PGE2‐EP4 signaling steers cDC2 maturation toward the induction of suppressive T‐cell responses

Dendritic cells (DCs) shape adaptive immunity in response to environmental cues such as cytokines or lipid mediators, including prostaglandin E2 (PGE2). In cancer, tumors are known to establish an enriched PGE2 microenvironment. Tumor‐derived PGE2 primes regulatory features across immune cells, including DCs, facilitating tumor progression. PGE2 shapes DC function by providing signaling via its two so‐called E‐prostanoid receptors (EPs) EP2 and EP4. Although studies with monocyte‐derived DCs have shown the importance of PGE2 signaling, the role of PGE2‐EP2/EP4 on conventional DCs type 2 (cDC2s), is still poorly defined. In this study, we investigated the function of EP2 and EP4 using specific EP antagonists on human cDC2s. Our results show that EP2 and EP4 exhibit different functions in cDC2s, with EP4 modulating the upregulation of activation markers (CD80, CD86, CD83, MHC class II) and the production of IL‐10 and IL‐23. Furthermore, PGE2‐EP4 boosts CCR type 7‐based migration as well as a higher T‐cell expansion capacity, characterized by the enrichment of suppressive rather than pro‐inflammatory T‐cell populations. Our findings are relevant to further understanding the role of EP receptors in cDC2s, underscoring the benefit of targeting the PGE2‐EP2/4 axis for therapeutic purposes in diseases such as cancer.


Introduction
DCs are a heterogeneous population of antigen-presenting cells that have the unique ability to initiate and shape adaptive immune responses.Activation of DCs is typically triggered by conserved pathogen-associated molecular patterns that induce a maturation process crucial for their ability to migrate and activate T Correspondence: Prof. I. Jolanda M. de Vries e-mail: Jolanda.deVries@radboudumc.nl cells.Besides these pathogen-associated molecules, the microenvironment in which DCs reside is usually rich in inflammatory mediators such as cytokines, growth factors, and lipid signaling molecules including prostaglandins, that can directly activate and steer DCs to induce suppressive or pro-inflammatory T-cell responses [1,2].
In human DCs, PGE2 exerts its immunomodulatory effects through E-prostanoid receptors (EP) EP2 and EP4 [14,28,31].PGE2 pleiotropic effects can be attributed to the existence of these receptors.Although both these receptors stimulate adenylate cyclase activity to increase the intracellular levels of cAMP, they possess different affinities for PGE2 [4,[32][33][34][35][36]. Furthermore, EP2 stimulates protein kinase A activity, whereas EP4 stimulates both protein kinase A and phosphoinositide 3-kinase [34].PGE2 signaling via EP4 results in Th17 differentiation [37,38] and DC migration toward lymph nodes [39].In contrast, PGE2-induced EP2 signaling promotes a suppressive phenotype characterized by the upregulation of IL-10 [14,15] and impaired T-cell stimulation [40].Yet, the molecular mechanisms as well as the specific role of each EP receptor in DCs are still poorly defined.To date, the function of EP2 and EP4 is primarily studied in mouse DCs or human moDCs.Considering that PGE2 is regarded as a major factor driving the development of suppressive DCs in pathologies such as cancer, it is crucial to increase our understanding of the role of EP2 and EP4 signaling in DC subsets present in the human body.
In blood, two major groups of DCs can be distinguished, the conventional DCs (cDCs) and the plasmacytoid DCs (pDCs) [41,42].The most prominent blood cDC subset is the CD1c + cDC, termed cDC2, characterized by their functional plasticity and versatility to modulate T-cell responses according to environmental cues [43,44].The tumor microenvironment can shift the cDC2 phenotype and function toward a regulatory state detrimental to antitumor immunity [8,44,45], and is associated with the failure of immunotherapy [46].
We here examine the individual contribution of EP2 and EP4 in PGE2-induced phenotype and function of cDCs type 2 (cDC2s).We first show that EP2 and EP4 gene expression is associated with a worse clinical outcome using a publicly available cancer patients' database.Next, we demonstrate that EP4 signaling results in the upregulation of maturation markers, IL-10 and IL-23 production, increased CCR7-based DC migration, and expansion of suppressive T-cell populations.EP2 signaling resulted in a modest upregulation of CCR7 and IL-23.Taken together, these results identify the differential function of EP2 and EP4 in human cDC2s and provide a rationale to intervene in the PGE2-EP2/4 axis for therapeutic purposes and prevent cDC2 skewing toward immunosuppression in PGE2-driven immune disorders such as cancer.

