Depending upon which TLRs are triggered, dendritic cells (DCs) may orient the differentiation of naive CD4+ T cells toward either Th1, Th2, regulatory T cells, or the recently defined Th17 lineage. In this study, we report that a dual stimulation of TLR4 and TLR7/8 with LPS plus R848 leads human monocyte-derived DCs (MoDCs) to produce multiple pro- and anti-inflammatory cytokines, including IL-10, IL-12, and IL-23. Surprisingly, a significant variability in the up-regulation of these cytokines is observed in DCs obtained from various healthy donors, with approximately one of three being “high responders.” High responding MoDCs stimulated via TLR4 and TLR7/8 induce naive allogeneic CD4+ T cell to secrete sequentially IL-10 and IFN-γ, and eventually IL-17A, whereas low responding MoDCs only stimulate IFN-γ production. Both TLR7 and TLR8 play a central role in this phenomenon: TLR4 triggering with LPS up-regulates TLR7 expression on human MoDCs from high responders, silencing of either TLR7 or TLR8 mRNAs inhibits cytokine production in LPS plus R848-treated MoDCs, and plasmacytoid DCs constitutively expressing high levels of TLR7 induce the production of IL-10, IFN-γ, and IL-17A by naive T cells when stimulated with R848 alone. Collectively, our results illustrate the synergy between TLR4 and TLR7/8 in controlling the sequential production of regulatory and proinflammatory cytokines by naive CD4+ T cells. The observed polymorphism in DC responses to such TLR-mediated stimuli could explain differences in the susceptibility to infectious pathogens or autoimmune diseases within the human population.
Following engagement with conserved microbial motifs (1, 2, 3), TLRs play a critical role in inducing appropriate immune responses against potential pathogenic agents (4). Most particularly, signals provided to dendritic cells (DCs)2 through selected TLRs are known to impact dramatically the cytokine milieu in which the Ag is presented, thus leading to distinct patterns of naive CD4+ T cell differentiation (5, 6, 7, 8, 9). Th1 cells are induced by IL-12-producing APCs, produce IFN-γ, and protect against intracellular bacteria or viruses. Chronic stimulation of Th1 cells may lead to autoimmunity (10). Th2 cells are elicited by IL-4-secreting APCs and produce IL-4, IL-5, and IL-13 cytokines. As such, they represent an effective defense mechanism against parasites, but are also associated with type I allergic inflammation (11). CD4+ regulatory T cells (Tregs) capable of suppressing both Th1 and Th2 immune responses are induced by DCs producing IL-10 and/or TGF-β. Naturally occurring Tregs express the transcription factor Foxp3 and produce IL-10 and/or TGF-β, whereas Th3 and Tr1 cells produce either TGF-β or IL-10, respectively (12, 13, 14). Recently, several studies established that DCs producing IL-1β, IL-6, IL-23, and/or TGF-β induce the differentiation of a new population of IL-17A-secreting CD4+ Th cells termed Th17 (5, 6, 7, 8, 9, 15, 16, 17, 18). Th17 cells contribute to defense mechanisms against extracellular infectious pathogens but may also cause autoimmune diseases (19, 20, 21, 22, 23). Immune responses to a pathogen or a danger signal are thus the end result of a balance between effector (proinflammatory) and regulatory mechanisms involving those various CD4+ T cell subsets at varying levels in the course of the response.
In the present study, we document that both human monocyte-derived (MoDCs) and plasmacytoid DCs (pDCs) produce high levels of IL-10, IL-12, and IL-23, albeit with unexpected variability depending upon individuals, when receiving a strong signal via TLR7 and/or TLR8. High responding DCs support a sequential differentiation of naive CD4+ T cells into IL-10, IFN-γ, and IL-17-producing cells, likely to mediate both regulatory and proinflammatory functional programs, respectively.
