Prostaglandin D₂ metabolites activate asthmatic patient-derived type 2 innate lymphoid cells and eosinophils via the DP₂ receptor

Abstract

Background

Prostaglandin D₂ (PGD₂) signaling via prostaglandin D₂ receptor 2 (DP₂) contributes to atopic and non-atopic asthma. Inhibiting DP₂ has shown therapeutic benefit in certain subsets of asthma patients, improving eosinophilic airway inflammation. PGD₂ metabolites prolong the inflammatory response in asthmatic patients via DP₂ signaling. The role of PGD₂ metabolites on eosinophil and ILC2 activity is not fully understood.

Methods

Eosinophils and ILC2s were isolated from peripheral blood of atopic asthmatic patients. Eosinophil shape change, ILC2 migration and IL-5/IL-13 cytokine secretion were measured after stimulation with seven PGD₂ metabolites in presence or absence of the selective DP₂ antagonist fevipiprant.

Results

Selected metabolites induced eosinophil shape change with similar nanomolar potencies except for 9α,11β-PGF₂. Maximal values in forward scatter of eosinophils were comparable between metabolites. ILC2s migrated dose-dependently in the presence of selected metabolites except for 9α,11β-PGF₂ with EC₅₀ values ranging from 17.4 to 91.7 nM. Compared to PGD₂, the absolute cell migration was enhanced in the presence of Δ¹²-PGD₂, 15-deoxy-Δ^12,14-PGD₂, PGJ₂, Δ¹²-PGJ₂ and 15-deoxy-Δ^12,14-PGJ₂. ILC2 cytokine production was dose dependent as well but with an average sixfold reduced potency compared to cell migration (IL-5 range 108.1 to 526.9 nM, IL-13 range: 125.2 to 788.3 nM). Compared to PGD₂, the absolute cytokine secretion was reduced in the presence of most metabolites. Fevipiprant dose-dependently inhibited eosinophil shape change, ILC2 migration and ILC2 cytokine secretion with (sub)-nanomolar potencies.

Conclusion

Prostaglandin D₂ metabolites initiate ILC2 migration and IL-5 and IL-13 cytokine secretion in a DP₂ dependent manner. Our data indicate that metabolites may be important for in vivo eosinophil activation and ILC2 migration and to a lesser extent for ILC2 cytokine secretion.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12931-021-01852-3.

Background

Allergic and non-allergic asthma are closely linked to increased production of prostaglandin D₂ (PGD₂) [1–3]. The main source of PGD₂ are mast cells, which thereby orchestrate early and late immune responses of the innate and adaptive immune system. PGD₂ primarily signals through prostaglandin D₂ receptors 1 (DP₁) and 2 (DP₂, also known as chemoattractant receptor-homologous molecule expressed on Th2 cells [CRTH2]). Activation of DP₁ is associated with increased smooth muscle relaxation, vasodilation, vascular permeability as well as reduced leukocyte chemotaxis and cytokine secretion [4, 5]. In contrast, DP₂ signaling leads to activation, chemokinesis, migration, and cytokine production of various leukocytes such as eosinophils, basophils, T cells or type 2 innate lymphoid cells (ILC2s) [6–11].

In vivo and in vitro, PGD₂ is rapidly degraded either enzymatically or spontaneously to various PGD₂ metabolites of the D, F and J series (Fig. 1) [12, 13]. The persistence of degradation products in plasma and urine may indicate a biological relevance either by preventing or prolonging in vivo receptor signaling [6, 14–16]. Enzymatic degradation of PGD₂ leads to 13,14-dihydro-15-keto-PGD₂ (DK-PGD₂) which is a highly selective DP₂ agonist [14, 15]. In plasma, PGD₂ is mainly metabolized to ∆¹²-PGD₂ and ∆¹²-PGJ₂ which preferably bind to DP₂ [12, 15]. Dehydration of PGD₂ and its spontaneous degradation product 9-deoxy-PGD₂ (PGJ₂) lead to formation of 15-deoxy-∆^12,14-PGD₂^12,17 and 15-deoxy-∆^12,14-PGJ₂ [7, 17, 18], respectively, which accumulate in low concentrations and as well preferably bind to DP₂ [15]. A major enzymatically derived product of the PGD₂ metabolism in vivo is 9α,11β-PGF₂ which is e.g. found in urine and plasma of asthmatics following allergen challenge [14, 19, 20]. Since most metabolites have selective agonistic properties for DP₂ over DP₁ they are considered to contribute to inflammatory DP₂ signaling [15, 21].

Fig. 1.