EP2 and EP4 expression correlates with poor survival in cancer patients
The notion that the presence of PGE2 coincides with a suppressive tumor microenvironment prompted us to investigate the relevance of EP2 and EP4 in tumors of cancer patients.Using the publicly available cancer genome atlas (TCGA) dataset, we assessed how PTGER2 (EP2 gene) and PTGER4 (EP4 gene) expression was associated with the clinical outcome of cancer patients.For the analysis, patient samples from the Pan-Cancer patient dataset with an enriched expression of the cDC2 marker CD1c were selected.In this patient dataset (4813 patients), a high PTGER2 (gene encoding EP2), and PTGER4 (gene encoding EP4) expression was significantly associated with lower overall survival in cancer patients (Fig. 1A).
Furthermore, we performed an expression correlation analysis between PTGER2 and PTGER4 and relevant genes associated with immunomodulation.Interestingly, expressions of several genes including CD86, PDL-1, CCR7, IL-10, and FOXP3 were positively correlated with PTGER4 and PTGER2 gene levels, whereas LAMP1, indicative of active cytotoxicity and tumor killing, was negatively correlated with both PTGER2 and PTGER4 gene (Fig. 1B).The genes analyzed correlated more strongly with PTGER4 than with PTGER2.
This database analysis suggests that high EP2/EP4 gene expression is correlated with the expression of genes related to immunomodulation or suppression and is related to an unfavorable clinical outcome in cancer patients.This prompted us to further investigate the role of EP2 and EP4 on cDC2s in the context of cancer.

PGE2 modulates cDC2 phenotype via EP2 and EP4
DC maturation is characterized by the upregulation of costimulatory markers, antigen-presenting molecules, and cytokines which enable DCs to initiate and shape T-cell responses.To establish the effects of PGE2 on cDC2 maturation, cDC2s were matured with a maturation cocktail (MC, consisting of TNFa, IL-1β, and IL-6) either in the absence or presence of PGE2.The rationale for investigating PGE2 together with the proposed MC sets on early studies demonstrating PGE2 alone can only modestly affect the phenotype and function of DCs compared with when combined with proinflammatory cytokines (e.g.TNFα, IL-6, or IL-1β) or TLR agonists [23,24,27,28,47].Phenotype analysis showed that addition of PGE2 to the MC (MC+PGE2) enhanced the upregulation of the proinflammatory markers CD80, CD86, and CCR7 (Fig. 2A).Furthermore, the presence of PGE2 increased the production of the suppressive cytokine IL-10 as well as the proinflammatory cytokine IL-23 (Fig. 2B).These results indicate that PGE2 induces immunomodulatory effects in cDC2s.
PGE2 induces immunomodulatory effects through EP2 and EP4 in DCs [14,28,31].To identify whether EP2 and EP4 are expressed by cDC2s, cDC2s were isolated and EP2 and EP4 transcription levels were determined by quantitative RT-PCR.This demonstrated that EP2, as well as EP4, are expressed by cDC2s and that EP2 receptor gene expression is higher than EP4 receptor gene expression (Fig. 2C).

PGE2 signaling via EP4 modulates the upregulation of maturation markers and cytokines on cDC2s
To distinguish the effects of PGE2 signaling through either EP2 or EP4 on cDC2 maturation, MC+PGE2 was added to cDC2s that were preincubated for 2 hours with either an EP2 antagonist (aEP2) or an EP4 antagonist (aEP4) (Fig. 3A).Specifically, we employed the aEP2 AH6809 and the aEP4 L161-982, which prevent EP specific PGE2 binding and have previously been used on immune cells [10,15,26,35,48].MC+PGE2 resulted in the upregulation of CD80, CD83, CD86, HLA-DR, and the co-inhibitory molecule PDL-1 on cDC2s (Fig. 3B).Blockade of EP4 by aEP4 impeded PGE2-dependent upregulation of the costimulatory molecules CD80, CD83, and CD86 as well as PDL-1 and HLA-DR.EP2 blockade by aEP2 did not result in an effect on the phenotype of cDC2s (Fig. 3B).Overall, we conclude that the PGE2-induced expression of membrane markers associated with DC maturation occurs via EP4 and not via EP2.
Depending on the environmental stimuli, DCs can secrete different immune-skewing cytokines.We observed that addition of PGE2 to the MC led to an increased production of both IL-10 and IL-23 (Fig. 3C).EP receptor blockade showed that aEP4 decreased PGE2-triggered IL-10 production, whereas aEP2 did not (Fig. 3C).Interestingly, EP4 blockade significantly inhibited IL-23 secretion, whereas EP2 blockade enhanced IL-23 production, suggesting that EP2 either influences EP4 signaling, or it directly inhibits IL-23 production.