Materials and Methods
Cells, reagents, and DC/T cocultures
Heparinized blood from healthy volunteers (obtained from Etablissement Français du Sang) was centrifuged over a Ficoll-Paque plus gradient (GE Healthcare) to isolate PBMCs. To generate MoDCs, 108 cells were cultured at 37°C in 5% CO2 in a 75-ml plastic flask in 25 ml of culture medium (RPMI 1640 supplemented with 2 mM l-glutamine, 20 μg/ml gentamicin, 50 μM 2-ME, 1% nonessential amino acids (all obtained from Invitrogen) and 10% FBS (Gentaur)). After 2 h, nonadherent cells were removed, and adherent cells were further cultivated for 7 days in presence of recombinant human GM-CSF and IL-4, 80 and 25 ng/ml, respectively (Gentaur). After 7 days at 37°C, a pure population of DCs was obtained, with more than 95% CD1a+ cells detected by flow cytometry using a FC500 cytometer and CXP analysis software (Beckman Coulter). To test the effect of TLR ligands on DCs, 106 MoDCs were plated in a 24-well plate in 1 ml of culture medium in presence of either medium, highly purified LPS from Escherichia coli (1 μg/ml; InvivoGen), Resiquimod (R848, 1 rusha μg/ml; InvivoGen) or a combination of LPS plus R848. After 24 h at 37°C and 5% CO2, MoDCs were lysed to isolate total RNA or washed and cultured in a 24-well plate with allogeneic CD4+ naive T cells at a 1:10 DC/T ratio in 1 ml of culture medium. Naive CD4+ T cells were isolated from PBMCs by negative selection using a Dynal CD4+ isolation kit (Invitrogen) according to the manufacturer’s instructions. Such naive CD4+ + T cells at a 1:10 pDC/T ratio.
Measurement of cytokine production
+ T cells were measured using the Flex Set cytometric beads array (CBA) (BD Bioscience), according to the manufacturer’s instructions. Analysis of intracellular cytokine levels was performed on naive CD4+ T cells cultured with MoDCs pretreated with TLR ligands. CD4+ T cells were recovered and restimulated with plate-bound anti-CD3 (1 μg/ml X35; Beckman Coulter) for 4 h in presence of GolgiPlug (BD Bioscience). Cells were surface-stained with anti-CD4 and anti-CD25 Abs (Beckman Coulter), then fixed and permeabilized using the Intraprep reagent (Beckman Coulter). Cells were subsequently washed and double-stained with either PE-labeled anti-IL-10 (BD Bioscience) or anti-IL-17A (eBioscience) Abs, as well as a FITC-conjugated anti-IFN-γ Ab (BD Bioscience). Samples were analyzed on a FC500 flow cytometer (Beckman Coulter).
RNA isolation and quantitative real-time PCR analysis
Total RNA was extracted from DCs or T cells using a RNeasy mini kit (Qiagen), and cDNAs were synthesized using TaqMan reverse transcription reagents (Applied Biosystems) as per the manufacturer’s instructions. mRNA expression was evaluated by quantitative PCR on a 7300 real-time PCR system (Applied Biosystems) with predesigned Taqman gene expression assays and reagents, according to the manufacturer’s instructions. The expression of the following genes was assessed in MoDCs: IL-12p35 (Hs00168405_m1), IL-12p40 (Hs00233688_m1), IL-23p19 (Hs00372324_m1), IL-1β (Hs00174097_m1), IL-6 (Hs00174131_ m1), TNF-α (Hs00174128_m1), and IL-4 (Hs00174122_m1). To monitor T cell polarization, the expression of the following genes was evaluated: Tbet (Hs00203436_m1), GATA-3 (Hs00231122_m1), Fox p3 (Hs00203958_m1), RORc2 (Hs01076112_m1), IFN-γ (Ηs00174143_m1), IL-4 (Hs00174122_ m1), IL-10 (Hs00174086_m1), TGF-β (Hs00171257_m1), and IL-17A (Hs00174383_m1). Data were interpreted, for each target gene monitored in MoDCs or T cells, in comparison with endogenous β-actin (Hs99999903_m1) as a control. RNA 18S (Hs99999901_s1) was used as a control for gene expression analysis in pDCs. The relative amount of target genes in each sample was calculated in comparison with the calibrator sample using the ΔΔCt method. The magnitude of gene induction was calculated using the formula 2−ΔΔCt = 2−(ΔCt for unstimulated cells − ΔCt for stimulated cells).
