Ceramide Inhibits Antigen Uptake and Presentation by Dendritic Cells

PBMCs obtained by standard Ficoll–Paque method (Organon Teknika, Durham, NC) were separated on multistep Percoll gradients (Pharmacia Fine Chemicals, Uppsala, Sweden) and the light density fraction from the 42.5–50% interface was recovered and depleted of CD19⁺ and CD2⁺ cells using magnetic beads coated with specific antibodies (Dynal, Oslo, Norway). The remaining cells were cultured in RPMI-1640 supplemented with 2 mM l-glutamine, 1% nonessential aminoacids, 1% pyruvate, 50 μg/ml kanamicin, 5 × 10⁻⁵ M 2-ME (GIBCO BRL, Gaithersburg, MD) + 10% FCS (Hyclone Laboratories, Inc., Logan, UT) in the presence of 50 ng/ml GM–CSF and 1000 U/ml IL-4 (provided by Dr. A. Lanzavecchia, Basel Institute for Immunology, Switzerland). Cultured DCs were routinely >90% CD1⁺, HLA-DR⁺, CD14⁻, and were used after 5–7 d of culture.

After stimulation, lipids were extracted and then incubated with Escherichia coli diacylglycerol kinase. Ceramide phosphate was then isolated by TLC using CHCl₃/CH₃OH/CH₃COOH (65:15:5) as solvent (13, 14). Authentic ceramide-1-phosphate was identified by autoradiography at Rf 0.25. Quantitative results for ceramide production are expressed as pmol ceramide-1-phosphate/10⁶ cells.

2 × 10⁵ DCs were resuspended in 200 μl RPMI buffered with 25 mM Hepes + 10% FCS. C2-ceramide and C2-dihydroceramide (Sigma Immunochemicals, St. Louis, MO) were reconstituted in EtOH at 15 mM and stored at −20°C until use. Diacylglycerol was purchased from Amersham (Buckingham, England). Lucifer yellow (LY), FITC–dextran (DX) (Molecular Probes, Inc., Eugene, OR), or HRP (Sigma Immunochemicals) were reconstituted in PBS, stored at 4°C, and spun in a microfuge before use to eliminate aggregates. To quantify LY and FITC–DX, cells were washed four times with cold PBS containing 1% FCS and 0.01% NaN₃ and analyzed on a FACScan^® (Becton Dickinson, Mountain View, CA), using propidium iodide to exclude dead cells. For horseradish peroxidase (HRP) quantification, cells were washed four times as above then four times with PBS alone with one tube change, lysed with 0.05% Triton X-100 in 10 mM Tris buffer pH 7.4 for 30 min, and the enzyme activity of the lysate was measured using O-phenilendiamine and H₂O₂ as substrates with reference to a standard curve. Tetanus toxoid (TT)-specific T cell clones KS140 and KB24 and TT peptide P2 (residues 830–843) were provided by Dr. A. Lanzavecchia. TT antigen was purchased from Connaught (Ontario, Canada).

For antigen presentation assays, 4 × 10⁴ T cells were cultured with 10⁴ irradiated DCs in 200 μl RPMI with 10% FCS in flatbottomed microplates. [³H]thymidine incorporation was measured at day 2. For MLR, 1.5 × 10⁵ responding cells from allogeneic adult PBMCs were cultured with different numbers of irradiated DCs. [³H]thymidine incorporation was measured at day 5. Parietaria judaica pollen (Allergon, Angelholm, Sweden) was extracted with bicarbonate buffer 0.125 M pH 8. PjE-specific clone P6.2 was isolated from an atopic patient (15).

2 × 10⁵ DCs were treated in 200 μl RPMI 10% FCS with 80 μM C2-ceramide for 10 min on ice, then for 1 h at 37°C. After incubation, cells were washed and left in culture for 48 h. Cells were then recovered and processed for propidium iodide staining and FACS^® analysis as previously described (13).

Supernatant from J558L cells stably expressing a chimeric mCD40L–mCD8α construct, provided by Dr. P. Lane, (Basel Institute for Immunology) was used as a source of CD40L. Recombinant human TNF-α with R32W and S86T substitutions (TNF-R1 specific) and TNF-α with D143N and A145R substitutions (TNF-R2 specific) were provided by Drs. W. Lesslauer and H. Loetscher (Hoffman La Roche, Ltd., Basel, Switzerland). IL-1β was provided by Dr. L. Melli (IRIS, Siena, Italy).

IL-1β and TNF-α have been shown to induce transient ceramide accumulation in tumor cell lines (16, 17). It was not known whether CD40, like other TNF receptor family members such as TNF-R1 p55, Fas/APO-1, or NGF-R p75 (13, 16, 18), also signaled through ceramide generation. Therefore, we investigated whether cross-linking of CD40 was able to induce ceramide accumulation in cultured immature DCs. Fig. 1 shows that CD40L, as well as IL-1β and TNF-α, engaging TNF-R1 p55 but not TNF-R2 p75, were all potent inducers of ceramide generation in cultured DCs. Because CD40L, IL-1β, and TNF-α trigger in vitro maturation of DCs (5), as does LPS, which is structurally analogous to ceramide itself (11), these results raised the possibility that a common ceramide-mediated pathway, mimicked by LPS, could be responsible for some of the functional changes observed during in vitro maturation of DCs.