EP4 signaling enhances CCR7 expression and the migratory capacity of cDC2s
DC maturation is also defined by the acquisition of a migratory phenotype, enabling DCs to migrate toward lymph nodes to initiate T-cell responses.To assess the influence of PGE2 and EP2/EP4 on the migratory phenotype and function of cDC2s, we measured the expression of chemokine receptor CCR7 which binds to the lymph node homing chemokines C-C motif ligand (CCL)19 and CCL21.The addition of PGE2 during maturation led to CCR7 upregulation on cDC2s which was significantly reduced in the presence of aEP4 (Fig. 4A, B).
To assess whether PGE2-induced CCR7 expression also led to functionally migratory cDC2s, we performed a transwell migration assay toward CCL19 and CCL21 (Fig. 4C).In accordance with the expression level of CCR7, the addition of PGE2 during maturation increased the chemotactic migration of cDC2s.Importantly, blockade of EP4 but not of EP2, resulted in a comparable migratory capacity as DC matured without PGE2 (Fig. 4D, E).This data demonstrates that PGE2-EP4 signaling is responsible for the induction of the cDC2 migratory capacity.

PGE2-EP4 matured cDC2s induce a superior T-cell proliferation
Efficient adaptive immunity relies on the ability of DCs to induce T-cell proliferation.The expression of maturation markers on DCs is crucial for the initiation and expansion of T cells.To test the capacity of PGE2-activated cDC2s to expand T cells, cDC2s were co-cultured with allogeneic T cells labeled with the CFSE dye.
Compared with cDC2s stimulated with MC only, T-cell proliferation was increased when MC+PGE2-stimulated cDC2s were used (Fig. 5).Although EP2 blockade did not have an effect, EP4 blockade significantly prevented the PGE2-enhanced ability of cDC2s to expand T cells (Fig. 5).In line with the expression of co-stimulatory molecules, we observed that PGE2 triggering of EP4 and not EP2 leads to an enhanced capacity of cDC2s to induce T-cell expansion.

EP2 and EP4 matured cDC2s lead to different T-cell populations
DCs are key players during the polarization of naïve T cells into different T-cell subsets.This T-cell polarizing ability is mediated by the variety of cytokines that can be produced by DCs.For instance, production of IL-12 promotes Th1 differentiation, whereas IL-23 and IL-10 facilitate the expansion of Th17 and Tregs [49][50][51].cDC2s were cocultured with allogeneic naïve T cells and the generation of Th1, Th17, Th2, and Tregs was assessed by determining the expression of T-bet (Th1), GATA3 (Th2), RORγt (Th17), and CD127, CD25, FOXP3 (Treg) (Fig. 6).T-cell analysis showed that MC+PGE2 matured cDC2s expanded Tregs.No differences were observed regarding Th17 cells.We noticed a significant but modest downregulation of Th1 and Th2 cells.Interestingly, while no effect of EP blockade was observed for the markers of Th17, Th1, or Th2 cells, aEP4, but not aEP2, abolished the enhanced induction of Tregs (Fig. 6A), indicating that EP4-signaling is responsible for the PGE2-induced expansion of Tregs.
Next, we determined the production of cytokines characteristic for different T-cell subsets (Fig. 6B).T cells expanded upon stimulation with cDC2s matured in the presence of PGE2 produced significantly more IL-10, slightly more IL-5 and IL-17, and similar levels of IFN-γ as upon stimulation with cDC2s matured with MC only (Fig. 6B).EP4 blockade during MC+PGE2 maturation of cDC2s prevented the increased IL-10 production and slightly affected IL-5 and IL-17 production.
To investigate the effect of the presence of PGE2 during cDC2 maturation on the capacity of these cells to induce CD8 T-cell activation, we analyzed the intracellular levels of the cytotoxic T-cell markers perforin and granzyme-B in expanded CD8 T cells.Interestingly, both perforin and granzyme-B levels were downregulated in T cells activated by MC+PGE2-matured cDC2s compared with MC-matured cDC2s.By adding aEP4 during MC+PGE2 maturation of cDC2s, we observed a tendency for EP4 signaling to account for the downregulation of these markers (Fig. 6C).Collectively, these results show that PGE2 signaling in cDC2s via EP4 and not EP2 skews the T-cell polarizing capacity of cDC2s toward the expansion of suppressive T-cell populations (e.g.increased Tregs and IL-10 producing T cells).

