The quorum-sensing molecule farnesol is produced by the opportunistic human fungal pathogen Candida albicans. Aside from its primary function of blocking the transition from yeast to hyphal morphotype, it has an immunomodulatory role on human dendritic cells (DC) through the alteration of surface markers, cytokine secretion, and their ability to activate T cells. Nonetheless, the molecular mechanisms by which farnesol modulates DC differentiation and maturation remained unknown. In this study, we demonstrate through transcriptional and functional assays that farnesol influences several signaling pathways during DC differentiation and in response to TLR agonists. In particular, farnesol increases the expression of the Ag-presenting glycoprotein CD1d through the nuclear receptors PPARγ and RARα, as well as p38 MAPK. However, the higher expression of CD1d did not confer these DC with an enhanced capacity to activate CD1d-restricted invariant NKT cells. In the presence of farnesol, there is reduced secretion of the Th1-inducing cytokine, IL-12, and increased release of proinflammatory cytokines, as well as the anti-inflammatory cytokine IL-10. These changes are partially independent of nuclear receptor activity but, in the case of TNF-α and IL-10, dependent on NF-κB and MAPK pathways. Interestingly, renewal of the IL-12/IL-10 milieu restores the ability of farnesol-differentiated DC to activate invariant NKT, Th1, and FOXP3+ regulatory T cells. Our results show that farnesol modulates nuclear receptors, NF-κB, and MAPK-signaling pathways, thereby impairing the capacity of DC to activate several T cells subsets and potentially conferring C. albicans, an advantage in overcoming DC-mediated immunity.
Dendritic cells (DC) are professional APCs. After taking up invading pathogens or other nonself-material in their immature form, DC migrate to draining lymph nodes and undergo a maturation process. As mature DC (mDC), these cells play a key role in the immune system by instructing T cells to elicit a response (1, 2). Optimal maturation of human DC requires the coordinated activation of several pathways, including NF-κB, a family of transcription factors that regulates inflammatory processes (3, 4). Indeed, LPS-induced maturation of DC involves the nuclear translocation of NF-κB, leading to the upregulation of costimulatory molecules (3, 4). Moreover, activation of NF-κB is known to be essential for the release of proinflammatory cytokines, such as IL-6, TNF-α, and IL-12, resulting in elevated Ag presentation and T cell activation (5, 6). The transcriptional activities of NF-κB can be modulated through their phosphorylation by members of the MAPK signaling pathway (7–9). Accordingly, inhibition of p38 and JNK MAPK block the upregulation of the costimulatory molecules CD40, CD80, and CD86 and the release of proinflammatory cytokines (IL-1α, IL-1β, TNF-α, and IL-12) induced by LPS stimulation, resulting in a reduced capacity to prime T cells (3, 10, 11).
Although DC maturation is tightly regulated by NF-κB and MAPK signaling pathways, this process is also influenced by the extra- and intracellular lipid milieu. These environments are monitored by a range of nuclear receptors, including the peroxisome proliferator activated receptor γ (PPARγ), retinoic acid receptor α (RARα), and liver X receptor α (LXRα). PPARγ activation in particular can alter the expression of CD1, CD80, and CD86 molecules (2), resulting in reduced secretion of IL-12 and the T cell–inducing chemokines CXCL10 and RANTES (12, 13). There are conflicting results regarding whether RARα activation enhances or represses DC maturation. In some studies, retinoic acid (all-trans retinoic acid [ATRA]) treatment promotes NF-κB binding, triggering expression of costimulatory molecules and production of TGF-β and IL-6 (14, 15). In contrast, Wada et al. (16) showed that ATRA-instructed DC have altered expression of CD1, CD80, and CD86 together with a lower capacity to produce IL-12 and activate T cells. The role of LXRα in DC maturation is similarly conflicting. Although one report found that LXRα activation strongly interfered with LPS-induced maturation via suppression of CD40 surface expression, IL-12 release, and T cell stimulatory capacity (17), Töröcsik et al. (18) demonstrated that LXRα activation synergistically enhanced CD80 and CD86 expression, proinflammatory cytokine release, and T cell activation.
