SEA and Δω1-SEA from S. mansoni eggs were prepared and isolated as described previously (22).

SEA and Δω1-SEA were treated with Endo-H (NEB #P0702S) according to their non-denaturing protocol conditions. Succinctly, for both SEA and Δω1-SEA, 500 µg was treated with Endo-H and 500 µg was mock-treated. The samples were not denatured. Buffer was directly added along with 12.500 units of Endo-H. The samples were then incubated at 37°C for 24 h. Removal of oligomannose N-glycans was confirmed by mass spectrometry (Supplementary Figure S1).

High-mannose glycans were cleaved from human serum proteins using Endo-H (NEB #P0702S) and incubated at 37°C for 48 h. The released glycans were purified by the application of C18 SPE (J.T.Baker, #7020-03) and carbon SPE columns (Supelco, #57088) and labeled with anthranilic acid (AA) by reductive amination. To remove the excess labeling reagent, acetonitrile (ACN) was added to a final concentration of 75%, and the sample was loaded onto Bio-Gel P10 Gel resin (catalog no.: 1504144; Bio-Rad) previously conditioned with 80% ACN. The glycans were eluted with MQ and dried using Speedvac. The glycans were then further fractionated using reverse-phase HPLC (RP-HPLC) and analyzed by MALDI-TOF MS, yielding pure Man9 glycan. The above-mentioned methodology is described in more detail by Petralia et al. (23)

2-Amino-N-(2-aminoethyl)-benzamide (AEAB)-labeled Man9 glycans were generated from AA-labeled Man9 as previously described (24). In short, 20 µg of both Maltopentose-AA [generated from maltopentose (Sigma #SMB01321) as described for Man9 above] and Man9-AA were mixed with 50 µL EDC (10 mg/mL in DMSO) and 50 µL HOBt (10mg/mL in DMSO). Subsequently, we added 10 µL of 5% (v/v) EDA and 0.5 M MES buffer (pH = 6.5). The samples were then vortexed for 1 min and incubated at room temperature for 3 h and then quenched with 1.1 mL of cold ACN. The samples were then vortexed and stored at -20°C for 30 min. The cloudy reaction mixture was centrifuged for 10 min at 10,000 × g. The supernatant was discarded and the precipitate was dried under vacuum. Once dry, it was dissolved in 100 µL of Mili-Q and applied to a RP-HPLC C18 column for purification.

Both Man9-AEAB and Maltopentose-AEAB were conjugated to 100 nm NHS-activated gold nanoparticles using a kit (Cytodiagnostics #CGN10K-100-1) and following the manufacturer’s protocol. As a deviation to the protocol, the glycans were diluted to 0.1 μg/μL using protein re-suspension buffer, and 1× PBS was used instead of conjugate storage buffer. The reaction efficiency was estimated to be ≈58% via RP-HPLC with fluorescence detection to quantify the percentage of recovered uncoupled material.

Peripheral blood mononuclear cells were isolated from the venous blood of healthy volunteers by density centrifugation in Ficoll as described before (25). Monocytes were isolated by positive magnetic cell sorting using CD14-microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany) and cultured in 10% FCS RPMI medium supplemented with 20 ng/mL rGM-CSF (BioSource/Invitrogen, Carlsbad, CA, USA) and 0.86 ng/mL of rIL-4 (R&D Systems, Minneapolis, MN, USA). On days 2 and 3, the medium, including the supplements, was replaced.

Immature moDCs were stimulated on days 5 and 6 in the presence or absence (if indicated) of 25 ng/mL ultrapure LPS (Escherichia coli 0111 B4 strain; InvivoGen, San Diego, CA, USA) along with the indicated reagents: SEA (20 μg/mL), Δω1-SEA (20 μg/mL), 50 μg/mL Zymosan (Z4250; Sigma-Aldrich, St. Louis, MO, USA), CIII (10 µM), indomethacin (Sigma-Aldrich #I7378, 50 µM), and 5, 2, or 0.2 μg/mL of either Man9- or Maltopentose-bound nanoparticles or inactivated nanoparticles.