To inhibit TLR7 and/or TLR8 expression, Silencer Select Pre-designed siRNAs (Applied Biosystems) specific for TLR7 or TLR8 were used. In each experiment, two 105 MoDCs were plated in 24-well plates in 400 μl of growth medium containing 10% FBS without antibiotic. Lipofectamine 2000 (1 μl/well; Invitrogen) was first diluted in 50 μl of Opti-MEM (Invitrogen) for 5 min before mixing with an equal volume of Opti-MEM containing 20 pmol of siRNA. After 20 min, 100 μl of the Lipofectamine/siRNA mix were added to the cells to perform siRNA transfection. Fresh growth medium was added 4 h after transfection. Cells were cultured for 36 h at 37°C to obtain optimal silencing of targeted genes. The efficacy of gene silencing was tested by real-time PCR using siRNA control and siRNA GAPDH as negative and positive controls, respectively.
Levels of cytokine produced were compared using a Student’s t test. A value of p < 0.05 was considered statistically significant. The multiple regression correlation coefficient R2 was calculated to estimate the linear relationship between cytokine gene expression by MoDCs and cytokine gene expression by CD4+ T cells.
Polarization of naive CD4+ T cells induced by MoDCs treated with LPS plus R848
We assessed the impact of DCs treated with LPS with or without R848 on Th cell polarization. MoDCs were treated for 24 h with either medium, LPS, R848, or LPS plus R848, then washed and cultured with purified allogeneic naive CD4+ T cells. After 1, 3, and 6 days, the expression of genes encoding for specific cytokines and transcription factors associated with either Th1, Th2, Treg, or Th17 differentiation was evaluated by quantitative PCR. As shown in Fig. 1⇓A, two distinct patterns of naive CD4+ T cell differentiation were observed after coculture with LPS plus R848-treated MoDCs. In three of nine donors, a strong (i.e., at least 50-fold) induction of IL-10, IFN-γ, and IL-17A genes was detected after 1, 3, and 6 days, respectively, of coculture (Fig. 1⇓A, open triangles). This pattern of cytokine gene expression was not observed in T cells cocultured with MoDCs treated with LPS or R848 alone (Fig. 1⇓A). In cocultures made between MoDCs/T cells from most donors, only IFN-γ gene expression was up-regulated to a lower level after 3 days (Fig. 1⇓A, black dots). In contrast to such low responding patients, high responders are thus defined based on the capacity of their MoDCs to induce a strong and sequential expression of IL-10/IFN-γ/IL-17A genes by naive CD4+ T cells, after LPS plus R848 stimulation.
Using intracellular cytokine staining, IL-10-secreting CD4+ T cells were detected at day 3 only in cocultures obtained with high responding MoDCs treated with LPS plus R848 (Fig. 1⇑B). IFN-γ production was up-regulated in T cells with a peak at day 5 whereas IL-10 was no longer detected in culture supernatants (Fig. 1⇑B). Multiple labeling experiments established that IL-10 and IFN-γ-secreting cells represent two distinct populations (Fig. 1⇑B), and that IL-10-producing T cells are CD25 negative (data not shown). No significant differences were observed regarding Foxp3 expression by CD4+ T cells, whatever the treatment of DCs (data not shown). CD4+ T cells produced IL-17A starting from day 7, with 35% of these cells secreting both IL-17A and IFN-γ. In cocultures performed with low responding MoDCs, only IFN-γ was detected, as early as day 3 (Fig. 1⇑B, right panel). All results obtained by intracellular cytokine staining were confirmed when measuring secreted cytokines in culture supernatants (data not shown).