Because of the central importance of DC maturation in triggering immune responses, pathogens have developed mechanisms to interfere with this process. One such strategy is the secretion of microbial metabolites, such as quorum-sensing (QS) molecules. QS is an important communication mechanism by which microbes coordinate population density–dependent changes in behavior. In QS, microbial-secreted molecules from growing microorganisms accumulate in the local environment and modulate specific responses once a critical threshold or concentration has been reached (19). Although originally identified in bacteria, QS-like systems have also been described in eukaryotic microbes (20, 21). Farnesol (FOH) is a QS molecule produced by the opportunistic, human-pathogenic fungus Candida albicans and controls the yeast-to-hyphal transition, one of the central virulence traits in C. albicans (22). Although FOH inhibits filamentation, existing data suggest that FOH is likely also a virulence factor that promotes disease progression during systemic infections (23, 24). Indeed, treatment with a subinhibitory concentration of fluconazole was sufficient to induce FOH production in C. albicans. Additionally, administration of FOH in a murine model of disseminated C. albicans infection resulted in increased mortality and kidney fungal burden, whereas chemical treatment to inhibit FOH synthesis reduced these effects (23, 24).
Further evidence supports a role for FOH as an immune modulator. FOH treatment stimulates the production of chemokines and proinflammatory cytokines by macrophages in vitro, and these effects are further enhanced by the presence of fungal cell wall components (25, 26). In a previous study, we found that FOH treatment triggers the activation of neutrophils and monocytes without increasing fungal uptake and killing (27). In contrast, FOH dramatically impaired the differentiation of monocytes into DC by reducing the surface exposure of Ag-presenting and costimulatory molecules. These changes in surface markers were accompanied by diminished IL-12 and increased IL-10 release, correlating with an impaired capability to promote T cell proliferation (27). Interestingly, the lipid Ag-presenting molecules CD1a and CD1d also showed altered expression. Given that MAPK and NF-κB signaling pathways regulate the expression of costimulatory molecules and cytokine release during DC maturation (3–6, 10, 11) and that the expression of CD1 molecules in these cells is tightly regulated by the nuclear receptors PPARγ and RARα (2); in this study, we sought to elucidate the molecular mechanisms affected during the differentiation and maturation of DC in the presence of FOH, particularly with regards to the modulation of these signaling pathways. Our data show that whereas FOH promotes nuclear receptor activation, the immunophenotypic changes present during DC differentiation from monocytes in the presence of FOH are largely independent of nuclear receptors. As an exception, we found the increased CD1d expression on DC following FOH treatment to be mediated via the PPARγ and RARα pathways and partially through p38 MAPK activation. Nevertheless, this CD1d upregulation did not confer an enhancement in the capacity of DC to activate CD1d-restricted invariant NKT cells (iNKT). In contrast, FOH enhances the release of proinflammatory cytokines and anti-inflammatory IL-10, partially through MAPK and NF-κB signaling pathways while also reducing IL-12 secretion. Reconstitution of the IL-12/IL-10 milieu was sufficient to restore the capacity of DC to activate different T cells subsets.
Materials and Methods
Trans, trans-FOH was obtained as a 4 M stock solution (Sigma-Aldrich). This 4 M FOH stock solution was first diluted in 100% methanol, followed by subsequent dilutions in RPMI 1640 + 5% FBS to the working concentrations of 50 and 100 μM. The concentrations of FOH used in this study were selected based on the observations that C. albicans can release up to 55 μM FOH in in vitro culture and that concentrations of 10–250 μM can modulate the morphological transition of C. albicans (28, 29). The following nuclear receptors antagonists were obtained from Tocris Bioscience: GW9662 (PPARγ antagonist, 10 μM), AGN193109 (RARα antagonist, 10 μM), and GSK2033 (LXRα antagonist, 1 μM). Methanol was used as a solvent control and indicated mock treatment. LPS, peptidoglycan, flagellin (Sigma-Aldrich), and Pam3CSK4 (R&D Systems) were diluted in water before application. α-Galactosylceramide (α-GalCer) was obtained from Abcam and diluted in DMSO.