Discussion
In this study, PGE2-mediated EP2 and EP4 signaling in the most abundant blood DC subset, the cDC2s was investigated.PGE2 is an important modulator of DC function and a hallmark of cancer progression.In fact, we observed that high EP2 or EP4 expression is associated with unfavorable survival of cancer patients.Here, we report that PGE2 contributes to the acquisition of a mature cDC2 state, which is characterized by the upregulation of costimulatory molecules and the chemokine receptor CCR7 together with a functional migratory phenotype.Furthermore, PGE2 stimulation during maturation of cDC2s resulted in an increased expansion of allogeneic T cells.cDC2s matured in the presence of PGE2 secreted high levels of IL-10 and IL-23 and skewed T-cell polarization toward the differentiation of suppressive T cells such as Tregs rather than proinflammatory populations.EP receptor antagonism revealed a central role for PGE2 signaling via EP4 and not via EP2 during the acquisition of such phenotype and function.Interestingly, correlation analysis of the cancer patient database showed a stronger correlation among genes such as IL-10, CCR7, FOXP3, or PDL1 with EP4 rather than EP2, arguing for the dominant role of EP4 in PGE2-mediated immunomodulation also in cancer patients.
To investigate the individual role of EP2 and EP4, we used EPspecific antagonists as they recapitulate the significance of blocking EP receptors in a PGE2-rich environment.In fact, several EP antagonists are currently being explored in phase 1 clinical trials to interfere with PGE2-driven mechanisms in disorders such as cancer (53)NCT04344795).Previously, EP-specific agonists have been employed to study individual EP receptor function.However, we have previously shown that EP receptor function can be influenced by triggering them with EP-specific agonists or PGE2 in combination with EP antagonists.This suggests a not-yet-defined crosstalk between EP2 and EP4 during PGE2 engagement in the cell membrane [35].Hence, we evaluated the role of EP2 and EP4 on the phenotype and function of cDC2s upon exposure to PGE2 with EP-specific antagonists.
The expression of co-stimulatory molecules is a hallmark of DC maturation.In moDCs, addition of PGE2 during maturation has been demonstrated to upregulate co-stimulatory markers including CD80, CD86, CD83, and HLA-DR [15,20,23,26].In line with these reports, we observed that MC+PGE2 upregulated these markers also on human cDC2s.EP receptor blockade revealed that EP4 accounted for the upregulation of these maturation markers, whereas no effect was detected upon EP2 blockade.This observation contrasts previous studies using moDCs treated with EPspecific agonists that showed that both EP2 and EP4 are important for the upregulation of CD80, CD83, CD86, and HLA-DR [15,20].This dissimilarity might be explained by the difference in type of DC studied and the use of agonist instead of antagonist system as described above.Expansion of T cells is highly dependent on the full maturation state of cDC2s [43].Like what was previously reported for moDCs [24,25], cDC2s treated with MC+PGE2 exhibited a superior T-cell expansion capacity as compared with cDC2s matured in the absence of PGE2.This superior expansion was to be expected, as T-cell expansion is mostly dependent on the expression of co-stimulatory molecules by DCs.In line with the expression pattern of co-stimulatory molecules, we observed that EP4 blockade diminished the superior T-cell expansion capacity of cDC2s matured in the presence of PGE2.
The trafficking of DCs from the periphery to the T-cell areas in secondary lymphoid organs is achieved by the expression of chemokine receptors, primarily the CCL19/CCL21 receptor CCR7.We report that PGE2 via EP4 mediates the upregulation of CCR7 as well as a migratory function of cDC2s toward the chemokine ligands CCL19/CCL21.Previous studies using moDCs have demonstrated that PGE2 is a mandatory factor for the induction of CCR7-based migration [27][28][29].Regarding human cDC2s, one study has evaluated the role of PGE2 in cDC2 migration [28].The authors reported that PGE2 enhances the cDC2s migratory capacity during TLR-mediated maturation.Regarding the role of EP receptors, previous literature has demonstrated that, in human moDCs, EP2 and EP4 contribute equally to CCR7-mediated migration (using agonists), whereas human tolerogenic moDCs and mice DCs rely on EP4 for this functional feature [15,28,39].Like tolerogenic moDCs [15,39], we observed that CCR7-mediated migration is mediated via EP4 signaling.Moreover, we observed that EP2 signaling inhibits EP4-mediated CCR7 upregulation on cDC2s.cDC2s are known for their ability to produce a variety of polarizing cytokines such as IL-12, IL-23, or IL-10 [44].PGE2 has been reported to modulate the cytokine production of human cDC2s and moDCs.PGE2-matured DCs are characterized by