Selected metabolism of prostaglandin D₂. Major metabolites are represented in bold. PGD₂, metabolite 2 and 3 belong to the D-series while metabolites 5, 6 and 7 belong to the J-series of prostaglandins. Metabolite 4 belongs to the F-series of prostaglandins

Targeting the DP₂ pathway is of significant interest for severe, uncontrolled asthmatic patients who do not or only partly respond to current available treatment options [1, 11, 22, 23]. Eosinophils and ILC2s are two major players in atopic asthma expressing DP₂. PGD₂ and metabolites activate eosinophils changing their cellular shapes [15, 18, 19], as well as inducing actin polymerization [21], CD11b expression [21] and migration [6, 12]. ILC2 numbers are increased in the airways of asthmatics and promote eosinophilia by secretion of type 2 cytokines [24, 25]. PGD₂ induces interleukin 5 (IL-5) and IL-13 secretion, cell migration, cell aggregation as well as the expression of adhesion molecules in ILC2s in a DP₂ dependent manner [26, 27]. Further, ILC2s produce endogenous PGD₂ which is degraded to PGJ₂, ∆¹²-PGJ₂, 15-deoxy-∆^12,14-PGJ₂ and DK-PGD₂ in culture supernatants [28].

Fevipiprant is a potent and selective DP₂ antagonist that has shown therapeutic benefit in certain subsets of asthma patients in phase 2 clinical trials [29–31]. In a single center mechanistic phase 2 clinical trial in patients with persistent eosinophilic asthma, fevipiprant not only reduced airway inflammation, but also improved epithelial integrity and reduced airway smooth muscle mass [31, 32]. In two recently published phase 3 studies in severe asthmatics, although neither trial showed a statistically significant reduction in asthma exacerbations, consistent and modest reductions in exacerbations rates were observed in both studies with a high dose of fevipiprant [33]. In two Phase 3 studies in moderate asthmatics, no significant improvements were observed in lung function or other asthma related outcomes, such as daytime symptom score or quality of life [34]. Previously, the inhibitory effects of fevipiprant in primary human cells have been characterized with PGD₂ activation [26, 35, 36].

Here, we demonstrate the potencies of seven PGD₂ metabolites to induce ILC2 migration and IL-5 and IL-13 cytokine secretion. We also reproduce previously reported effects of the PGD₂ metabolites on eosinophil shape change. ILC2 as well as eosinophil activation was then blocked with fevipiprant demonstrating the DP₂ dependency of cell activation.

Methods

Study design

Fifteen atopic asthmatic volunteers (eight male/seven female; aged 18–65 years, average 34.5 ± 10.6 years; BMI 19–32 kg/m², no oral steroids for > four weeks) were enrolled into the study. Subjects differed between eosinophil (n = 8) and ILC2 (n = 7) experiments and also between activation and inhibition experiments, respectively. Whole blood (200–500 ml) was collected in 3.8% trisodium citrate and was processed within one hour after blood withdrawal. Eosinophil levels in peripheral blood had to be > 0.15 × 10⁶/ml for the eosinophil shape change.

Reagents

PGD₂, 13,14-dihydro-15-keto-PGD₂, PGJ₂, Δ¹²-PGJ₂, Δ¹²-PGD₂, 15-deoxy-Δ^12,14-PGJ₂, 15-deoxy-Δ^12,14-PGD₂, 9α,11β-PGF₂ were purchased from Cayman Chemicals (Biomol GmBH, Hamburg, Germany). Fevipiprant (GST0000013789) was provided by Novartis Pharma AG (Basel, Switzerland). Reagents were dissolved in sterile-filtered Hybri-Max dimethylsulfoxide (DMSO, Sigma-Aldrich, Taufkirchen, Germany).

Granulocyte isolation

Blood was stored on ice until processing. Granulocytes were isolated as described [18, 37]. In brief, 200 ml blood was diluted 1:3 in Dulbecco’s Phosphate Buffered Saline (DPBS). The blood suspension was incubated with 4% (w/v) dextran-T500 (dilution 5:1, VWR, Hannover, Germany) for 30 min on ice. The upper phase was layered on Ficoll-Paque® (Sigma-Aldrich, Taufkirchen, Germany) and centrifuged (25 min, 300×g, 18 °C). Granulocyte pellets each were resuspended in 500 µl DPBS and erythrocytes were lysed for 40 s with 20 ml ice cold, sterile, endotoxin-free distilled water. The reaction was stopped with 20 ml DPBS (2 × concentrated). After centrifugation (10 min, 300×g, 18 °C), cell pellets were washed with 50 ml DPBS. Granulocytes were resuspended in assay buffer (0.1% bovine serum albumin (BSA) in DPBS, Sigma-Aldrich, Taufkirchen, Germany) to a cell density of 6.25 × 10⁶/ml.

Eosinophil shape change

Granulocytes (80 µl) and assay buffer (10 µl) were incubated in a water bath (5 min, 37 °C). Metabolite solutions (10 µl, final concentration (conc.) 0.01 nM, 0.1 nM, 0.5 nM, 1 nM, 5 nM, 10 nM, 100 nM, 1 µM, 1% DMSO in assay buffer) were added (5 min, 37 °C). The incubation was stopped with 250 µl of 0.25% BD Cell-Fix™ Solution (1:10 with sterile water followed by 1:4 with assay buffer, BD Biosciences, Heidelberg, Germany) and immediate placement on ice (≥ 5 min). Flow cytometric analysis was performed with a Navios 3/10 (Beckman Coulter). Granulocytes were determined by FSC and SSC properties. Eosinophils were discriminated from neutrophils by autofluorescence properties at 560 nm (FL2). 1000 eosinophils were acquired. Eosinophil shape change was calculated as percentage increase of the mean FSC units.