DC generation and maturation
Human monocytes were isolated from buffy coats of healthy volunteers (provided by the Transfusion Medicine Department of the Jena University Hospital). First, PBMC were obtained by density gradient centrifugation using Biocoll (Biochrom). Monocytes were positively selected from PBMC using CD14 MicroBeads coupled to a magnetic cell sorting system (Miltenyi Biotec). Freshly isolated monocytes were seeded into six-well plates at a density of 2 × 106 cells per well and cultured in RPMI 1640 (Biochrom) supplemented with 10% heat-inactivated FBS (Biochrom), 800 U/ml GM-CSF (Leukine [sargramostim]), and 1000 U/ml IL-4 (Miltenyi Biotec). Cells were cultured for 6 d at 37°C and 5% CO2, and medium was changed on day 3, accompanied by the addition of fresh GM-CSF and IL-4. Ligands or solvent controls were added to the cells starting from the first day. Nuclear receptor antagonists were added 1 h prior to FOH or mock treatment. The viability of immature DC (iDC) was not affected by these treatments (Supplemental Fig. 1). More than 85% of the starting monocytes differentiated into iDC, as determined by negative CD14 and positive CD1a staining. To obtain mDC, day 6 iDC were stimulated with 10 ng/ml LPS (Sigma-Aldrich) for 24 h. To evaluate TLR1/2, TLR2/6, and TLR5 activation, iDC were stimulated with Pam3CSK4 (10 μg/ml), peptidoglycan (10 μg/ml), or flagellin (100 ng/ml), respectively, for 24 h. To block NF-κB and MAPK activation, iDC were preincubated with the following inhibitors for 30 min prior to LPS stimulation: SC75741 (NF-κB inhibitor, 5 μM; Cayman), SP600125 (JNK inhibitor, 10 μM) and SB203580 (p38 inhibitor, 20 μM), both obtained from InvivoGen, and FR180204 (ERK1/2 inhibitor, 10 μM; Tocris Bioscience).
Phenotypic characterization of DC, iNKT, and T cells was performed by flow cytometry using fluorochrome-conjugated Abs. One hundred microliters of cell suspension was stained with the following Abs: mouse anti-human CD14/FITC, CD1a/PE, CD80/PE, CD40/FITC, CD1d/APC, CD3/PerCp, CD4/Pacific Blue, CD25/PE, CD85k (Ig-like transcript 3 [ILT3])/APC, CD85d (ILT4)/PE, CD274 (programmed cell death ligand 1 [PD-L1])/Brillian Violet 421, FOXP3/FITC, 6B11 (Vα24/Jα18)/FITC, and IFN-γ/FITC (BioLegend) as well as CD85j (ILT2)/FITC (BD Biosciences). After 20 min of staining, cells were collected by centrifugation (300 × g for 5 min) and resuspended in CellWASH (BD Biosciences). For determination of viability, iDC were harvested and stained with Fixable Viability Stain V450 (BD Biosciences) for 15 min prior to Ab staining. Isotype and fluorescence minus one were used for proper gating. Cells were evaluated on a FACSCanto II (BD Biosciences) and data analyzed using FlowJo 7.6.4 software.
Human iNKT cell expansion
mDC were treated with 100 ng/ml α-GalCer for 24 h to obtain α-GalCer–loaded mDC. Autologous T cells were collected by positive selection of CD3+ cells using CD3 MicroBeads (Miltenyi Biotec) and cocultured with α-GalCer–loaded mDC in a 10:1 ratio (1 × 106 T cells/1 × 105 mDC) in 24-well plates. After 24 h, 100 U/ml human rIL-2 (ImmunoTools) was added to the coculture. Following 7 d of coculture, iNKT expansion was quantified by the measurement of CD3+Vα24/Jα18+ cells by FACS. For cytokine reconstitution assays, human rIL-12 (10 ng/ml) (ImmunoTools) and a mouse anti-human IL-10–neutralizing Ab (10 μg/ml) or a mouse IgG2B isotype control (both obtained from R&D Systems) were added starting on the first day of coculture.
After 7 d of mDC/T cell coculture, cells were collected and restimulated for 6 h with PMA (50 ng/ml) and ionomycin (1 μg/ml) in the presence of brefeldin A (5 μg/ml) and monensin (5 μg/ml) (Sigma-Aldrich). Intracellular staining of IFN-γ was performed after fixation and permeabilization using the Intracellular Fixation and Permeabilization Buffer Set according to the manufacturer's protocol (eBioscience). For FOXP3 staining, the Foxp3/Transcription Factor Staining Buffer Set from eBioscience was used.
iDC were incubated in the presence of either LPS, Pam3CSK4, peptidoglycan, or flagellin for 24 h and then supernatants collected and stored at −80°C. The concentrations of secreted cytokines were analyzed using the ProcartaPlex Immunoassay from Thermo Fisher Scientific according to the manufacturer’s instructions.