impaired IL-12 production and an augmented IL-10 secretion [13][14][15] We also observed elevated IL-10 production, but no IL-12 was detected in the cDC2 supernatants.The absence of IL-12 production is most likely due to the cytokine cocktail used for maturation as opposed to TLR-induced maturation [53].EP4 blockade reduced the production of IL-10, identifying EP4 as the main receptor mediating IL-10 production by PGE2-matured cDC2s.A previous report showed that IL-10 production by cDC2s was dependent on EP2 triggering and not EP4 triggering.This discrepancy might be explained by using agonists to dissect EP2/EP4 receptor function or the different basal stimulation compared with the one used in this study [15] In accordance with previous studies with moDCs [14,15], we observed that PGE2 boosted the production of IL-23 by cDC2s.Like moDCs ( [14,15], our data revealed that EP4 accounts for the upregulation of IL-23 and that EP2 blockade further increased IL-23 production.Overall, the increased IL-10 and IL-23 production by cDC2s is due to EP4 signaling. Previous studies showed that PGE2-matured moDCs skew Tcell responses toward the differentiation of Th2/Tregs rather than Th1/17-cell responses [18][19][20][21].Here, we show that PGE2matured cDC2s yield high numbers of CD127-CD25+FOXP3+ T cells, and high levels of IL-10 in the supernatant, indicative of Tregs.Our results are in line with previous observations where Treg expansion is facilitated by DC-derived IL-10 [7,19].T-cell skewing by PGE2-matured cDC2s toward Treg was mediated via EP4.This seems in contrast with a previous publication showing that signaling through both EP2 and EP4 is responsible for the upregulation of IL-10 in T-cell cultures [20].The contrasting results could be explained as a different DC type that was used (moDCs) together with an agonist system without PGE2 for the engaging of the EP receptors.
Activation of CD8 T cells by cDC2s that were matured in the presence of PGE2 resulted in the expansion of CD8 T cells exhibiting reduced levels of intracellular cytotoxic molecules perforin and granzyme-B, indicative of CD8 T cells with lower cytotoxic potential.For CD4 T cells it is shown that PGE2-matured moDCs suppress polarization toward Th1 and favor Th2 skewing [19,20].Here, stimulation of T cells with PGE2-matured cDC2s resulted in a significant decrease in T-bet expression and GATA3 but no significant changes in IFN-γ production nor in IL-5 production.Taken together with the observed minimal higher IL-17 production, we conclude that PGE2-cytokine matured cDC2s do not consistently skew T-helper cells toward Th1, Th2, or Th17.Of interest, slightly increased IL-17 production was observed by T cells stimulated with PGE2-matured cDC2s in the presence of aEP2.This is likely due to the superior release of IL-23 by cDC2s upon aEP2 treatment, a feature previously reported as well for moDCs [14,15].
Our data identifies PGE2 as modulator of activating and suppressive features in human cDC2s.Using a cytokine cocktail for cDC2 maturation, we observed that PGE2 signaling via EP4 is responsible for cDC2 maturation as well as the production of IL-23 and IL-10, which facilitated the development of Tregs.In contrast, EP2 receptor blockade resulted in a modest upregulation of IL-23 and CCR7.We propose that PGE2 signaling via EP4 and not EP2 accounts for cDC2 modulation, favoring suppressive rather than proinflammatory responses.Considering that a high EP2/EP4 gene score associates with a worse clinical outcome in cancer patients together with the notion that EP4 is the main receptor responsible for cDC2 modulation, it is conceivable that EP4 signaling might recapitulate a similar phenotype and function in cDC2s present in the TME of cancer patients.In fact, a recent mice study reported the central role of PGE2 signaling in the development of suppressive DCs characterized by the high expression of EP4, CCR7, and chemokines involved in Treg recruitment [11].Given our observation that in human cDC2s, EP4 mediates the upregulation of CCR7, IL-10, and the Treg differentiation capacity, we propose that the PGE2-EP4 axis might shift human cDC2s toward a similar suppressive state in the TME, arguing for the clinical benefit of targeting PGE2-EP4 signaling in human DCs for the treatment of cancer.In this sense, recent mouse models have illustrated the therapeutic value of targeting the PGE2-EP2/4 axis to prevent tumor-derived immunosuppression and thus, prevent tumor growth [10,11].However, further research is needed to fully understand the relevance of EP2 and EP4 functions on other human DC subsets such as cDC1 and pDCs.Identifying the specific role of EP2 and EP4 receptors in human DCs may help engineer novel strategies to modulate PGE2 signaling, alleviate tumor-derived suppression, and thus, facilitate the development of antitumor immune responses in cancer patients.