After 48 h of stimulation, the surface expression of co-stimulatory molecules was determined by flow cytometry (FACS-Canto; BD Biosciences, Breda, The Netherlands or Aurora; Cytek, Amsterdam, The Netherlands) using the following antibodies: CD1a (clone HI149), CD14 (clone MΦP9), CD86 (clone 2331 FUN-1), CD40 (clone 5C3), and CD80 (clone L307.4) (all BD Biosciences); HLA-DR (clone LN3) CD83 (clone HB15e) (both eBioscience, San Diego, CA, USA); and CD252/OX40L (clone ANC10G1; Ancell, Bayport, MN, USA). Only live cells that were negative for Zombie NIR (BioLegend Europe BV, Amsterdam, The Netherlands) were included in the analysis. The acquired samples were unmixed using SpectroFlo version 3 (if measured with Aurora) and analyzed with FlowJo.

For the analysis of T cell polarization, 5 × 10³ moDCs pulsed for 48 h were cultured with 2 × 10⁴ allogenic naïve CD4+ T cells for 7 days in the presence of staphylococcal enterotoxin B (10 pg/mL). On day 7, the T cells were replated and rhuIL-2 (10 U/mL; R&D Systems) was added to expand the T cells. On days 9 and 10, the T cells were split with medium containing the same concentration of rhuIL-2. Intracellular cytokine production was analyzed on day 12 after restimulation with 100 ng/mL phorbol myristate acetate, 2 μg/mL ionomycin, and 10 μg/mL brefeldin A for 4 h. Subsequently, the cells were fixed with 2% paraformaldehyde (all Sigma-Aldrich). The cells were permeabilized with permeabilization buffer (eBioscience #00-5523-00) and stained with antibodies against IL-4 and IFN-γ, respectively (BD Biosciences). The acquired samples were unmixed using SpectroFlo version 3 (if measured with Aurora) and analyzed with FlowJo.

Lipid mediators (LM) and polyunsaturated fatty acids (PUFA) were measured using reverse-phase liquid chromatography coupled to tandem mass spectrometry (RPLC-MS/MS) as previously described (26), with some modifications. Briefly, 2 µL internal standard (IS) mix of deuterated lipid standards consisting of PGE2-d₄, 15-HETE-d₈, Leukotriene B₄-d₄, DHA-d₅, 8-iso-PGF2a-d₄, and 14 (15)-EET-d₁₁ (50 ng/mL in MeOH) was added to 400 µL culture supernatants. Lipids were extracted and purified by solid-phase extraction (SPE) after protein precipitation with 1.2 mL MeOH. The dried extracts were reconstituted in 100 µL 40% MeOH and transferred into a micro-vial glass insert. Furthermore, a 40-µL sample was injected and analyzed using a Shimadzu Nexera LC40 system with an autosampler coupled to a QTrap 6500 mass spectrometer (Sciex). Kinetex C18 50 × 2.1 mm, 1.7 µm column, and C8 precolumn (Phenomenex) were used for LC separation. LC–MS/MS chromatograms were integrated manually using Sciex OS (Sciex). The results were reported as relative peak area of lipids to the internal standards. PGE2-d₄ IS was used for reporting the area ratios of PGE2, TxB2, and PGF2a.

For the detection of ERK phosphorylation (pERK), 2.5 × 10⁴ immature moDCs were seeded overnight in a 96-well flat-bottom plate. moDCs were stimulated with SEA (25 μg/mL) and Δω1-SEA (25 μg/mL) for the indicated periods, and the moDCs were fixed for 15 min with 4% ultrapure formaldehyde (Polysciences, Warrington, PA, USA) directly in the plate. The cells were harvested and washed first with PBS and then with 0.5% of saponin for permeabilization. The cells were intracellularly stained with anti-phospo-p44/42 MAPK (Erk1/2) (clone E10) (both Cell Signalling Technology). Following 2 h of incubation at room temperature, the cells were washed with 0.5% of saponin, and ERK phosphorylation was determined by flow cytometry.

