of costimulatory molecules on freshly purified, unstimulated pDCs and TRAIL+ pDCs generated by overnight TLR7/8 (IMQ, CL-075) or TLR9 (CpG2216) stimulation was analyzed by flow cytometry. Cells were gated on BDCA-2 and HLA-DR. Representative histograms (A) depict the baseline expression of costimulatory molecules on freshly isolated pDCs and their upregulation after overnight stimulation with IMQ. MFI values of IMQ-activated cells are shown. Quantitative analysis (B) shows mean percentages of TRAIL-expressing cells before and after stimulation with TLR7/8 and TLR9 agonists (n = 3). C, MLRs were performed with titrated numbers of unstimulated or IMQ-activated pDCs incubated with 105 allogeneic T cells per well. The effects of TRAIL and type I IFNs were investigated using neutralizing anti-TRAIL, anti–IFN-α, anti–IFN-β (5 μg/ml each) or isotype-matched control Abs that were preincubated with pDCs for 30 min prior to addition of responder cells. T cell proliferation on day 6 was measured by [3H]thymidine incorporation. Data are expressed as mean cpm of duplicate wells ± SD. One representative example of three independent experiments giving similar results is shown. D, Flow cytometric analysis of CD69, TRAIL, and TRAIL-R1-R4 expression on freshly isolated and anti–CD3-/CD28-activated CD3+ T cells. Cells were gated on CD3 and histograms shown are representative of three different donors analyzed.

FIGURE 2.

TRAIL+ pDCs express high levels of costimulatory molecules and exhibit allostimulatory capacity. A and B, The expression of costimulatory molecules on freshly purified, unstimulated pDCs and TRAIL+ pDCs generated by overnight TLR7/8 (IMQ, CL-075) or TLR9 (CpG2216) stimulation was analyzed by flow cytometry. Cells were gated on BDCA-2 and HLA-DR. Representative histograms (A) depict the baseline expression of costimulatory molecules on freshly isolated pDCs and their upregulation after overnight stimulation with IMQ. MFI values of IMQ-activated cells are shown. Quantitative analysis (B) shows mean percentages of TRAIL-expressing cells before and after stimulation with TLR7/8 and TLR9 agonists (n = 3). C, MLRs were performed with titrated numbers of unstimulated or IMQ-activated pDCs incubated with 105 allogeneic T cells per well. The effects of TRAIL and type I IFNs were investigated using neutralizing anti-TRAIL, anti–IFN-α, anti–IFN-β (5 μg/ml each) or isotype-matched control Abs that were preincubated with pDCs for 30 min prior to addition of responder cells. T cell proliferation on day 6 was measured by [3H]thymidine incorporation. Data are expressed as mean cpm of duplicate wells ± SD. One representative example of three independent experiments giving similar results is shown. D, Flow cytometric analysis of CD69, TRAIL, and TRAIL-R1-R4 expression on freshly isolated and anti–CD3-/CD28-activated CD3+ T cells. Cells were gated on CD3 and histograms shown are representative of three different donors analyzed.

Close modal

pDCs are known to produce vast amounts of type I IFNs upon encounter of natural (viruses) and synthetic TLR7 ligands (3). TRAIL is an IFN-stimulated gene (18), and the same TLR7 ligands that induce IFN-α production can also induce TRAIL expression on pDCs (8, 9). In agreement with these observations, we found that TRAIL expression on pDCs induced by TLR7/8 (IMQ, CL-075) and also TLR9 (CpG2216) agonists was paralleled by an increase of IFN-α and TNF-α, but not IFN-γ in the supernatant (Fig. 3A). In contrast to pDCs, we found that PMA/ionomycin-treated PBMCs produce large amounts of IFN-γ as well as TNF-α, but not IFN-α. When we investigated the role of type I IFNs in the development of a cytotoxic phenotype, we found that: 1) the IFN-α/β receptor expressed on nonstimulated pDCs (Fig. 3B); and 2) pDC stimulation with TLR7/8 and TLR9 agonists in the presence of anti–IFN-α/β Abs prevents their TRAIL expression (Fig. 3C). To test whether TLR stimulation is required or, alternatively, whether IFN-α alone may be sufficient for TRAIL induction, we stimulated pDCs with different amounts of IFN-α overnight. Fig. 3D shows that IFN-α stimulation alone induces TRAIL expression on pDCs (Fig. 3D, upper panel), albeit at a lesser extent than what is seen with IMQ (Fig. 1B). It should also be noted that IFN-α, in contrast to IMQ stimulation, does not lead to an upregulation of costimulatory molecules, not even at high doses (Fig. 3E). These findings suggest that TLR agonists activate signaling pathways in addition to those engaged by IFN-α. To investigate this hypothesis, we added minimal doses of IMQ (0.05 μg/ml) that alone did not induce relevant levels of TRAIL to pDCs cultured with IFN-α and, indeed, found that TRAIL expression was significantly enhanced under such conditions (Fig. 3D, lower panel). We further observed that in the presence of suboptimal IFN-α concentrations (250 IU/ml), agonists to TLR2/1 (Pam3CSK4) and TLR2/6 (FSL-1) that were unable to stimulate TRAIL expression when used alone now induced low-level TRAIL expression (Supplemental Fig. 1).