To test this hypothesis directly, we investigated whether exposure to exogenous cell-permeant C2-ceramide could down-modulate DC antigen uptake ability. DCs capture antigen either via macropinocytosis, a cytoskeleton-dependent type of fluid phase endocytosis initiated by membrane ruffling and formation of large vesicles, or receptor-mediated endocytosis through Fcγ and mannose receptors (5). As shown in Fig. 2, C2-ceramide could inhibit the uptake of three different classical endocytosis markers and their time-dependent accumulation into DCs. Both macropinocytosis, as assessed by LY and FITC-DX, and receptormediated endocytosis, as assessed by limiting amounts of HRP, were significantly affected. Comparable results were also obtained using C6-ceramide, a longer acyl chain ceramide analogue (data not shown). By contrast, C2-dihydroceramide, a structural analogue of C2-ceramide that lacks a double bond at the 4–5 position in the sphingoid base, was ineffective. Similarly, exposure to other diffusible signaltransducing lipid mediators such as diacylglycerol, did not affect macromolecule uptake ability of DCs (Fig. 2, A, C, E).

We then tested whether the endogenous production of ceramide would result in a similar inhibition of the endocytic ability of DCs. Exposure of cultured DCs to exogenous sphingomyelinase, which results in intracellular ceramide accumulation (data not shown), also induced a dose-dependent inhibition of HRP uptake (Fig. 3,A) and substantially retarded its time-dependent accumulation (Fig. 3 B). Taken together, these results indicated that ceramide could specifically mediate inhibition of macromolecules uptake by DCs.

Cultured DCs are extremely efficient at presenting soluble antigen to specific T cells (2). In vitro maturation of DCs promoted by short term exposure to TNF-α results in a severalfold decrease of the antigen presentation capacity, associated with an increase in T cell stimulatory ability (4). Therefore, we tested whether ceramide could be sufficient for effectively modulating antigen presentation to T cells by using two different soluble antigens, TT and a soluble extract of P. judaica pollen (PjE). Cultured DCs were exposed to C2-ceramide and then pulsed with TT or PjE, before being used to challenge antigen-specific T cell clones (15, 19). Fig. 4 shows that C2-ceramide induced a ∼50fold reduction in the ability of DCs to present PjE, and ∼100-fold reduction in the ability to present TT to their respective T cell clones (Fig. 4, A and B). By contrast, DCs treated with C2-ceramide were at least as efficient as untreated DCs in presenting nonprocessed antigen, i.e., in presenting an immunogenic TT peptide to the same TTspecific T cell clone (Fig 4,C). Ceramide analogue C2-dihydroceramide was ineffective in blocking the response to soluble antigens (Figs. 4, A, B, and C).

C2-ceramide is known to induce apoptotic cell death when administered to hemopoietic tumor cell lines or to in vivoactivated primary lymphoid cells within 6–12 h (20–22). Therefore, we checked whether the observed changes in antigen-processing capacity were due to loss of cell viability. C2-ceramide–treated DCs cultured for as long as 48 h excluded Trypan blue, displayed normal morphology, and did not show any DNA fragmentation by propidium iodide staining and FACS^® analysis (Fig. 4, D and E). Moreover, C2-ceramide treatment did not affect the ability of DCs to stimulate allogeneic T cells (Fig. 4 F).

Finally, we investigated whether the endogenous production of ceramide would affect the ability of DCs to present soluble antigen to T cells. Cultured DCs were treated with exogenous sphingomyelinase before being pulsed with TT, or with a TT peptide, and used to challenge a TT-specific T cell clone. Fig. 5 shows that endogenous ceramide production almost completely prevented presentation of soluble TT antigen, but had no inhibitory effect on TT peptide presentation by DCs, to the same TT-specific T cell clone.

In this paper, we provide evidence that ceramides inhibit the antigen-capturing ability of cultured DCs, thereby suggesting a common molecular basis for CD40L, TNF-α, and IL-1β, or bacterial products such as LPS, to downmodulate antigen presentation by professional APC (5). In fact, we show that CD40L, as well as TNF-α or IL-1β, were all strong inducers of ceramide accumulation in DCs. Ceramides may specifically control antigen capturing and processing by DCs, as other cytokine-mediated differentiation events, i.e., upregulation of LFA1, B7-1, ICAM-1, and MHC molecules, were not consistently affected by ceramide exposure (data not shown). Accordingly, the enhanced immunostimulatory ability of mature DCs could not be promoted by exogenous ceramides, suggesting that additional intracellular mediators participate in the maturative process. Importantly, specific immunoefficiency of DCs can be inhibited without affecting cell viability or the ability to present nonprocessed antigen.

A possible explanation for these findings may reside in the capacity of endogenously released ceramides to interfere with vesicular trafficking. In fact, ceramides have been shown to directly inhibit endocytosis (23) and glycoprotein transport through the Golgi complex in CHO cells (24). Perturbing anterograde transport through the Golgi may prevent newly synthesized MHC class II molecules to reach endosomal compartments to be loaded with peptides derived from hydrolyzed antigen. Interestingly, the fungal antibiotic brefeldin A (BFA), a classic inhibitor of both endogenous and exogenous antigen processing and presentation (25, 26), which causes disassembly of the Golgi apparatus (27) and its fusion with the ER and with early endosomes (28, 29), also triggers sphingomyelin hydrolysis resulting in ceramide production (30). Therefore, it is likely that the capacity of BFA to modulate antigen presentation is mediated by endogenously released ceramide.

Ceramides are emerging as intramembrane messengers involved in a variety of cellular adaptive and differentiative responses. Here, we provide evidence for a novel important function of ceramides in highly specialized cells such as DCs, which is modulation of soluble antigen presentation. Moreover, our data suggest that the capacity of ceramides to perturb intracellular membrane trafficking may be exploited by extracellular ligands able to trigger sphingomyelin hydrolysis, or by bacterial products that mimic ceramide, such as LPS, in order to regulate professional APC function and antigen-specific immune responses.

We thank Drs. A. Lanzavecchia, P. Lane, W. Lesslauer, and H. Loetscher for reagents.