RNA isolation and quantitative real-time PCR
Preparation of whole cell RNA was carried out with the RNeasy Plus Mini Kit (QIAGEN) according to the manufacturer’s instructions. The resulting RNA concentration was quantified via NanoDrop (Thermo Fisher Scientific). cDNA was generated from 25 ng of total RNA using the Precision nanoScript 2 Reverse Transcription Kit and quantitative RT-PCR (qRT-PCR) performed using the PrecisionPLUS Master Mix premixed with SYBR Green kit (PrimerDesign) and Quantitect Primer Assay oligonucleotides (QIAGEN) on a Stratagene Mx3500P cycler. Gene expression was normalized to the TBP reference gene and relative expression was calculated using the 2−ΔΔCT method.
Data were analyzed with GraphPad Prism software version 6. A Shapiro–Wilk normality test was used to assess Gaussian distribution. Comparison among groups was performed using a one-way ANOVA with Tukey posttest for multiple comparisons; otherwise, Kruskal–Wallis was used as a nonparametric test. A p value <0.05 was considered significant and marked in graphs using the following designations: *p < 0.05, **p < 0.01, and ***p < 0.001.
Differentiation of DC in the presence of FOH increases the expression of genes involved in nuclear receptor activity
Previous experiments showed that FOH treatment impairs DC differentiation from monocytes by altering surface molecule expression and the secretion of Th1-inducing cytokines (27). However, all the demonstrated effects of FOH on mDC thus far were determined following LPS-induced stimulation of TLR4. To get a broader view regarding the impact of FOH, we extended our analysis to stimulation of other TLRs. Similar to LPS-mediated maturation, cytokine secretion and surface expression of both CD80 and CD40 were altered in FOH-differentiated DC stimulated with either Pam3CSK4 (TLR1/2 agonist), peptidoglycan (TLR2/6 agonist), or flagellin (TLR5 agonist) (Supplemental Fig. 2), indicating that FOH-induced changes are independent of any specific TLR and that the effects of FOH are the result of downstream signaling cascades.
Interestingly, nuclear receptors agonists induce a similar immunophenotype as FOH in DC (13, 30). In agreement with this, pathway analysis of differentially expressed genes during differentiation of monocytes into iDC, and following LPS stimulation to generate mDC, identified numerous processes to be influenced by 50 μM FOH stimulation, including the PPAR signaling pathway (27). Several genes linked to nuclear receptor activation, including target genes of PPARγ, RARα, and LXRα were upregulated during DC differentiation in the presence of FOH (Fig. 1A). To confirm these data, we quantified gene expression levels of these nuclear receptors and their target genes by qRT-PCR (Fig. 1B). In accordance with previously published microarray data, FOH treatment increased the expression of PPARγ and its target gene FABP4, as well as of LXRα and its target gene APOC1. In addition, the target gene of RARα, TGM2, was upregulated, whereas gene expression of the receptor itself did not change as a result of FOH treatment. The effects of FOH on gene transcription were concentration dependent, and a higher expression of these genes was generally observed after treatment with 100 μM FOH. Because we detected no differences in viability (see Supplemental Fig. 1), subsequent experiments were performed using both 50 and 100 μM FOH concentrations. Taken together, our results show that DC generated in the presence of FOH have an altered transcriptional pattern, including increased expression of genes involved in nuclear receptor pathways.
Modulatory effects of FOH on DC surface molecule expression are partially mediated by nuclear receptor activation
To evaluate the role of nuclear receptor activation in the immunophenotypic changes induced by FOH, cells were exposed to various nuclear receptors antagonists during DC differentiation and analyzed for their surface expression of costimulatory and Ag-presenting molecules. One of the most prominent effects we observed in our previous study was in CD1 transmembrane glycoproteins (27). DC generated in the presence of FOH were characterized by markedly lower surface CD1a but increased expression of CD1d. None of the nuclear receptor antagonists tested in this study were sufficient to recover CD1a expression (Fig. 2A). In contrast, blocking the activity of PPARγ with receptor antagonist GW9662 in iDC differentiated in the presence of 50 and 100 μM FOH returned CD1d expression to basal levels (Fig. 2B; mock iDC, median of 266 ± 70; iDC with 50 μM FOH, median of 481 ± 145; iDC with 50 μM FOH and PPARγ antagonist, median of 265 ± 53). A similar return to basal CD1d expression levels DC was observed in the presence of RARα antagonist AGN193109 (mock iDC, median of 289 ± 71; iDC with 50 μM FOH, median of 474 ± 149; iDC with 50 μM FOH and RARα antagonist, median of 243 ± 64). These effects were also maintained during the maturation process from iDC to mDC, suggesting that the increased CD1d surface exposure induced by FOH is regulated by a common pathway that involves both nuclear receptors. However, blocking LXRα receptor activity using the antagonist GSK2033 downregulated CD1a (mock iDC, median of 12,731 ± 6598; with LXRα antagonist, median of 2845 ± 2469) and promoted the surface expression of CD1d (mock iDC, median of 259 ± 85; with LXRα antagonist, median of 581 ± 203) in mock-treated control cells, comparable to the effects induced by FOH (iDC with 50 μM FOH: CD1a, median of 255 ± 98; CD1d, median of 525 ± 227). The LXRα antagonist GSK2033 affected neither CD1a nor CD1d surface expression during FOH treatment.