The sequential polarization of CD4+ T cell responses is due to the synergistic up-regulation of cytokine production by MoDCs following stimulation via TLR4 and TLR7/8
To determine the cause of the differential CD4+ T cell polarization depending on MoDC donors, we further tested the effect of a single or combined TLR4 or TLR7/8 stimulation on MoDCs, evaluating changes in surface phenotype, gene expression, and cytokine production as read-outs. Whether used alone or in combination, both LPS (TLR4 ligand) and R848 (TLR7/8 ligand) molecules induced a comparable maturation of MoDCs, with a clear up-regulation of CD40, CD83, CD86, and HLA-DR surface molecules within 24 h (data not shown). The expression of various cytokine genes was quantified by real-time PCR after stimulation of MoDCs obtained from a total of 26 healthy donors (Fig. 2⇓A). In these experiments, a clear synergy between LPS and R848 was observed in enhancing concomitantly the expression of IL-12p35, IL-12p40, IL-23p19, IL-6, IL-1β, IL-10, and ΤΝF-α genes in all donors, even if the magnitude of the up-regulation varied dramatically between donors (Fig. 2⇓A). A lower but substantial up-regulation of these genes was still observed following stimulation with a single TLR agonist. MoDCs from eight high responding donors, i.e., confirmed to elicit a strong cytokine production by T cells, exhibited a major increase in both IL-12p35 (>2115-fold), IL-12p40 (>4286-fold), IL-23p19 (>1108-fold), IL-6 (>587-fold), IL-1β (>37-fold), IL-10 (>13-fold), and ΤΝF-α (>16-fold) gene expression following stimulation via TLR4 and TLR7/8 receptors (Fig. 2⇓A, open triangles). In MoDCs from low responders, the up-regulation of cytokine genes following stimulation was usually significantly lower (Fig. 2⇓A, black dots). Such differences between high and low responders were confirmed following direct measurement of cytokine levels in culture supernatants from DCs (Fig. 2⇓B).
We subsequently related cytokine gene expression by LPS plus R848-treated MoDCs with cytokine induction in naive CD4+ T cells. Specifically, we noticed four clear correlations (i.e., between IL-10 in MoDCs/IL-10 in T cells at day 1, IL-12p35 and IL-12p40 in MoDCs/IFN-γ in T cells at day 3, and IL-23 in MoDCs/IL-17 in T cells at day 6) (Fig. 3⇓A). These experiments also suggested differences in terms of cytokine gene expression by LPS plus R848-treated MoDCs from low vs high responders (Fig. 3⇓A, black dots and open triangles, respectively).
To highlight the role of IL-10, IL-12, and IL-23 produced by LPS plus R848-treated MoDCs in the sequential differentiation of naive CD4+ T cells, various blocking Abs were added to the cocultures. As shown in Fig. 3⇑B, the anti-IL-10 Ab up-regulated both IL-17A and IFN-γ gene expression by T cells, the anti-IL-12 Ab decreased IFN-γ gene expression whereas the anti-IL-23 Ab down-regulated IL-17A gene expression.
To inquire about possible mechanisms, we further analyzed TLR4, 7, and 8 gene expression by MoDCs after LPS, R848, or LPS plus R848 stimulation. Engagement of TLR4 or TLR7/8, alone or in combination, up-regulated both TLR7 and TLR8 genes. Interestingly, TLR up-regulation after DC stimulation was consistently higher in high vs low responders (Fig. 4⇓A, open triangles vs black dots). Similarly, high responders exhibited higher baseline levels in TLR4 and TLR8 gene expression when compared with low responders (Fig. 4⇓B), and the latter correlated with the level of induction of IL-10, IL-12A, IL-12B, and IL-1β genes following stimulation with either LPS or R848 (data not shown).
Altogether, these results illustrated that a combined TLR4 and TLR7/8 stimulation of MoDCs triggers a major up-regulation of both IL-12, IL-23, IL-6, IL-1β, IL-10, and TNF-α production as well as an increased TLR7 and TLR8 expression in approximately one of three healthy donors tested. Polymorphisms observed in those responses are probably related to a variation among donors in the capacity of TLR4 engagement to up-regulate TLR7 and TLR8 expression.
Role of TLR7 and TLR8 in the induction of IL-10, IL-12, and IL-23 production by human DCs
To investigate the respective contribution of TLR7 and TLR8 in the up-regulation of cytokines aforementioned, blocking experiments with siRNAs specific for either TLR7 or TLR8 were performed using high responding MoDCs treated with LPS plus R848. The specificity of RNA silencing was confirmed using an irrelevant siRNA as a control (Fig. 5⇓A). Silencing of either TLR7 or TLR8 resulted in a decreased production of IL-10 and IL-12p70, whereas it had no significant effect on IL-23 production (Fig. 5⇓B). Dual silencing of TLR7 and 8 induced a strong down-regulation of IL-10, IL-12p70, and IL-23, suggesting an additive and possibly synergistic role of these two TLRs in controlling the production of these cytokines (Fig. 5⇓B). Although not shown, silencing of TLR7 and/or 8 had no impact on patterns of cytokines produced by MoDCs stimulated with LPS.