FIGURE 3.

PDCs upregulate TRAIL in an IFN-α/β–dependent fashion. A, Measurement of IFN-α, TNF-α, and IFN-γ in the supernatants of TLR7/8- (IMQ, CL-075) and TLR9-activated (CpG2216) pDCs and PBMCs activated with PMA/ionomycin. Data represent mean cytokine levels ± SEM (n = 2–5). B, Flow cytometric analysis of IFN-α/β receptor expression (open histogram) on pDCs compared with an isotype Ab (filled histogram) using fresh PBMCs and gating on HLA-DR+BDCA-2+ cells. C, Freshly isolated pDCs were preincubated with neutralizing anti–IFN-α, anti–IFN-β (5 μg/ml each), or isotype-matched control Abs for 30 min prior to TLR stimulation. After overnight culture, the level of surface TRAIL was assessed by flow cytometry. Data represent mean percentages of TRAIL+ cells ± SD (n = 3). D, Histograms depict MFI of surface TRAIL expression on pDCs after overnight stimulation with different concentrations of IFN-α (0–50,000 IU/ml) alone or in combination with low-dose IMQ and are representative of three independent experiments performed. E, Flow cytometric analysis of costimulatory molecule expression on TRAIL+ pDCs generated by IFN-α activation (25,000 IU/ml). One representative staining series is shown (n = 4). Numbers indicate the percentage of positive cells. CE, Cells were gated on BDCA-2 and HLA-DR.

FIGURE 3.

PDCs upregulate TRAIL in an IFN-α/β–dependent fashion. A, Measurement of IFN-α, TNF-α, and IFN-γ in the supernatants of TLR7/8- (IMQ, CL-075) and TLR9-activated (CpG2216) pDCs and PBMCs activated with PMA/ionomycin. Data represent mean cytokine levels ± SEM (n = 2–5). B, Flow cytometric analysis of IFN-α/β receptor expression (open histogram) on pDCs compared with an isotype Ab (filled histogram) using fresh PBMCs and gating on HLA-DR+BDCA-2+ cells. C, Freshly isolated pDCs were preincubated with neutralizing anti–IFN-α, anti–IFN-β (5 μg/ml each), or isotype-matched control Abs for 30 min prior to TLR stimulation. After overnight culture, the level of surface TRAIL was assessed by flow cytometry. Data represent mean percentages of TRAIL+ cells ± SD (n = 3). D, Histograms depict MFI of surface TRAIL expression on pDCs after overnight stimulation with different concentrations of IFN-α (0–50,000 IU/ml) alone or in combination with low-dose IMQ and are representative of three independent experiments performed. E, Flow cytometric analysis of costimulatory molecule expression on TRAIL+ pDCs generated by IFN-α activation (25,000 IU/ml). One representative staining series is shown (n = 4). Numbers indicate the percentage of positive cells. CE, Cells were gated on BDCA-2 and HLA-DR.

Close modal

To investigate whether TRAIL is functionally active on pDCs, we performed 2-h Europium-TDA release cytotoxicity assays. TDA release from target cells occurs early and rapidly, allowing for the detection of cell death already after shorter periods of culture compared with other commonly used cell death assays (Supplemental Fig. 2). pDCs were incubated with the leukemic T cell line Jurkat, which expresses the proapoptotic receptor TRAIL-R2 (DR5/TNFRSF10B) (Fig. 4A) and is highly susceptible to TRAIL-mediated killing. We found that pDCs that had been stimulated overnight with IMQ, but not untreated pDCs, kill Jurkat cells in a TRAIL-dependent fashion, as demonstrated by the lack of killing in the presence of neutralizing anti-TRAIL Abs (Fig. 4B). In keeping with earlier observations (Fig. 2D), TRAIL-expressing pDCs were unable to lyse unstimulated CD3+ T cells that had been isolated from the same donor as the pDC effector population (Supplemental Fig. 3). Having shown that TRAIL is also induced through TLR9 (Fig. 1B), we performed cytotoxicity assays using pDCs that had been activated with CpG2216 overnight. Cytotoxic activity of TLR9-activated pDCs was less pronounced than that of their TLR7/8-activated counterparts isolated from the same donor (Fig. 4C). This could, but must not necessarily, be due to a lower expression of TRAIL (Fig. 1B, upper panel).

FIGURE 4.

