DC-SIGN-Fc is composed of the extracellular domains of DC-SIGN (amino acid residues 64–404) fused to the Fc domain of human IgG1 (Geijtenbeek et al., 2002). DC-SIGN-Fc was produced in CHO cells and purified with a Hi Trap Protein A HP column (GE Healthcare). The following antibodies were used: α-DC-SIGN [clone AZN-D1 (Geijtenbeek et al., 2000b)], α-dendritic cell immunoreceptor [DCIR, clone 111F8.04 (Dendritics)], and α-mannose receptor [MR, clone 19.2 (BD Bioscience)]. Stainings were performed using 5 μg/ml of primary antibody and 5 μg/ml of AlexaFluor 488 F(ab′)₂ fragment of goat anti-mouse IgG (Invitrogen) as secondary antibody.

PSA was purified from B. fragilis as previously described (Cobb et al., 2004). Labeling was performed by oxidation of 20% of the galactofuranose side chain sugar. AlexaFluor 488 (AF488) was added via the aldehyde group (Cobb et al., 2004). Sp1 was prepared as described previously (Tzianabos et al., 2000b).

Raji and Raji-DC-SIGN cells (Geijtenbeek et al., 2000a) were cultured in RPMI1640 (Invitrogen) supplemented with 10% FBS (BioWhittaker), 1000 U/ml penicillin/streptomycin (Lonza), and 2 mM glutamine (Lonza). Monocytes were isolated from PBMCs from buffy coats of healthy donors (Sanquin) by a lymphoprep gradient (Axis-Shield) and a subsequent percoll gradient centrifugation (Amersham). DCs were generated by culturing purified monocytes in RPMI1640 supplemented with 10% FBS, 1000 U/ml penicillin/streptomycin, and 2 mM glutamine in combination with IL-4 (262.5 U/ml; Biosource) and GM-CSF (112.5 U/ml; Biosource) for 4–7 days. CD4⁺ T cells were isolated from the PBL fraction of buffy coats of healthy donors via negative selection using MACS beads (Milteny Biotec) according to the manufacturer’s instructions.

PSA-AF488 was incubated with DCs for 3 h at 37°C at the indicated concentrations in PBS containing calcium. For blocking experiments, 30 μg/ml PSA-AF488 was incubated with DCs for 45 min at 37°C in the presence of blocking agents and monitored by flow cytometry (FACScan, BD Biosciences). All flow cytometric analysis was performed with Flowjo software (Tree star). DC-SIGN expressing Raji cells were incubated with the indicated concentrations of PSA-AF488 for 3 h at 37°C. Unbound PSA was washed away and the amount of bound/internalized PSA was measured by flow cytometry. CLRs were blocked with a final concentration of 20 μg/ml of CLR-specific blocking antibodies, 50 mM monosaccharides [α-N-acetyl-d-glucosamine, α-l-Fucose, d-galactose, d-glucose, N-acetyl-d-glucosamine (Sigma Aldrich)], 25 μg/ml mannan (Sigma Aldrich), 25 μg/ml Laminarin (Sigma Aldrich), 25 μg/ml mannose-BSA (Sigma Aldrich), or 10 mM EGTA for 30 min at 37°C prior to the PSA binding assay.

PSA-AF488 was incubated with DCs for 3 h at 37°C at the indicated concentrations in PBS buffer containing calcium. As a positive and negative control, cells were incubated with a targeting antibody against DC-SIGN (AZN-D1) for 3 h at either 37 or 4°C respectively. Cells were washed in ice-cold PBS, fixated in 4% paraformaldehyde (Electron Microscopy Sciences), and measured by imaging flow cytometry (AMNIS Inc.) as previously described (García-Vallejo et al., 2012). Cells that have internalized antigen typically have positive internalization scores while cells that show the antigen still on the membrane have negative scores. Cells with scores around 0 have similar amounts of antigen on the membrane and in intracellular compartments.

Ninety six-well flat-bottomed ELISA plates (Maxisorp, Nunc) were coated with 10 μg/ml or indicated concentrations of unlabeled PSA or Sp1 in PBS or Lewis^X-polyacrylamide conjugate (PAA, Lectinity) in coating buffer (50 mM Na₂CO₃, pH 9.7) overnight at room temperature. DC-SIGN-Fc binding was measured as previously described (García-Vallejo et al., 2012).

Irradiated (3500 rad) DCs were pre-incubated in the presence or absence of 20 μg/ml AZN-D1 for 30 min at 37°C. PSA was added at a concentration of 30 μg/ml and incubated for 3 h at 37°C. Unbound PSA and AZN-D1 were washed away and DCs were incubated with autologous CD4⁺ T cells for 6 days at a ratio of 1:10 in combination with 100 μM of the NO substrate glyco-SNAP-2 (Calbiochem). PSA binding and uptake and the blocking effect of α-DC-SIGN were monitored by simultaneously incubation of DCs with PSA-AF488 and measured by flow cytometry. Proliferation was measured using [³H]-thymidine incorporation (1 μCi/well; Amersham Bioscience) for 16 h. Cells were harvested onto filters and the [³H]-thymidine incorporation was measured with the use of a beta counter (Perkin Elmer).

P values were calculated with a Student’s t-test. P values < 0.05 were considered to be significant.