To extend these results, we assessed the impact of TLR7 stimulation on purified pDCs, which constitutively express high levels of this receptor, in the absence of TLR4 and TLR8. Stimulation with R848 alone was sufficient to enhance IL-10, IL-12p35, IL-12p40, IL-23, and IL-10 gene expression by pDCs (Fig. 6⇓A), as well as the production of corresponding cytokines (Fig. 6⇓B). Based on the observed variability in the responses among donors, it was possible to distinguish high and low responders (Fig. 6⇓, open triangles and black dots, respectively). High responding pDCs exhibited a parallel up-regulation of IL-10 (>19.5-fold), IL-12p35 (>10-fold), IL-12p40 (>15-fold), IL-23 (>47-fold), and TLR7 (>12-fold) genes (Fig. 6⇓A, open triangles). Noteworthy, levels of cytokine mRNAs in TLR7 stimulated-pDCs were much lower when compared with MoDCs stimulated with TLR4 plus TLR7/8 ligands.
We also compared patterns of cytokine gene expression observed in MoDCs stimulated by LPS+R848 and pDCs incubated with R848 in four healthy donors. A clear correlation was observed for all donors, suggesting that the status of high/low responsiveness for a given donor extends to both MoDC and pDC subsets (Fig. 6⇑C).
R848-treated pDCs trigger a differentiation of naive CD4+ T cells toward IFN-γ, IL-10, and IL-17 secreting cells
To assess the impact of R848-treated pDCs on Th polarization, pDCs were stimulated with medium or R848 for 24 h, then washed and cocultured for up to 6 days with naive allogeneic CD4+ T cells. In cocultures performed with R848-treated pDCs from high responders (50% of donors), a strong induction of IFN-γ gene expression was detected at day 1, whereas IL-10 and IL-17A genes were only significantly up-regulated after 6 days of coculture (Fig. 7⇓A, upper panel). Secretion of IL-10, IFN-γ, and IL-17A cytokines was confirmed in 6-day culture supernatants (Fig. 7⇓B, upper panel). As previously observed for MoDCs, low responding pDCs treated with R848 drove the differentiation of naive CD4+ T cells mostly toward IFN-γ-producing cells, with only small amounts of IL-10 and IL-17A detected (Fig. 7⇓, A and B, lower panels). In these experiments, neither gene expression nor cytokine production was detected in control conditions lacking T lymphocytes (data not shown).
Following recognition of viral or microbial pathogen associated molecular patterns, TLRs expressed by DCs critically influence the polarization of Ag-specific CD4+ T cell responses toward either the Th1, Th2, Th17, or Treg lineages (3, 4). In this context, selected TLR ligands can be used alone or in combination, as potential vaccine adjuvants to elicit the most appropriate immune response in humans.
In the present study, we observed, in agreement with a previous report (24), that a combination of TLR4 and TLR7/8 ligands acts in synergy to induce the production of multiple cytokines in MoDCs. The latter include IL-10, IL-12, and IL-23, known to subsequently regulate distinct functional programs in naive CD4+ T cells associated with Treg (10), Th1 (13, 14), or Th17 differentiation (19), respectively. Levels of cytokine up-regulation in MoDCs following joint stimulation via TLR4/TLR7/TLR8 differed dramatically between healthy donors, even if some level of synergy was always observed. One of three donors had MoDCs exhibiting a strong up-regulation of cytokine genes. Such high responding DCs were capable, when cultured with naive CD4+ T cells, to induce a sequential production of IL-10, IFN-γ, and, eventually, IL-17A, with peaks after 3, 5, and 7 days, respectively. pDCs stimulated through TLR7 alone also produced IL-10, IL-12, and IL-23. Interestingly, high responding pDCs elicited a distinct sequence of CD4+ T cell polarization when cultured with naive CD4+ T cells, compared with LPS plus R848-treated MoDCs, with an initial IFN-γ-production followed by IL-17A and IL-10 secretion after 3 and 6 days, respectively. The reason for such a difference in the kinetics of cytokine secretion is presently unclear, but could be linked with differences in the regulation of IL-10, IL-12, and IL-23 gene expression in those two DC subpopulations. Nonetheless, for a given donor the status of high/low responsiveness was found to extend to both MoDC and pDC subsets.