TLR- and IFN-α–activated pDCs kill Jurkat cells in a TRAIL-dependent fashion. A, Expression of TRAIL-R1 and -R2 on Jurkat T cells was assessed by flow cytometry. BE, Cytotoxic activity of IMQ- (5 μg/ml), CpG2216- (2.5 μM), and IFN-α–activated (25,000 IU/ml) pDCs or pDCs that were left unstimulated overnight against Jurkat T cells was determined after 2 h. In BD, activated pDCs were preincubated with neutralizing anti-TRAIL, anti–IFN-α, anti–IFN-β (10 μg/ml each), or isotype-matched control Abs for 30 min prior to addition of target cells. In E, freshly isolated pDCs were preincubated with neutralizing anti–IFN-α/β or isotype-matched control Abs for 30 min prior to induction of TRAIL with IMQ. Data in B, D, and E represent mean of duplicates of specific lysis ± SD, and one representative example of three independent experiments yielding similar results is shown. Data in C show mean of specific lysis ± SD of results from two independent experiments.

FIGURE 4.

TLR- and IFN-α–activated pDCs kill Jurkat cells in a TRAIL-dependent fashion. A, Expression of TRAIL-R1 and -R2 on Jurkat T cells was assessed by flow cytometry. BE, Cytotoxic activity of IMQ- (5 μg/ml), CpG2216- (2.5 μM), and IFN-α–activated (25,000 IU/ml) pDCs or pDCs that were left unstimulated overnight against Jurkat T cells was determined after 2 h. In BD, activated pDCs were preincubated with neutralizing anti-TRAIL, anti–IFN-α, anti–IFN-β (10 μg/ml each), or isotype-matched control Abs for 30 min prior to addition of target cells. In E, freshly isolated pDCs were preincubated with neutralizing anti–IFN-α/β or isotype-matched control Abs for 30 min prior to induction of TRAIL with IMQ. Data in B, D, and E represent mean of duplicates of specific lysis ± SD, and one representative example of three independent experiments yielding similar results is shown. Data in C show mean of specific lysis ± SD of results from two independent experiments.

Close modal

Because IFN-α itself (i.e., in the absence of TLR ligands) can induce TRAIL expression on pDCs (Fig. 3D, upper panel), we performed additional cytotoxicity assays using IFN-α–activated pDCs to test their lytic capacity. Notably, IFN-α–activated pDCs were very efficient killers (Fig. 4D). Despite their often only marginal expression of TRAIL (Fig. 3D, upper panel), IFN-α–activated pDCs lysed Jurkat targets in a strictly TRAIL-dependent fashion (Fig. 4D). pDC cytotoxicity is apparently independent of DC maturation, as IFN-α–activated TRAIL+ pDCs had not upregulated costimulatory molecules to the same extent as seen after IMQ stimulation (Figs. 3E and 2A, respectively). As expected, pDCs that had been incubated with neutralizing anti–IFN-α/β Abs prior to IMQ stimulation, and therefore could not upregulate TRAIL, were also not cytotoxic to Jurkat cells (Fig. 4E). In contrast, the addition of neutralizing anti–IFN-α/β Abs during the cytotoxicity assay did not affect tumor cell lysis (Fig. 4B), indicating that IFN-α produced by pDCs that already express TRAIL is dispensable for their cytotoxic function. Taken together, our data show that both IMQ- and IFN-α–activated pDCs kill Jurkat cells in a TRAIL-dependent fashion and may therefore be actively involved in clearing TRAIL-sensitive tumor targets.

Once distributed to the cell membrane, the type II membrane protein TRAIL may be subject to cleavage by metalloproteases (19). Because evidence exists that TRAIL is functional in both its membrane-bound and soluble (sTRAIL) form, we now asked how TRAIL mediates pDC cytotoxicity. Based on our results, we believe that Jurkat cell lysis via TRAIL+ pDCs involves mainly, or even exclusively, membrane TRAIL for two reasons: we observed that 1) levels of sTRAIL in supernatants collected from unstimulated, TLR7/8- (IMQ), and TLR9-stimulated (CpG2216) pDCs are negligible and below those found in normal human serum (Fig. 5A); and 2) Jurkat cells express TRAIL-R2 exclusively (Fig. 4A). The latter finding is of particular relevance as membrane TRAIL can activate both TRAIL-R1 and -R2, whereas sTRAIL signals only via TRAIL-R1, unless secondarily cross-linked by Abs (20). It was therefore not unexpected that supernatants retrieved from IMQ-activated pDCs (Fig. 5B) and pDCs separated from their targets using a transwell (Supplemental Fig. 4) were unable to induce significant Jurkat cell lysis. Together, these findings indicate that the rapid lysis observed in our 2-h cytotoxicity assays requires cell contact and that membrane-bound TRAIL, but not sTRAIL, is the primary mediator of target cell lysis.

FIGURE 5.