FOH-differentiated DC showed reduced cell surface expression of the costimulatory molecules CD40 and CD80 in a concentration-dependent manner (Fig. 2C, 2D). However, blocking nuclear receptors did not prevent the FOH-induced effects on both CD40 and CD80 expression on iDC and mDC. Uniquely, the surface exposure of CD80 was restored in presence of the PPARγ antagonist at the low 50 μM concentration of FOH (iDC with 50 μM FOH: median of 276 ± 65; iDC with FOH and PPARγ antagonist: median of 454 ± 120) and reached levels similar to mock-treated iDC (median of 385 ± 140).
Altogether, our results confirmed alterations in the immunophenotype of DC differentiated in the presence of FOH. However, the majority of these effects could not be restored by treatment of FOH-instructed cells in concert with different nuclear receptor antagonists, indicating that FOH acts independent of the nuclear receptors PPARγ, RARα, or LXRα. The chief exception to this was the role for PPARγ and RARα activity in FOH-induced CD1d overexpression.
Despite enhanced CD1d levels, FOH-mDC have a reduced capacity to induce iNKT and FOXP3+ regulatory T cell expansion
In our previous work, we performed mixed allogeneic lymphocyte reactions and showed that FOH-treated mDC were significantly less potent at inducing T cell proliferation (27). However, CD1d is a MHC class 1–like molecule that presents lipid Ags to iNKT cells to activate them (13, 31). Therefore, we tested whether the elevated CD1d surface levels induced by FOH could specifically enhance the capability of mDC to drive iNKT expansion. For this, we cocultured autologous T cells with mDC for 7 d and then analyzed iNKT expansion by flow cytometry. Surprisingly, mDC differentiated in the presence of FOH had a reduced ability to induce iNKT expansion (50 μM FOH: fold change of 2.5 ± 1.2) compared with the mock-treated cells (fold change of 7.3 ± 1.7), despite their significantly increased CD1d levels (Fig. 3A). Although the CD1d expression induced by FOH was dependent on both PPARγ and RARα activity, blocking these nuclear receptors in FOH-treated mDC-using antagonists resulted in comparably lower iNKT expansion.
We further addressed if FOH could promote a DC shift toward a tolerogenic phenotype by analyzing the expression of inhibitory receptors, including ILTs (also called leukocyte Ig-like receptors) ILT2, ILT3, ILT4, and PD-L1 on mDC under the influence of FOH. Stimulation of these receptors alters the cytokine secretion profile of DC while activating natural FOXP3+ regulatory T cells (32). Although FOH-instructed mDC had significantly elevated surface levels of ILT3 and PD-L1, regulatory T cell expansion was reduced (Fig. 3B, 3C), indicating that FOH treatment leads to a paralysis of DC despite the upregulation of immunomodulatory surface molecules.
FOH modulates cytokine release by TLR4-treated DC partially through nuclear receptors, NF-κB, and MAPK activation
The cytokine milieu generated by DC is crucial for mediating T cell activation and instructing the type of T cell response. Previously published data show that FOH induces elevated levels of proinflammatory cytokines (IL-8 and TNF-α) and IL-10, as well as reduced secretion of IL-12 (27). Thus, we tested the contribution of nuclear receptor activity to these immunophenotypic changes. Although none of the nuclear receptor antagonists tested prevented the FOH-induced increase in IL-8 (Fig. 4A), the presence of the LXRα antagonist GSK2033 blocked the increased release of TNF-α by FOH-treated mDC (50 μM FOH: without LXRα antagonist, 8.7 ± 3.5 ng/ml; with LXRα antagonist, 3.0 ± 1.4 ng/ml) and restored TNF-α levels to those observed in mock-treated samples (4.4 ± 1.0 ng/ml) (Fig. 4B). In contrast, the RARα antagonist AGN193109 reduced the LPS-induced release of TNF-α in general and independent of FOH under all tested conditions. Treatment of mDC with either the PPARγ, RARα, or LXRα antagonists in the absence of FOH resulted in elevated IL-10 levels, which were synergistically increased in the presence of 50 μM FOH (Fig. 4C). However, the strong IL-10 release induced by the higher 100 μM concentration of FOH was not further enhanced through the blockade of nuclear receptor activity. Similarly, the profound FOH-induced reduction in IL-12 was independent of nuclear receptors. Treatment with the respective antagonists was not sufficient to restore the secretion of IL-12 by mDC (Fig. 4D). Treatment with the LXRα antagonist alone inhibited IL-12 secretion in mock-treated mDC and during exposure to the lower FOH concentration.