Although the unique feature of PSA to activate specific CD4⁺ T cell responses in humans has long been demonstrated (Tzianabos et al., 2000b), the molecular mechanisms leading to recognition and internalization, a crucial step in the processing and loading of PSA onto MHC class II, are still unclear. Since PSA is a complex carbohydrate and CLRs on DCs are involved in the recognition of glycans, CLRs appear to be the prime candidates for the recognition of PSA by APCs. In addition, several CLRs are antigen-uptake receptors for presentation on MHC class II or MHC class I (Sancho and Reis e Sousa, 2012). Therefore, we first analyzed whether DCs, known to express a wide variety of CLRs, were able to bind PSA. Our results show that human DCs efficiently bound PSA in a dose-dependent manner (Figure 1A). Although DCs exposed to PSA for a prolonged period of time at 37°C are expected to internalize any bound material, we could not deduce this from classical flow cytometry data. We therefore measured PSA-incubated DCs using imaging flow cytometry. As a positive and negative control we treated DCs with a monoclonal antibody against DC-SIGN, known to induce DC-SIGN internalization (Engering et al., 2002a), at 37 or 4°C, respectively. Triggering of DC-SIGN induced internalization only at 37°C. The signal corresponding to PSA perfectly overlapped with that of the positive control, indicating that PSA was efficiently internalized under these conditions (Figure 1B). To determine whether CLRs play a role in the binding/internalization of PSA we analyzed binding/internalization by flow cytometry in the presence of EGTA, a Ca²⁺ chelator. Results indicate that PSA binding/internalization by DCs was indeed Ca²⁺-dependent, a typical characteristic of CLRs (Figure 1C). In order to elucidate the identity of the CLR(s) involved in the recognition of PSA, we incubated DCs with PSA in the presence of different mono and polysaccharides as competitive inhibitors for CLRs. Binding/internalization of PSA could be blocked in the presence of fucose and mannan (Figure 1D), in concentrations previously shown to block DC-SIGN-dependent ligand binding (Curtis et al., 1992). In addition, high concentrations of glucose and N-acetyl-d-glucosamine (GlcNAc) could also block binding/internalization of PSA, although to a lesser extent (Curtis et al., 1992; Lee et al., 2011). Altogether these results indicate the involvement of a mannose/fucose-specific CLR, such as DC-SIGN or MR. Other mono and polysaccharides, like laminarin (a dectin-1-specific glycan) and GalNAc (a macrophage galactose-type lectin-specific sugar) did not alter PSA binding/internalization (Figure 1D). Also, Mannose-BSA (a MR-specific ligand) did not inhibit PSA binding/internalization (Figure 1D), suggesting that MR is not involved in the recognition of PSA by DCs. Therefore, we concluded that the CLR DC-SIGN is a likely PSA receptor on DCs.

To examine whether DC-SIGN directly binds to PSA, we used a DC-SIGN-Fc chimeric protein consisting of the extracellular domains, including the carbohydrate recognition domain of DC-SIGN fused to the human IgG1 Fc tail. We compared DC-SIGN-binding to PSA to the well-known DC-SIGN ligand Lewis^X. Concentration-dependent binding of DC-SIGN-Fc to both PSA and Lewis^X was observed and the interaction appeared to be Ca²⁺-dependent, since it could be blocked by the addition of EGTA (Figure 2A). To confirm the binding of PSA at the cellular level, we used a Raji cell line transduced with DC-SIGN (Geijtenbeek et al., 2000a). Raji-DC-SIGN cells expressed high levels of DC-SIGN, which was absent on the parental Raji cells (Figure 2B). PSA bound the Raji-DC-SIGN cells in a concentration-dependent manner, which could be blocked by EGTA to the level of the parental Raji cells (Figure 2C). Several ZPSs have similar functional characteristics as PSA, including the induction of CD4⁺ T cell proliferation and the development of peritoneal abscesses (Velez et al., 2009). One of these ZPSs is Sp1, the capsular polysaccharide from type 1 Streptococcus pneumoniae. Although the biological properties of Sp1 mimic those of PSA, Sp1 was not recognized by DC-SIGN (Figure 2D), suggesting that DC-SIGN is a PSA-specific receptor and not a general binding partner for all ZPSs.

Since we identified DC-SIGN as a specific PSA receptor, we continued our research to determine the contribution of DC-SIGN in the presentation of PSA by DCs to CD4⁺ T cells. DCs express different CLRs, such as DCIR and MR, that share overlapping glycan specificity with DC-SIGN (Figure 3A). To rule out a possible involvement of these CLRs in the recognition of PSA, we tested the ability of different CLR blocking antibodies to inhibit PSA binding. Based on titration of the DC-SIGN antibody in previously published experiments (Geijtenbeek et al., 2000b), a concentration of 20 μg/ml blocking antibody was used. We observed that only the DC-SIGN blocking antibody decreased binding/internalization significantly, which was not further enhanced in combination with anti-MR or DCIR blocking antibodies (Figure 3B). These data confirmed DC-SIGN as the main receptor on DCs for the recognition of PSA. DC-SIGN functions as internalization receptor that mediates antigen routing to MHC class II loading compartments (Engering et al., 2002a). To assess the contribution of DC-SIGN in the MHC class II presentation of PSA to T cells, DCs were pulsed with PSA in the presence or absence of DC-SIGN blocking antibodies for 3 h, after which unbound PSA was washed away. PSA binding/uptake by DCs was clearly decreased in the presence of α-DC-SIGN (Figure 3C). The observed increase in CD4⁺ T cell proliferation by PSA-incubated DCs confirms the presence of PSA-specific T cells in blood (Figure 3D). Furthermore, PSA-specific T cell proliferation was completely inhibited in the presence of blocking antibodies to DC-SIGN (Figure 3D); demonstrating a DC-SIGN-dependent role in the uptake and processing of PSA for MHC class II presentation and subsequent T cell activation.

In conclusion, we have shown that DC-SIGN plays a crucial role in the binding and internalization of the T cell activating ZPSs PSA by human DCs. In addition, inhibiting DC-SIGN function decreases MHC class II presentation and subsequent T cell activation.