Real-Time quantitative PCR
Total RNA isolation was performed from freshly isolated cDC2s with RNeasy Mini Kit columns (Qiagen) following the manufacturer's instructions.RNA reverse transcription was performed using the High-Capacity cDNA RT kit (Thermo Fisher) using a Veriti 96w thermocycler (Thermo Fisher Scientific).The applied program consisted of 25°C, 10 min; 37°C, 120 min using the Taq-Man real-time PCR (Thermo Fisher) for the detection of PTGER2 and PTGER4 transcripts, using previously used primers [15].

DC phenotype analysis
Flow cytometry was used to assess the phenotype and matured status of DCs.cDC2s were collected from the culture plated, washed, and incubated with blocking buffer (PBS + 5% FBS + 0.01% NaH3 + 5% HS) for 15 min.After blocking, cDC2s were stained with antibodies binding to CD80 PerCP

Transwell migration assay
Migration assays were performed using 96-well plates with 5 μm pore size (CLS3388-2EA, Corning Inc.).Lower compartment of the transwell plate was loaded with 200 μLs X-vivo 2% HS containing 100 ng/mL of CCL19 and CCL21 (582102, 582202, BioLegend).To assess passive DC migration, medium without CCL19/21 was used.3×10 4 DCs were seeded in 30 μL volume in the upper compartment.The plate was incubated for 3 h at 37°C, 5% CO 2 to allow DC migration.Migrated cells were collected from the lower compartment and quantified on a MACSQuant Analyzer 10 Flow Cytometer (Miltenyi Biotec) together with the initially loaded population of added cDC2s to the upper compartment of the transwell plate.DC migration was calculated in % migrated cells as (number of migrated DCs in condition)/(initially loaded DC in conditioned) × 100.

Mixed lymphocyte reaction
The ability of stimulated DCs to induce T-cell proliferation was investigated in a mixed lymphocyte reaction.Allogeneic Pan T cells were stained with CFSE (C34554, ThermoFisher), 5 μM; Life Technologies) for 10 min; thereafter, reaction was stopped by blocking with FCS.Cells were washed and counted.cDC2s were cultured with CFSE-labeled T cells in a ratio 1:10 for 6 days at 37°C, 5% CO 2 .After the 6 days, T-cells were measured on the FACs to evaluate the loss of CFSE expression.T-cell stimulatory capacity was calculated in fold change as (% proliferated T cells per condition)/(% proliferated T cells under "Non-treated" DC condition].