Jurkat cell lysis by pDCs requires cell contact. A, The concentration of sTRAIL in the supernatants of pDCs (5 × 105 cells/ml) cultured with TLR7/8 (IMQ; n = 5) or TLR9 (CpG2216; n = 4) ligands overnight and in the serum of five healthy donors was determined by ELISA; limit of detection 32 pg/ml. Each dot denotes one individual measurement. B, Cytotoxic activity of IMQ-stimulated pDCs (E:T ratio 20:1) or their supernatants (from cultures of 5 × 105 cells/ml) against Jurkat target cells was determined after 2 h. Data represent mean percentage of specific lysis ± SD (n = 3).

FIGURE 5.

Jurkat cell lysis by pDCs requires cell contact. A, The concentration of sTRAIL in the supernatants of pDCs (5 × 105 cells/ml) cultured with TLR7/8 (IMQ; n = 5) or TLR9 (CpG2216; n = 4) ligands overnight and in the serum of five healthy donors was determined by ELISA; limit of detection 32 pg/ml. Each dot denotes one individual measurement. B, Cytotoxic activity of IMQ-stimulated pDCs (E:T ratio 20:1) or their supernatants (from cultures of 5 × 105 cells/ml) against Jurkat target cells was determined after 2 h. Data represent mean percentage of specific lysis ± SD (n = 3).

Close modal

Both in man and mice, tumor regression of melanocytic neoplasms upon treatment with topical IMQ is associated with an influx of pDCs (14, 21), but evidence that activated pDCs may directly exert antimelanoma activity is lacking. To address this issue, we performed cytotoxicity assays with the melanoma cell lines SKMel2 and WM793 as targets. Similar to Jurkat cells, these cell lines express TRAIL-R2, but not TRAIL-R1, as determined by FACS analysis (Fig. 6A). After 4–6 h of coculture, IMQ-activated pDCs effectively lysed WM793 and, less so, SKMel2 melanoma cells in a TRAIL-dependent fashion (Fig. 6B, 6C). The degree of melanoma cell lysis correlated roughly with their expression levels of TRAIL-R2 (Fig. 6A–C). In general, though, the extent of pDC-induced melanoma cell lysis was lower and required prolonged periods of coculture compared with Jurkat cells. This was even true when optimal killing conditions (skTRAIL) were used (Fig. 6D).

FIGURE 6.

Melanoma cells are sensitive to TRAIL-induced apoptosis mediated by IMQ- and/or IFN-α–activated pDCs. A, Expression of TRAIL-R1 and -R2 on melanoma target cells (SKMel2, WM793) was assessed by flow cytometry. B, C, E, and F, Cytotoxic activity of IMQ- (5 μg/ml) and IFN-α–activated (25,000 IU/ml) pDCs against the melanoma cell lines SKMel2 (B, E) and WM793 (C, F). D, The maximum level of TRAIL-induced lysis of SKMel2, WM793, and Jurkat cells was evaluated using recombinant skTRAIL (100 ng/ml). G, Evaluation of TRAIL expression on pDCs (left panel) and cytotoxic activity of the same cells (right panel) against WM793 targets after stimulation with IMQ (5 μg/ml; low-dose 0.05 μg/ml), IFN-α (25,000 IU/ml; low-dose 250 IU/ml), and combinations thereof. In all cytotoxicity assays (B, C, E–G), pDCs were preincubated with neutralizing anti-TRAIL (10 μg/ml) or isotype-matched control Abs for 30 min prior to performing the assay at E:T target ratios of 20:1. Data represent mean of duplicates of specific lysis ± SD. One representative example of three experiments performed and yielding similar results is shown for all assays.

FIGURE 6.

Melanoma cells are sensitive to TRAIL-induced apoptosis mediated by IMQ- and/or IFN-α–activated pDCs. A, Expression of TRAIL-R1 and -R2 on melanoma target cells (SKMel2, WM793) was assessed by flow cytometry. B, C, E, and F, Cytotoxic activity of IMQ- (5 μg/ml) and IFN-α–activated (25,000 IU/ml) pDCs against the melanoma cell lines SKMel2 (B, E) and WM793 (C, F). D, The maximum level of TRAIL-induced lysis of SKMel2, WM793, and Jurkat cells was evaluated using recombinant skTRAIL (100 ng/ml). G, Evaluation of TRAIL expression on pDCs (left panel) and cytotoxic activity of the same cells (right panel) against WM793 targets after stimulation with IMQ (5 μg/ml; low-dose 0.05 μg/ml), IFN-α (25,000 IU/ml; low-dose 250 IU/ml), and combinations thereof. In all cytotoxicity assays (B, C, E–G), pDCs were preincubated with neutralizing anti-TRAIL (10 μg/ml) or isotype-matched control Abs for 30 min prior to performing the assay at E:T target ratios of 20:1. Data represent mean of duplicates of specific lysis ± SD. One representative example of three experiments performed and yielding similar results is shown for all assays.