To obtain further insight into the molecular mechanisms that regulate altered cytokine release by FOH-differentiated DC, we evaluated the role of MAPK and NF-κB signaling. Both pathways are known regulators of cytokine release in response to TLR stimulation (6, 10, 11, 33) (Fig. 5A). Although the release of IL-8 was not blocked by any of the inhibitors tested, the elevated TNF-α levels observed in FOH-differentiated DC were reduced in the presence of the NF-κB and JNK inhibitors (100 μM FOH: without inhibitors, 15.8 ± 2.5 ng/ml; with NF-κB inhibitor, 5.9 ± 2.9 ng/ml; with JNK inhibitor, 3.4 ± 2.7 ng/ml), reaching levels below mock-treated mDC (without inhibitor, 9.1 ± 3.1 ng/ml). Likewise, we found the secretion of IL-10 to be dependent on NF-κB and p38 MAPK activation in every condition tested, indicating that the increased secretion of this anti-inflammatory cytokine in both the presence and absence of FOH was regulated by these two signaling pathways. Both NF-κB and p38 MAPK signaling pathways also regulated the release of IL-12 in mock-treated mDC (without inhibitor, 3217 ± 1379 pg/ml; with NF-κB inhibitor, 1206 ± 796 pg/ml; with p38 inhibitor, 61 ± 24 pg/ml). In contrast, during FOH treatment, we observed only a synergistic reduction in the presence of the p38 inhibitor (50 μM FOH: without p38 inhibitor, 402 ± 221 pg/ml; with p38 inhibitor, 27 ± 13 pg/ml) and no reversion to the reduced IL-12 release by FOH-instructed mDC.
We next examined the impact of these signaling pathways on the cell surface expression of Ag-presenting and costimulatory molecules. The reduced expression of CD1a in FOH-differentiated mDC was not restored when MAPK and NF-κB were inhibited during LPS stimulation (Fig. 5B). Interestingly, the CD1d overexpression induced by 100 μM FOH was partially blocked in the presence of the p38 MAPK inhibitor (100 μM FOH: without p38 inhibitor, median of 473 ± 84; with p38 inhibitor, median of 341 ± 52, p: 0.07), suggesting an interaction between p38 MAPK– and PPARγ/RARα-signaling pathways. As expected, the expression of CD80 in mock-treated cells after LPS stimulation was blocked in the presence of NF-κB and p38 inhibitors (mock-treated: without inhibitor, median of 1102 ± 165; with NF-κB inhibitor, median of 778 ± 171; with p38 inhibitor, median of 645 ± 100). Furthermore, we observed a synergistic reduction in CD80 expression on FOH-differentiated mDC (50 μM FOH: without inhibitor, median of 810 ± 168; with NF-κB inhibitor, median of 559 ± 150; with p38 inhibitor, median of 409 ± 86). Finally, the presence of the NF-κB inhibitor blocked the expression of CD40 following LPS stimulation in mock-treated and FOH-differentiated mDC (mock-treated: without inhibitor, median of 15,050 ± 1706; with NF-κB inhibitor, median of 9729 ± 3482).
Our results confirm the substantial impact of FOH on the cytokine profile released by mDC that also serves to promote inflammation while dampening Th1 responses. However, these effects are largely independent of nuclear receptor activity, especially with regard to changes in anti-inflammatory IL-10 and Th1-inducing IL-12 levels. In contrast, the enhanced secretion of proinflammatory TNF-α and anti-inflammatory IL-10 in the presence of FOH is dependent on NF-κB, JNK, and p38 MAPK activation. Interestingly, we found that CD1d overexpression induced by FOH during LPS stimulation might require p38 MAPK signaling pathway activation.