Naïve T-cell polarization
cDC2s and allogeneic naïve T cells were cocultured at a ratio 1:5 (DC:T cell) in a final volume of 200 μLs in a 96 U bottom culture plate.On day 6, T-cell cultures were provided with 20 U/mL of recombinant IL-2 (130-097-744, Miltenyi Biotec), afterward, IL-2 was added every 2 days until day 12. Resting T cells were counted with the MACSQuant Analyzer 10 Flow Cytometer (Miltenyi Biotec), and either replated for cytokine production or phenotypical analysis.Phenotypical analysis was performed by flow cytometry.Live/dead staining was performed with eFluor780 viability day (1:2000, ThermoFisher) in PBS for 15 min at 4°C.Cells were later fixed and permeabilized using the FOXP3 transcription factor staining kit (00-5523-00, ThermoFisher) according to the manufacturer's instructions.Cells were blocked using permeabilization buffer containing 5% HS.Later, intracellular staining was performed in 25 μL for 20 min in permeabilization buffer with antibodies binding to T-bet PerCP-Cy5.

Publicly
available gene expression and survival datasets were analyzed using the UCSC Xena browser (https://xenabrowser.net/datapages/) [54].The Kaplan-Meier overall survival curves were generated from the PAN-CANCER dataset, selecting for the patient samples enriched for the expression cDC2 marker gene CD1c samples (13.97-20.75,n = 4813) and grouped based on PTGER2 and PTGER4 scores.The calculated P value using the log-rank test was provided by the UCSC Xena browser.Relationship between the expression levels of PTGER2/PTGER4 and CD86, PDL1, CCR7, FOXP3, IL10, and LAMP1 genes, and generated Pearson correlation values and P values (see Supporting information Fig. S5) were performed in the above-indicated dataset of cancer patients using the UCSC Xena browser.

Statistical analysis
Data analysis was performed using GraphPad Prism 8 (Version 8.0.2.GraphPad Software).Significance among the different conditions was assessed by one-way ANOVA with Tukey multiple comparison corrections, Friedman test followed by Dunn's testing, or a paired t-test.Statistical significance was defined as * P < 0.05; ** P < 0.01; * * * P < 0.001).analyzed the data.Jorge Cuenca-Escalona wrote the manuscript with input from all authors.Alessandra Cambi and I. Jolanda M. de Vries supervised the entire project.

Figure 1 .
Figure 1.Association of PTGER2 and PGER4 with different immune markers and clinical outcome of cancer patients.(A) Kaplan-Meier overall survival curve of Pan-Cancer (n = 4813) patients with a high expression of CD1c from the PAN-CAN TCGA datasets sorted by PTGER2 and PTGER4 expression.P value was calculated using the log-rank test.(B) Heatmap correlation plot of the expression of PTGER2 or PTGER4 and the indicated genes in the selected PAN-CAN human patient samples (n = 4813).

Figure 2 .
Figure 2. PGE2 modulates the phenotype of cDC2s which express EP2 and EP4.(A) Histograms showing the expression levels of CD80, CD86, and CCR7 of a single representative donor analyzed by flow cytometry on cDC2s after treatment with the cytokine maturation cocktail (MC) and the MC+PGE2.(B) Heatmap showing the cytokine production levels (pg/mL) of IL-10 and IL-23 upon the indicated treatment for a representative donor.(C) Expression of EP2 and EP4 receptors analyzed by RT-PCR on freshly isolated cDC2s.Bar graph showing the relative expression to b-actin as arbitrary units (AU).Data shown are the mean ± SEM of three independent experiments with cDC2s from three different donors (n = 3).P value was calculated using the Mann-Whitney test for nonparametric data.

Figure 3 .
Figure 3. PGE2-EP4 signaling modulates the upregulation of maturation markers and cytokines on cDC2s.(A) To examine the role of individual EP2 and EP4 signaling cDC2s were treated with either aEP2 or aEP4 2 hours prior to treatment with the MC (containing IL-6, IL-1β, and TNFα) together with PGE2 (MC+PGE2).After 24 h of stimulation, expression of maturation markers on cDC2s and cytokine production by cDC2s were assessed.(B) CD80, CD86, CD83, PDL1, HLA-ABC, and HLA-DR on cDC2s were analyzed by flow cytometry and are displayed as either % positive DCs (CD80, CD86, CD83, PDL1) or gMFI (HLA-DR, HLA-ABC) (see Supporting information Fig. S1 for the gating strategy).Bar graphs show the mean ± SEM and each data point represents an individual cDC2 donor (n = 6) from a single independent experiment, representative of three independent experiments with similar results.(C) For cytokine production analysis DCs were treated with the indicated stimuli for 24 h followed by an overnight re-stimulation with 100 ng/mL with LPS.cDC2 supernatants were analyzed by standard sandwich ELISA for IL-10 and IL-23 production.Relative cytokine production was assessed by normalizing the cytokine production of the MC-treated cDC2s.Bar graphs show the mean ± SEM and each data point represents an individual cDC2 donor (n ≥ 7) combined from three independent experiments.P values were calculated on raw concentration values with one-way ANOVA with Tukey multiple comparison correction (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001).