Close modal

Systemic treatment with rIFN-α has shown to increase disease-free and, when used in high doses, overall survival of stage II and III melanoma patients (22). In keeping with our observation of functionally active TRAIL on IFN-α–stimulated pDCs (Fig. 4D), we found that IFN-α–activated, but not unstimulated, pDCs effectively lysed both SKMel2 and WM793 target cells after 4–6 h (Fig. 6E, 6F) and that the levels of lysis were similar to those observed with IMQ-activated cells. pDC cytotoxicity was largely TRAIL-dependent, as demonstrated by the reduction in killing to baseline levels after addition of a neutralizing TRAIL Ab. We should not forget that the IFN-α doses (25,000 IU/ml) used in these experiments are supraoptimal and far exceed the blood/tissue levels measured in IFN-α–treated patients. We therefore asked whether clinically relevant doses of IFN-α (250 IU/ml) in combination with (suboptimal) IMQ doses (0.05 μg/ml) would lead not only to optimal TRAIL expression (Fig. 3D), but also to optimal tumor cell lysis. Fig. 6G shows that this is indeed the case. To the best of our knowledge, we show for the first time that IFN-α–treated pDCs exhibit cytotoxic activity against melanoma cells and that the combination of suboptimal stimulation conditions for IFN-α and IMQ may result in effective tumor cell lysis.

The finding that DCs can express cytotoxic molecules after TLR7/8 and TLR9 stimulation (8, 10) and can use them to effectively lyse cancer cells (9, 15, 23) as well as virus-infected targets (12) adds a novel facet to our current concept of DC biology. In this study, we sought to unravel the molecular events governing pDC cytotoxicity and, thus, gain a better understanding of the potential role of cytotoxic DCs in health and disease.

We were able to extend our previous results that TLR7/8-activated (IMQ) pDCs express TRAIL (15) by the finding that in the absence of other costimulators (i.e., IFN-α), TLR-induced TRAIL expression is exclusively associated with the activation of intracellular TLRs. This was not entirely unexpected, as pDCs express only low levels of TLR1 and TLR6 mRNA, react poorly to ligation of these moieties, and lack mRNA for all other extracellular TLRs (57). The selective induction of TRAIL through TLRs that are involved in viral DNA and RNA recognition may, at least partly, be responsible for the prominent role of pDCs during viral infection (3). It is of interest, however, that pDCs costimulated with low-dose IFN-α acquire the ability to upregulate TRAIL in response to TLR2/1 and TLR2/6 agonists, suggesting that IFN-α and TLR agonists may exert combined and additive effects. Notably, we also found modest TRAIL expression on pDCs that had been cultured with IL-3 (Fig. 1F). Because IL-3 lacks IFN-α–inducing properties (24), the reason for this phenomenon is yet unknown. TNF-α and CD40L apparently have a bimodal effect on pDCs. They both induce maturational events (e.g., upregulation of CD83) but, in contrast, fail to trigger TRAIL positivity. The latter effect is probably due to downregulation of IFN-α (25). This demonstrates that TRAIL expression on pDCs is not simply the result of promiscuous cell activation but preferentially occurs after stimulation with IFN-α–inducing agents, such as viruses and antimicrobial peptide-complexed autologous RNA (3, 26).

Different from the situation with TRAIL, we found the lytic molecule granzyme B to be abundantly expressed on freshly isolated/IL-3–activated pDCs but considerably less so after stimulation with TLR7/8 and TLR9 agonists. The functional role of this lytic molecule on pDCs has yet to be clarified. Several observations from us and others (15, 27) argue against its involvement in pDC-induced killing. First, unstimulated granzyme B+ pDCs fail to lyse tumor targets (Fig. 4B). Second, the reported dependence of granzyme B on other lytic molecules (e.g., perforin, granulysin) for CTL and NK cell effector functions (28) is in agreement with our findings of pDCs being consistently devoid of perforin, granulysin, FasL, and lysozyme. Other investigators reported that, after activation, pDC-derived granzyme B limits T cell expansion (29) and effectively kills tumor cells (27), and they also detected lysozyme (17). Methodological differences may well account for these discrepancies, but one may also entertain the possibility that pDC-associated granzyme B alone may exert its lytic function only after prolonged contact with its target (27).

An important observation of our study was that TLR-activated, TRAIL-expressing pDCs exhibit a mature phenotype and are efficient inducers of T cell proliferation in an allogeneic MLR (Fig. 2A–C). The latter result was not necessarily expected, given that IFN-α has known antiproliferative properties and that both HIV-infected pDCs (12) and CMV-infected monocyte-derived DCs (30) can delete activated T lymphocytes. The reasons why IFN-α and TRAIL do not influence T cell proliferation could be that the immunostimulatory effect of a mature DC outweighs its cytotoxic capacity or, most likely, that once T cells become sensitive to TRAIL by upregulating their TRAIL-Rs (after 3 d), pDCs have already lost their TRAIL expression (usually after 48 h) and/or have ceased to produce IFN-α.