For the Dectin-1/2-hFc binding ELISAs, 96-well high-binding half-area microplates (Corning #10052511) were used. The appropriate antigens were coated in 50 μL TSM (20 mM Tris-HCl, 150 mM NaCl, 2 mM CaCl₂, and 2 mM MgCl₂ at pH 7.4) overnight at 4°C; SEA of Schistosoma mansoni (50 μg/mL), mock or Endo-H treated SEA of Schistosoma mansoni (50 μg/mL), Zymosan (20 μg/mL), NHS gold nanoparticles (5 μg/mL), or 1% BSA in TSM. After overnight coating, the plate was washed three times with an excessive amount of TSM/0.005% Tween. After washing, the plate was blocked for 1 h with 100 μL TSM/1% BSA at room temperature. After blocking, the plate was washed three times, again with an excessive amount of TSM/0.005% Tween. Following the washing, 50 μL Dectin-1-hFc (Sino Biological #10215-H01H) or Dectin-2-hFc (Sino Biological #10250-H01H) at a concentration of 10 μg/mL in TSM/0.005% Tween was incubated for 2 h at room temperature. After incubation, the plate was washed five times again with an excessive amount of TSM/0.005% Tween. Next, 50 μL of Monoclonal Biotin Mouse anti-human IgG1-Fc (Invitrogen #05-3340) (1:500 in TSM/0.005% Tween) was added to the plates, along with HRP-Strep (BD 51-9002813), and incubated for 1 h. The plate was then washed again for six times. We used 50 μL of TMB ELISA substrate solution (ThermoFisher #34021) for 30 min, followed by 25 μL of H₂SO₄ 1.8 M to stop the coloring reaction. ELISA readout was performed at 450 nm, with absorbance correction at 570 nm, using MultiskanTM FC Microplate Photometer (ThermoFisher, type 357).

Wild-type (WT) mice, both male and female and all on a C57BL/6J background, were bred under SPF conditions at the Leiden University Medical Center (LUMC), Leiden, The Netherlands. The mice were culled through cervical dislocation. Animal experiments were performed when the mice were between 8 and 16 weeks old. The animal experiments were performed in accordance with local government regulations, EU Directive 2010/63EU, and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes as well as approved by the Dutch Central Authority for Scientific Procedures on Animals (CCD) (animal license number AVD116002015253).

The mice were injected subcutaneously with 5,000 S. mansoni eggs in the hind footpad and injected i.p. with either vehicle or 1 mg CIII/20 g of body weight. The injections with either CIII or vehicle were then repeated on days 2 and 4 post-immunization. At 7 days later, the mice were sacrificed, and cells from both draining and nondraining lymph nodes were isolated and analyzed as described below.

Antigen-specific responses were determined by culturing 5 × 10⁵ LN cells per well in high-binding 96-well flat-bottom plates (Corning #3590) in 200 μL complete medium (RPMI containing 10% FCS, 100 U/mL penicillin/streptomycin, and 2 mM L-glutamine) in the presence of 20 μg/mL SEA along with 2.5 μg/mL IL-4R blocking antibody to retain IL-4 in culture supernatants. After 48 h, the culture supernatants were stored for cytokine determination. The cell culture supernatants were analyzed for cytokines using the Cytokine Bead Array (BD) according to the manufacturer’s recommendation. The samples were analyzed on BD Canto II Flow Cytometer. Alternatively, cytokine production was assessed by intracellular staining of T cells from LNs after polyclonal restimulation in 96-well flat-bottom plates for 4 h with phorbol 12-myristate 13-acetate (PMA; 50 ng/mL), ionomycin (1 μg/mL), and brefeldin A (10 μg/mL; all from Sigma-Aldrich). Afterward, the cells were fixed with 2% PFA and subsequently stained in eBioscience permeabilization buffer: IL-4 (11B11), IFN-γ (XMG1.2), IL-13 (eBio13A), IL-17A (TC11-18H10.1), and CD4 (RM4-5), IL-10 (JES5-16E3) (all BD Bioscience or BioLegend). The samples were analyzed on BD Canto II Flow Cytometer or Cytek Aurora. The acquired samples were unmixed using SpectroFlo version 3 (if measured with Aurora) and analyzed with FlowJo.