Using additional agonists specific for TLR7 (i.e., Imiquimod, Gardiquimod) or TLR8 (ssPolyU) together with LPS, we confirmed that a significant synergy in cytokine induction is consistently observed after joint engagement of TLR4 plus TLR7 and/or TLR8 (data not shown). We observed that TLR7, which is not present in MoDCs under normal conditions, is dramatically up-regulated in selected donors after stimulation by LPS, in agreement with a previous study (25). Thus, the observed polymorphism between high and low responders is probably due to differences in TLR7/8 up-regulation following TLR4 stimulation, suggesting that a threshold stimulation of TLR7 and/or TLR8 is required to activate the joint secretion of multiple cytokines by MoDCs. Consistent with this, single or dual silencing of TLR7 and TLR8 genes using siRNAs confirmed that the two TLRs act in parallel, and possibly in synergy, to up-regulate IL-10, IL-12p70, and IL-23 production in MoDCs. TLR7 plays a crucial role in controlling cytokine production, as seen in pDCs, which express only TLR7 following stimulation with R848 alone.
Pathogenic stimuli capable of engaging both TLR4 and TLR7/8 have yet to be described. Indeed, TLR4 is a surface receptor known to sense bacterial motifs (e.g., LPS), whereas endosomal TLR7/8 receptors rather detect viral components (e.g., ssRNA). One possible inference from the present study is that massive exposure to infectious pathogens leading to a strong engagement of TLR7 and/or TLR8 on MoDCs could result in a Th1 and Th17 protective immune response, after overcoming an initial Treg response. The distinct kinetics of cytokine secretion induced in T cells by TLR7-stimulated pDCs is more puzzling, because it eventually results in a mixed regulatory/proinflammatory T cell response associating IL-10 and IL-17A production.
That TLR7 and/or TLR8 stimulation on DCs leads to Th17 cell differentiation raises the possibility that this could also contribute to the pathogenicity of autoimmune diseases (21) such as rheumatoid arthritis (22) or systemic lupus erythematosus (23). In patients with rheumatoid arthritis, both TLR4 and TLR7 are highly expressed by synovial cells, and TLR4 ligands are increased in both serum and synovial fluid (22). Similarly, both TLR4 hyperresponsiveness as well as the presence of circulating TLR7 and TLR8 agonists, such as the small nuclear ribonucleoprotein particle U1, have been reported in patients with systemic lupus erythematosus (26, 27). Thus, combined stimulation of TLR4 and TLR7/8 could explain the production of proinflammatory cytokines in some patients with autoimmune diseases (28, 29).
Collectively, our results establish that a combination of TLR4 and a TLR7/8 ligands act in synergy to enhance the production of IL-10, IL-12, and IL-23 by MoDCs. The induction of these genes varies dramatically among healthy individuals, probably as a consequence of differences in levels of TLR7 and 8 expressed by DCs in a constitutive or inducible (e.g., following TLR4 engagement) manner. These data further document the role of DCs in integrating microbial or viral stimuli to drive appropriate adaptive immune responses associating both regulatory and proinflammatory cytokines in a coordinated manner.
The authors thank Pr. David Klatzmann for his useful comments.
The authors have no financial conflict of interest. All the coauthors are employees at the biopharmaceutical company Stallergènes S.A.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Address correspondence and reprint requests to Dr. Phillipe Moingeon, Research and Development, Stallergenes S.A., 6 Rue Alexis de Tocqueville, Antony Cedex 92183, France. E-mail address:
↵2 Abbreviations used in this paper: DC, dendritic cell; CBA, cytometric bead array; MoDC, monocyte-derived dendritic cell; pDC, plasmacytoid dendritic cell; Treg, regulatory T cell.
- Received June 17, 2008.
- Accepted January 6, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.