Altered secretion of IL-12 and IL-10 influences the capacity of FOH-differentiated DC to activate different T cell subsets
Although FOH induces an increase in CD1d surface exposure, we could not detect an effect on iNKT cell expansion. However, our results also show that FOH simultaneously promotes IL-10 secretion while diminishing IL-12 release. Because of their potent T cell–activating effects, the IL-12 family of cytokines is an important link between innate and adaptive immunity. However, IL-10 is known to inhibit the production of IL-12 by macrophages and DC (34). Therefore, we hypothesized that the observed shift in the ratio between IL-12 and IL-10 in FOH-differentiated DC might play a role in iNKT activation (Fig. 6A). To test our hypothesis, we performed T cell/mDC coculture experiments and modified the cytokine environment via the addition of rIL-12 and an IL-10 neutralizing Ab. Interestingly, the reduced iNKT expansion seen in the presence of FOH-differentiated mDC was significantly enhanced after the addition of exogenous IL-12 and the IL-10–blocking Ab (50 μM FOH: without IL-12/αIL-10, fold change of 4.7 ± 2.9; with IL-12/αIL-10, fold change of 12.1 ± 4.8). Even under mock-treated conditions, we could detect a stimulatory effect of the higher IL-12/IL-10 ratio on iNKT expansion (fold change of 15.4 ± 3.8) compared with the control T cell/mDC coculture (fold change of 10.1 ± 3.7) (Fig. 6B).
The reduced potential of FOH-treated mDC in inducing T cell proliferation is the result of altered IL-12 and IL-10 release (27). In line with the results for iNKT cell activation, the percentage of IFN-γ–producing T cells was higher when T cells were cocultured with mock-treated mDC compared with FOH-differentiated mDC, and the presence of exogenous IL-12 and IL-10 blockade was able to restore this function (Fig. 7). In agreement with this, the expansion of FOXP3+ regulatory T cells was also dependent on the altered IL-12/IL-10 ratio, despite the changes in mDC surface marker expression induced by FOH (Fig. 3D).
Human DC play a central role in inducing and directing immune responses against invading pathogens. Consequently, human-pathogenic microorganisms, including viruses, bacteria, fungi, and parasites, have developed ways to modulate DC function during infection (35, 36). In addition to contact-dependent mechanisms, microorganisms can actively release soluble mediators, such as QS molecules, to modify host immune responses (27, 37–40). For example, the bacterial QS molecule homoserine lactone (HSL) modulates DC maturation (37). Mechanistically, HSL blocks expression of the costimulatory molecules CD40 and CD80 and the Ag-presenting molecule HLA-DR on DC. Furthermore, HSL-instructed DC show reduced release of IL-12 while promoting immunosuppression via increased production of anti-inflammatory IL-10 and selective activation of FOXP3+ regulatory T cells (37). Similarly, we have shown that the fungal QS molecule FOH, produced by C. albicans, impairs the differentiation from monocytes into DC (27). As a result of this, FOH-instructed DC have a reduced capacity to induce T cell proliferation. Nevertheless, the molecular mechanisms by which FOH modulates DC differentiation and maturation remained poorly understood. In the current study, we identify several molecular pathways that are modulated by FOH during DC differentiation and maturation, supporting its role as a virulence factor.
FOH markedly alters the transcriptional response, surface phenotype, and cytokine release of DC during differentiation and after maturation, and some of these effects are regulated through nuclear receptor activity. Nuclear receptors modulate macrophage and DC functionality by regulating the release of proinflammatory cytokines and the surface exposure of Ag-presenting and costimulatory molecules (2). Indeed, the release of proinflammatory TNF-α induced by FOH is regulated by LXRα and RARα. Szanto et al. (41) showed that RARα activation by natural and synthetic ligands promotes the expression of genes such as CYP27A1, an enzyme involved in ligand production of LXRα. Activation of LXRα also activates NF-κB, which enhances the secretion of proinflammatory TNF-α (18). Thus, our results suggest that FOH modulates the release of TNF-α via a coordinated activation of RARα and LXRα, and involving both NF-κB and JNK signaling pathways.