Data were tested for normality using the Shapiro–Wilk test. The statistical tests used are indicated in the figure legends. Generally, data were compared using one-way ANOVA for more than two groups or two-way ANOVA for comparing multiple parameters across two or more groups, with Tukey’s post-hoc test for multiple comparison. If comparing parameters within the same sample, a paired or repeated-measures test with Geisser–Greenhouse correction was used. p-values <0.05 were considered significant (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). All statistical analyses were performed using GraphPad Prism v.9.0.

We previously reported that blocking antibodies against Dectin-1 and Dectin-2 in DCs were able to impair omega-1-independent Th2 priming by SEA (10). However, it was not assessed whether the components in SEA were directly interacting to Dectin-1 and/or Dectin-2 to promote this response. To determine this directly, we performed a binding ELISA with a construct consisting of the carbohydrate binding domain of Dectin-1 or Dectin-2 coupled to a human IgG1 Fc domain. Zymosan, a known ligand for Dectin-1 and Dectin-2, was taken along as positive control. We observed that both Dectin-1 and Dectin-2 were directly able to bind to the components present in SEA (Figures 1A, B).

These receptors are C-type lectins that preferentially bind to sugar residues. It was shown in previous studies (8) that high-mannose glycans (e.g., Man9), known to be Dectin-2 ligands (20, 27), are present in high frequency in the mixture. To assess if the presence of these high-mannose moieties was required for SEA to interact with Dectin-1 and Dectin-2, we treated SEA with endoglycosidase-H (EndoH) to specifically remove N-linked oligomannose glycans, including those with high mannose, without affecting complex N-linked glycans or O-linked glycans (Supplementary Figure S1). EndoH hydrolyzes the glycosidic linkage of high-mannose glycans between GlcNAc1 and GlcNAc2, resulting in free glycans with only one GlcNAc residue present in the core, while PNGase-A cleaves off glycans between GlcNAc 1 and asparagine, resulting in free glycans with two GlcNAc residues present in the core. The success of EndoH treatment can thus be confirmed in the top spectrum due to the absence of high-mannose glycans with two GlcNAc residues, indicating that all high-mannose glycans were released by Endo-H before the PNGase-A treatment.

As assessed in a Dectin-1 and Dectin-2 binding ELISA, the hydrolysis of high-mannose glycans in SEA did not affect binding by Dectin-1. However, this treatment did significantly decrease the binding of SEA by Dectin-2, suggesting that SEA directly interacts with the latter in a partly high-mannose-dependent manner (Figures 1C, D).

Considering that SEA requires the presence of oligomannose glycans, including Man9, to bind to Dectin-2, we wondered if the presence of these glycans was important for the induction of the signaling cascade leading to OX40L expression and Th2 polarization by moDCs. EndoH-treated SEA lost its ability to induce ERK phosphorylation (Figure 2A), OX40L expression (Figure 2B), and subsequent Th2 polarization by moDCs (Figure 2C). The latter two readouts were performed in the presence of LPS as neutral DC maturation factor (5). Correspondingly, the synthesis of PGE2 was also reduced in moDCs stimulated with EndoH-treated SEA (Figure 2D). These effects were not confounded by the presence of omega-1 (ω1) as moDCs stimulated with SEA depleted of ω1 (Δω1-SEA), which had undergone EndoH treatment, were also compromised in their ability to induce a Th2 response compared to DCs stimulated with mock-treated control Δω1-SEA (Figure 2E). These data suggest that high-mannose residues are important for ω1-independent Th2 polarization by SEA.