IFN-α exhibits known TRAIL-inducing properties on PBMCs (3133). In this study, we show that TLR-induced TRAIL expression on pDCs (8, 9) and also, as a consequence, TRAIL-mediated pDC cytotoxicity is IFN-α dependent. More importantly, we demonstrate that tumor targets lysed by IFN-α–stimulated pDCs also include melanoma cells (Fig. 6E, 6F). In this situation, IFN-α may even exert a dual effect, as it also sensitizes tumor cells to TRAIL-dependent apoptosis (9, 34). This mechanism could be, at least partly, responsible for the reported efficacy of IFN-α as adjuvant treatment in tumor patients (e.g., high-risk primary melanoma) (22).

Given our findings, one may ask whether there is any potential advantage in using IMQ alone or in combination with IFN-α. There exist several reasons why we believe that this is the case. Firstly, IMQ when applied topically or injected into the tissue stimulates the release of proinflammatory cytokines and chemoattractants (Fig. 3A) (13). DCs represent a major portion of such an IMQ-induced infiltrate (14, 15, 21). Secondly, IMQ may have additional benefits in a therapeutic setting when compared with IFN-α alone. As we show in this study, the TRAIL-inducing properties of IFN-α at physiological concentrations are rather limited, but can be greatly enhanced even by suboptimal doses of IMQ (Figs. 3D, 6G). Lastly, what was said above should be re-emphasized, that IMQ, in contrast to IFN-α, causes phenotypic and functional pDC maturation, providing these cells with the capacity to induce productive primary immune responses. The combination of their tumoricidal and immunostimulatory properties thus makes TLR-stimulated DCs a double-edged sword in its true and best sense. Following this reasoning, cancer cell lysis effectuated by these cells would result in the release of tumor-associated Ags which, when taken up and presented by the very same cell population, would hopefully result in protective cancer immunity. Data obtained in a humanized mouse model support this notion (35). Clinical trials are needed and currently being devised to test both safety and efficacy of this approach.

We thank Dr. Patrick Brunner for assistance with multiplex cytokine analysis and Bärbel Reininger for assistance with cell isolation.

This work was supported by a research grant from the Austrian Science Fund, Vienna, Austria (DK-W1212-B13).

The online version of this article contains supplemental material.

Abbreviations used in this article:

BDCA

blood dendritic cell Ag

DC

dendritic cell

FasL

Fas ligand

IMQ

imiquimod

mDC

myeloid dendritic cell

MFI

mean fluorescence intensity

pDC

plasmacytoid dendritic cell

sTRAIL

soluble TRAIL

skTRAIL

super killer TRAIL.