Figure 4 .
Figure 4. PGE2 induces CCR7-based migratory cDC2s via EP4 receptor signaling.cDC2s treated with the indicated stimuli for 24 h were analyzed for the expression of CCR7 and their migratory capacity toward the chemokine CCL19/21 in a transwell assay.(A) Histogram showing the CCR7 expression levels of a representative donor analyzed by flow cytometry on cDC2s after 24 h of stimulation with the indicated treatment.(B) Bar graph showing the percentage of CCR7 positive cDC2s (see Supporting information Fig. S1 for the gating strategy).Each data point represents an individual cDC2 donor from a single independent experiment (n = 5), representative of three independent experiments with similar results.(C) Schematic representation of the transwell migration assay.To assess cDC2 migratory capacity, cDC2s were added to the upper chamber of a transwell system.After 3 h, the number of migrated DCs toward the chamber containing the chemokine CCL19/21 were quantified by flow cytometry.(D) Pseudocolour dot plots showing the cDC2 population initially loaded (starting population) and the migrated cDC2s after 3 h of incubation from a single representative donor.(E) Bar graph showing the percentage of migrated cDC2s.Dotted grey line indicates the value of passive migration, non-CCL19/21 specific.Each data point represents an individual cDC2 donor (n = 7) combined from six independent experiments and bar graphs show the mean ± SEM.P values were calculated with one-way ANOVA with Tukey multiple comparison correction (*P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Figure 5 .
Figure 5. Triggering of EP4 by PGE2 during cDC2 maturation increases the capacity of cDC2s to expand T cells.cDC2s pretreated with the indicated components were cocultured with allogeneic T cells labeled with CFSE.On day 6, T-cell proliferation was assessed (see Supporting information Fig. S2 for the gating strategy).(A) Histogram showing the CFSE expression of a representative donor analyzed by flow cytometry.(B) Bar graph showing the relative proliferation of T cells.Relative proliferation was assessed by normalizing the T-cell proliferation of the MC-treated cDC2s.Each data point represents an individual cDC2 donor (n = 6) combined from six independent experiments, and the bar graphs show the mean ± SEM.P values were calculated with one-way ANOVA with Tukey multiple comparison correction (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001).

Figure 6 .
Figure 6.Triggering of EP4 by PGE2 during maturation of cDC2s leads to suppressive rather than proinflammatory T-cell populations.DCs treated with the indicated components were harvested and cocultured with allogenic naive T cells until day 12. (A) T cells were analyzed for the intracellular expression of transcription factors.The bar graph shows the percentage of Th1 (T-bet), Th2 (GATA3), Th17 (RORγt), and Treg cells (CD127 − CD25 + FOXP3 + ) (see Supporting information Fig. S3A and B for the gating strategy).Each data point represents individually expanded T cells from an individual cDC2 donor (n = 5) from a single independent experiment and the bar graphs show the mean ± SEM. (B) cDC2-expanded T-cell populations were re-stimulated for 48 h with anti-CD3/CD28 dynabeads and analyzed for cytokine production by multiplex assay (IL-10, IL-17, and IL-5) and ELISA (IFN-γ).Bar graph showing the concentration of IL-10, IL-5, IL-17, and IFN-γ.Each data point represents individually expanded T cells from an individual cDC2 donor (n ≥ 4) from a single-independent experiment and the bar graphs show the mean ± SEM (C) CD8 T cells were analyzed for the intracellular expression levels of perforin and granzyme-B (see Supporting information Fig. S3C for the gating strategy).Bar graph showing the percentage of positive cells for perforin and granzyme-B.Each data point represents individually expanded T cells from an individual cDC2 donor (n = 5) from a single independent experiment and the bar graphs show the mean ± SEM.P values were calculated with one-way ANOVA with Tukey multiple comparison correction or Friedman with Dunn's multiple comparison test (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001).