To investigate if high-mannose glycans themselves are sufficient to recapitulate the effects of SEA, we coupled AEAB-labeled mannose-9 (Man9) oligosaccharides isolated from human serum to N-hydroxysuccinimide (NHS)-activated gold nanoparticles. Interestingly, while these loaded nanoparticles were efficiently covered with high-mannose glycans (Supplementary Figures S2A–D) and were able to directly bind to Dectin-1 and Dectin-2 as determined by ELISA (Supplementary Figures S3A, B), they were not able to replicate the effects of SEA on moDCs in terms of inducing OX40L on moDCs or condition them for Th2 polarization in a concentration of glycans labeled to NPs ranging from 0.2 to 5 µg/mL (Figures 2F, G). This indicates that, while high-mannose glycans are required for the activation of the dectin-OX40L axis by SEA, they are not sufficient to activate this pathway.

We previously found that SEA-driven Th2 polarization via Dectin-1/2 is critically dependent on PGE2, one of the downstream products of COX. However, in contrast to PGE2 neutralization experiments, COX inhibition was only able to modestly decrease Th2 priming (10). We wondered whether this could be explained by the fact that COX inhibition does not only inhibit PGE2 synthesis but also affects the synthesis of other COX products that may affect the Th2 priming capacity of DCs. To test this and avoid this potentially confounding issue, we used CIII, an inhibitor that specifically targets microsomal Prostaglandin E synthase-1 (mPGES) without affecting other COX-derived products (28) (Supplementary Figure S4).

Corresponding with the inhibition of PGE2 synthesis by both drugs, their incubation reduced the Δω1-SEA-driven expression of OX40L by moDCs (Figure 3A). While no statistical difference was found when comparing CIII to indomethacin (p = 0.442) (Figure 3B), CIII treatment significantly lowered the Th2 priming ability of Δω1-SEA-treated moDCs. In contrast, this was not the case for treatment with indomethacin, suggesting that targeting PGE2 synthesis itself is superior in modulating Th2 priming in this setting to targeting COX further upstream.

In view of these results that we obtained in vitro, we wondered if PGE2 synthesis inhibition would also reduce Th2 polarization in vivo. To test this, we injected S. mansoni eggs in the footpad of wild-type mice, followed by i.p. injections of CIII on days 0, 2, and 4 post-challenge. On day 7, the mice were sacrificed, and the CD4⁺ T cell response in the draining and non-draining lymph nodes was characterized.

Mice challenged with S. mansoni eggs and treated with CIII displayed an overall lower number of CD4⁺ T cells in draining LNs (Figure 4A), resulting in a lower number of IL4- and IL13-producing Th2-polarized T cells upon polyclonal restimulation, when compared to egg-immunized mice injected with the vehicle control. However, this was also true for the number of IFNγ- and IL-17-producing Th cells. As a consequence, CIII treatment did not alter the ratio between IL4- and IFNγ-producing CD4⁺ T cells (Figures 4B–F). No difference was seen in IL10-producing CD4⁺ T cells (Figure 4G). To assess antigen-specific cytokine responses, cells that had been isolated from the LNs of immunized mice were restimulated with SEA. While no difference was seen in the IL-4 response (Figure 4H), cell cultures from egg-immunized mice receiving CIII displayed a nearly significant lower IL-13 (Figure 4I) and a decreased IFNγ (Figure 4J) response to the antigens, while the IL-17 (Figure 4K) and the IL-10 (Figure 4L) levels remained unaffected. This indicates that the optimal priming of Th2 cell responses upon S. mansoni egg challenge in vivo, as well as that of other concomitant egg-induced Th cell responses, relies on de novo PGE2 synthesis.