1
Banchereau
J.
,
Steinman
R. M.
.
1998
.
Dendritic cells and the control of immunity.
Nature
392
:
245
252
.
2
Iwasaki
A.
,
Medzhitov
R.
.
2004
.
Toll-like receptor control of the adaptive immune responses.
Nat. Immunol.
5
:
987
995
.
3
Gilliet
M.
,
Cao
W.
,
Liu
Y. J.
.
2008
.
Plasmacytoid dendritic cells: sensing nucleic acids in viral infection and autoimmune diseases.
Nat. Rev. Immunol.
8
:
594
606
.
4
MacDonald
K. P.
,
Munster
D. J.
,
Clark
G. J.
,
Dzionek
A.
,
Schmitz
J.
,
Hart
D. N.
.
2002
.
Characterization of human blood dendritic cell subsets.
Blood
100
:
4512
4520
.
5
Jarrossay
D.
,
Napolitani
G.
,
Colonna
M.
,
Sallusto
F.
,
Lanzavecchia
A.
.
2001
.
Specialization and complementarity in microbial molecule recognition by human myeloid and plasmacytoid dendritic cells.
Eur. J. Immunol.
31
:
3388
3393
.
6
Kadowaki
N.
,
Ho
S.
,
Antonenko
S.
,
Malefyt
R. W.
,
Kastelein
R. A.
,
Bazan
F.
,
Liu
Y. J.
.
2001
.
Subsets of human dendritic cell precursors express different toll-like receptors and respond to different microbial antigens.
J. Exp. Med.
194
:
863
869
.
7
Hornung
V.
,
Rothenfusser
S.
,
Britsch
S.
,
Krug
A.
,
Jahrsdörfer
B.
,
Giese
T.
,
Endres
S.
,
Hartmann
G.
.
2002
.
Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides.
J. Immunol.
168
:
4531
4537
.
8
Hardy
A. W.
,
Graham
D. R.
,
Shearer
G. M.
,
Herbeuval
J. P.
.
2007
.
HIV turns plasmacytoid dendritic cells (pDC) into TRAIL-expressing killer pDC and down-regulates HIV coreceptors by Toll-like receptor 7-induced IFN-alpha.
Proc. Natl. Acad. Sci. USA
104
:
17453
17458
.
9
Chaperot
L.
,
Blum
A.
,
Manches
O.
,
Lui
G.
,
Angel
J.
,
Molens
J. P.
,
Plumas
J.
.
2006
.
Virus or TLR agonists induce TRAIL-mediated cytotoxic activity of plasmacytoid dendritic cells.
J. Immunol.
176
:
248
255
.
10
Colisson
R.
,
Barblu
L.
,
Gras
C.
,
Raynaud
F.
,
Hadj-Slimane
R.
,
Pique
C.
,
Hermine
O.
,
Lepelletier
Y.
,
Herbeuval
J. P.
.
2010
.
Free HTLV-1 induces TLR7-dependent innate immune response and TRAIL relocalization in killer plasmacytoid dendritic cells.
Blood
115
:
2177
2185
.
11
Chehimi
J.
,
Papasavvas
E.
,
Tomescu
C.
,
Gekonge
B.
,
Abdulhaqq
S.
,
Raymond
A.
,
Hancock
A.
,
Vinekar
K.
,
Carty
C.
,
Reynolds
G.
, et al
.
2010
.
Inability of plasmacytoid dendritic cells to directly lyse HIV-infected autologous CD4+ T cells despite induction of tumor necrosis factor-related apoptosis-inducing ligand.
J. Virol.
84
:
2762
2773
.
12
Stary
G.
,
Klein
I.
,
Kohlhofer
S.
,
Koszik
F.
,
Scherzer
T.
,
Müllauer
L.
,
Quendler
H.
,
Kohrgruber
N.
,
Stingl
G.
.
2009
.
Plasmacytoid dendritic cells express TRAIL and induce CD4+ T-cell apoptosis in HIV-1 viremic patients.
Blood
114
:
3854
3863
.
13
Schön
M. P.
,
Schön
M.
.
2008
.
TLR7 and TLR8 as targets in cancer therapy.
Oncogene
27
:
190
199
.
14
Palamara
F.
,
Meindl
S.
,
Holcmann
M.
,
Lührs
P.
,
Stingl
G.
,
Sibilia
M.
.
2004
.
Identification and characterization of pDC-like cells in normal mouse skin and melanomas treated with imiquimod.
J. Immunol.
173
:
3051
3061
.
15
Stary
G.
,
Bangert
C.
,
Tauber
M.
,
Strohal
R.
,
Kopp
T.
,
Stingl
G.
.
2007
.
Tumoricidal activity of TLR7/8-activated inflammatory dendritic cells.
J. Exp. Med.
204
:
1441
1451
.
16
Blomberg
K.
,
Hautala
R.
,
Lövgren
J.
,
Mukkala
V. M.
,
Lindqvist
C.
,
Åkerman
K.
.
1996
.
Time-resolved fluorometric assay for natural killer activity using target cells labelled with a fluorescence enhancing ligand.
J. Immunol. Methods
193
:
199
206
.
17
Matsui
T.
,
Connolly
J. E.
,
Michnevitz
M.
,
Chaussabel
D.
,
Yu
C. I.
,
Glaser
C.
,
Tindle
S.
,
Pypaert
M.
,
Freitas
H.
,
Piqueras
B.
, et al
.
2009
.
CD2 distinguishes two subsets of human plasmacytoid dendritic cells with distinct phenotype and functions.
J. Immunol.
182
:
6815
6823
.
18
Chawla-Sarkar
M.
,
Lindner
D. J.
,
Liu
Y. F.
,
Williams
B. R.
,
Sen
G. C.
,
Silverman
R. H.
,
Borden
E. C.
.
2003
.
Apoptosis and interferons: role of interferon-stimulated genes as mediators of apoptosis.
Apoptosis
8
:
237
249
.
19
Gonzalvez
F.
,
Ashkenazi
A.
.
2010
.
New insights into apoptosis signaling by Apo2L/TRAIL.
Oncogene
29
:
4752
4765
.
20
Wajant
H.
,
Moosmayer
D.
,
Wüest
T.
,
Bartke
T.
,
Gerlach
E.
,
Schönherr
U.
,
Peters
N.
,
Scheurich
P.
,
Pfizenmaier
K.
.
2001
.
Differential activation of TRAIL-R1 and -2 by soluble and membrane TRAIL allows selective surface antigen-directed activation of TRAIL-R2 by a soluble TRAIL derivative.
Oncogene
20
:
4101
4106
.
21
Wolf
I. H.
,
Kodama
K.
,
Cerroni
L.
,
Kerl
H.
.
2007
.
Nature of inflammatory infiltrate in superficial cutaneous malignancies during topical imiquimod treatment.
Am. J. Dermatopathol.
29
:
237
241
.
22
Mocellin
S.
,
Pasquali
S.
,
Rossi
C. R.
,
Nitti
D.
.
2010
.
Interferon alpha adjuvant therapy in patients with high-risk melanoma: a systematic review and meta-analysis.
J. Natl. Cancer Inst.
102
:
493
501
.
23
Fanger
N. A.
,
Maliszewski
C. R.
,
Schooley
K.
,
Griffith
T. S.
.
1999
.
Human dendritic cells mediate cellular apoptosis via tumor necrosis factor-related apoptosis-inducing ligand (TRAIL).
J. Exp. Med.
190
:
1155
1164
.
24
Krug
A.
,
Rothenfusser
S.
,
Hornung
V.
,
Jahrsdörfer
B.
,
Blackwell
S.
,
Ballas
Z. K.
,
Endres
S.
,
Krieg
A. M.
,
Hartmann
G.
.
2001
.
Identification of CpG oligonucleotide sequences with high induction of IFN-alpha/beta in plasmacytoid dendritic cells.
Eur. J. Immunol.
31
:
2154
2163
.
25
Palucka
A. K.
,
Blanck
J. P.
,
Bennett
L.
,
Pascual
V.
,
Banchereau
J.
.
2005
.
Cross-regulation of TNF and IFN-alpha in autoimmune diseases.
Proc. Natl. Acad. Sci. USA
102
:
3372
3377
.
26
Ganguly
D.
,
Chamilos
G.
,
Lande
R.
,
Gregorio
J.
,
Meller
S.
,
Facchinetti
V.
,
Homey
B.
,
Barrat
F. J.
,
Zal
T.
,
Gilliet
M.
.
2009
.
Self-RNA-antimicrobial peptide complexes activate human dendritic cells through TLR7 and TLR8.
J. Exp. Med.
206
:
1983
1994
.
27
Bratke
K.
,
Nielsen
J.
,
Manig
F.
,
Klein
C.
,
Kuepper
M.
,
Geyer
S.
,
Julius
P.
,
Lommatzsch
M.
,
Virchow
J. C.
.
2010
.
Functional expression of granzyme B in human plasmacytoid dendritic cells: a role in allergic inflammation.
Clin. Exp. Allergy
40
:
1015
1024
.
28
Lieberman
J.
2003
.
The ABCs of granule-mediated cytotoxicity: new weapons in the arsenal.
Nat. Rev. Immunol.
3
:
361
370
.
29
Jahrsdörfer
B.
,
Vollmer
A.
,
Blackwell
S. E.
,
Maier
J.
,
Sontheimer
K.
,
Beyer
T.
,
Mandel
B.
,
Lunov
O.
,
Tron
K.
,
Nienhaus
G. U.
, et al
.
2010
.
Granzyme B produced by human plasmacytoid dendritic cells suppresses T-cell expansion.
Blood
115
:
1156
1165
.
30
Raftery
M. J.
,
Schwab
M.
,
Eibert
S. M.
,
Samstag
Y.
,
Walczak
H.
,
Schönrich
G.
.
2001
.
Targeting the function of mature dendritic cells by human cytomegalovirus: a multilayered viral defense strategy.
Immunity
15
:
997
1009
.
31
Kayagaki
N.
,
Yamaguchi
N.
,
Nakayama
M.
,
Eto
H.
,
Okumura
K.
,
Yagita
H.
.
1999
.
Type I interferons (IFNs) regulate tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) expression on human T cells: A novel mechanism for the antitumor effects of type I IFNs.
J. Exp. Med.
189
:
1451
1460
.
32
Sato
K.
,
Hida
S.
,
Takayanagi
H.
,
Yokochi
T.
,
Kayagaki
N.
,
Takeda
K.
,
Yagita
H.
,
Okumura
K.
,
Tanaka
N.
,
Taniguchi
T.
,
Ogasawara
K.
.
2001
.
Antiviral response by natural killer cells through TRAIL gene induction by IFN-alpha/beta.
Eur. J. Immunol.
31
:
3138
3146
.
33
Herbeuval
J. P.
,
Boasso
A.
,
Grivel
J. C.
,
Hardy
A. W.
,
Anderson
S. A.
,
Dolan
M. J.
,
Chougnet
C.
,
Lifson
J. D.
,
Shearer
G. M.
.
2005
.
TNF-related apoptosis-inducing ligand (TRAIL) in HIV-1-infected patients and its in vitro production by antigen-presenting cells.
Blood
105
:
2458
2464
.
34
Shigeno
M.
,
Nakao
K.
,
Ichikawa
T.
,
Suzuki
K.
,
Kawakami
A.
,
Abiru
S.
,
Miyazoe
S.
,
Nakagawa
Y.
,
Ishikawa
H.
,
Hamasaki
K.
, et al
.
2003
.
Interferon-alpha sensitizes human hepatoma cells to TRAIL-induced apoptosis through DR5 upregulation and NF-kappa B inactivation.
Oncogene
22
:
1653
1662
.
35
Aspord
C.
,
Charles
J.
,
Leccia
M. T.
,
Laurin
D.
,
Richard
M. J.
,
Chaperot
L.
,
Plumas
J.
.
2010
.
A novel cancer vaccine strategy based on HLA-A*0201 matched allogeneic plasmacytoid dendritic cells.
PLoS ONE
5
:
e10458
.

The authors have no financial conflicts of interest.