One of the most prominent effects induced by FOH treatment was observed for CD1 expression. CD1a and CD1d are MHC-like molecules that present lipid Ags to T cells and are important for priming the immune response against pathogenic bacteria and fungi (42, 43). Although CD1a expression was reduced by FOH and not restored in the presence of nuclear receptor antagonists, CD1d upregulation by FOH was blocked by PPARγ and RARα antagonists, suggesting that CD1d expression is regulated by these nuclear receptors and epistatic to the effects of FOH. Szatmari et al. (30) showed that the activation of PPARγ leads to an upregulation of retinaldehyde dehydrogenases (RALDH; ALDH1), catalyzing the production of ATRA. ATRA itself modulates CD1d expression via the activation of RARα in human DC. Concurrent with the higher CD1d surface exposure on FOH-differentiated DC, we also observed increased gene expression of ALDH1A1. In agreement with our data, FOH has been shown to activate nuclear receptors in various cell types (44–48). Shchepin et al. (49) showed that FOH can accumulate in the nuclei of C. albicans, potentially interacting directly with nuclear receptors. Our results indicate that, in addition to a direct interaction, FOH could be modulating nuclear receptors activity via regulation of p38 MAPK–signaling pathway. Indeed, inhibition of p38 MAPK during the maturation process partially reversed the upregulation of CD1d molecules induced by higher FOH concentrations. In line with our findings, other studies have shown that PPARγ transcriptional activity is modulated by p38 MAPK phosphorylation (50, 51). Taken together, our results indicate that FOH regulates CD1d expression via activation of PPARγ and RARα nuclear receptors in a process partially dependent on p38 MAPK signaling.
CD1d presents lipid Ags to iNKT, and the upregulation of this molecule is associated with an enhanced ability to activate these cells (13). However, we failed to observe higher iNKT expansion in the presence of FOH-differentiated DC. In concert with the signals triggered by CD1d-mediated lipid presentation, costimulatory molecule interactions, such as CD40/CD40L and CD28/CD80/CD86, and the release of IL-12 by DC are required for optimal iNKT activation (52, 53). Indeed, treatment with FOH resulted in reduced CD40 and CD80 surface expression on DC, as well as a shift in IL-12/IL-10 secretion independent of nuclear receptor pathways. We found that IL-10 secretion was dependent on the activation of NF-κB and p38 MAPK. Although IL-10 is generally able to suppress the release of IL-12 in cells stimulated with LPS (54, 55), we did not observe reconstitution of IL-12 secretion when IL-10 release was blocked, suggesting an alternative mechanism in the regulation of IL-12 production by FOH.
The altered secretion of IL-12 and IL-10 by DC differentiated in the presence of FOH plays a key role in their capacity to induce iNKT expansion. Reconstitution of IL-12 and IL-10 levels to those observed in the absence of FOH resulted in increased expansion of iNKT cells. Therefore, the paralytic effects of FOH on DC are driven primarily by the altered secretion of these cytokines, rather than the expression of surface markers. These effects may influence the outcome of infection, as CD1d-restricted iNKT cells are important during the immune response to various pathogens (42, 56, 57). Cohen et al. (42) showed that iNKT activation is important for controlling C. albicans infection, in a process dependent on IL-12 secretion by DC. Furthermore, iNKT cells are activated after recognition of α-manosyl residues on C. albicans in a CD1d-dependent manner, and this activation increases the survival of mice coinfected with Streptococcus pneumoniae (58). Moreover, we observed that altered secretion of IL-12 and IL-10 by FOH-differentiated DC also regulates their capacity to activate Th1 and FOXP3+ regulatory T cells, both of which play an important role in anti-Candida immunity by modulating macrophages and Th17 cells activation (59–61).
In summary, our data provide new insight into the molecular mechanisms by which FOH modulates DC function. FOH affects DC differentiation and function by modifying multiple, diverse signaling events during DC maturation. Among the signaling cascades modulated by FOH are MAPK, NF-κB, and nuclear receptors, cumulatively resulting in the modified expression of key surface molecules and cytokine release. However, despite individual effects of FOH that could, in isolation, suggest activation of DC, the net effect of FOH exposure on DC is paralysis. FOH-exposed DC are unable to respond properly to TLR activation, and the altered secretion of IL-12 and IL-10 impairs their capacity to activate iNKT, Th1, and FOXP3+ regulatory T cells. Because of the key role these cell types play in antifungal immunity, FOH can be viewed as a virulence factor enabling C. albicans to overcome immune surveillance by DC.
We are grateful to Amelia Barber for critical reading of the manuscript and to all the anonymous blood donors who contributed to this study.
This work was supported by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center CRC124 FungiNet (Project C3) and by the Excellence Graduate School, Jena School for Microbial Communication (Jena, Germany).
The online version of this article contains supplemental material.
Abbreviations used in this article:
all-trans retinoic acid
Ig-like transcript 3
liver X receptor α
programmed cell death ligand 1
peroxisome proliferator activated receptor γ
retinoic acid receptor α.
The authors have no financial conflicts of interest.