For DP₂ inhibition experiments, granulocytes (80 µl) and fevipiprant (10 µl, final conc. 0.01 nM, 0.05 nM, 0.1 nM, 1 nM, 5 nM, 10 nM, 100 nM, 500 nM, 1 µM, 10 µM) were incubated in a water bath (5 min, 37 °C). Metabolite solution (10 µl, EC₇₀ concentration) was added for 5 min (37 °C). The reaction was stopped and cells were analyzed as described above.

Type 2 innate lymphoid cell isolation

Blood was stored at RT until processing. Whole blood (500 ml) was diluted 1:2 in DPBS. PBMCs were isolated using Ficoll-Paque® and SepMate-50 PBMC Isolation tubes (STEMCELL Technologies, Grenoble, France) following manufacturer’s protocol. PBMCs were pooled, washed with 50 ml DPBS and centrifuged (5 min, 300×g, 4 °C). T cells, B cells and monocytes depletion enriched ILC2s using CD3, CD14 and CD19 MACS separation beads (Miltenyi Biotech, Bergisch Gladbach, Germany) and LD columns following manufacturer’s protocol. Enriched cells were centrifuged (5 min, 300×g, 4 °C) and resuspended in staining buffer containing 5% fetal bovine serum (FBS) and 2 mM ethylenediaminetetraacetic acid (EDTA) in DPBS. Cells were stained for 15 min at RT with a PerCP-Cy5.5-labeled lineage cocktail (CD4, CD8, CD14, CD16, CD19, CD34, CD123, FcεRI), CD11b-FITC, CD56-FITC, CD3-BV510, CD127-BV421, CD45-Alexa Fluor 700 and CD294-PE. View supplement for more detailed information. CD45+, Lineage-, CD11b-, CD56-, CD3-, CD127+, CD294+ cells were sorted with an FACS ARIA Fusion (BD Bioscience) into 96 U bottom well plates (Corning, Amsterdam, Netherlands).

Type 2 innate lymphoid cell culture

Sorted ILC2s (100 cells/well) were expanded with human feeder PBMCs (100,000/well; 37 °C, 5% CO₂) for three to five weeks. Culture medium contained RPMI 1640 Glutamax medium (Life Technologies, Darmstadt, Germany), 1% Pen/Strep (Life Technologies, Darmstadt, Germany), 10% h.i. human AB serum (Sigma-Aldrich, Taufkirchen, Germany), and 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Lonza, Wakersville, USA). The medium was supplemented with 100 U/ml rh-IL-2 (Life Technologies, Darmstadt, Germany), 25 ng/ml rh-IL-4 (Miltenyi Biotech, Bergisch Gladbach, Germany), 5 µg/ml phytohemagglutinin-M (PHA-M; Sigma-Aldrich, Taufkirchen, Germany).

Cell migration assay

Migration was assessed using 5.0 µm pore size, polycarbonate membrane, polystyrene 96-transwell plates (Corning, Amsterdam, Netherlands; Hölzel, Köln, Germany) following manufacturer’s protocols. Metabolites (lower well, final conc. 5 nM, 10 nM, 50 nM, 100 nM, 500 nM, 1 µM, 2 µM, 3 µM and 5 µM) and cells (upper well, ≤ 100,000/well) were incubated for 6 h (37 °C, 5% CO₂) in culture medium without IL-2, IL-4 and PHA. Migrated cells (25 µl) were incubated with CellTiter-Glo® Luminescence Cell Viability Assay (Promega, Mannheim, Germany) following manufacturer’s protocol. Luminescence was measured using a Tecan infinite F200 pro.

For DP₂ inhibition experiments, ILC2s were incubated with fevipiprant (final conc. 0.01 nM, 0.1 nM, 0.5 nM, 1 nM, 5 nM, 10 nM, 100 nM, 1 µM, 10 µM, for 1 h, 37 °C, 5% CO₂) in culture medium without IL-2, IL-4 and PHA and cell migration was measured as described above using the EC₇₀ concentration of respective metabolites.

Cytokine measurement

Cells (≤ 150,000/well) and metabolite solutions (final conc. 2.5 nM, 5 nM, 25 nM, 50 nM, 250 nM, 500 nM, 1 µM, 1.5 µM, 2.5 µM) were incubated for 24 h in culture medium without IL-2, IL-4 and PHA (U-bottom 96-well plates, 37 °C, 5% CO₂). After centrifugation (5 min, 300×g, RT), the supernatant was collected and stored at − 80 °C until measurement.

For DP₂ inhibition experiments, ILC2s were incubated with fevipiprant (final conc. 0.01 nM, 0.1 nM, 0.5 nM, 1 nM, 5 nM, 10 nM, 50 nM, 100 nM, 1 µM) for 1 h (37 °C, 5% CO₂) and afterwards with the EC₇₀ concentration of respective metabolites for 24 h. Cells were incubated as described above.

Supernatants were diluted 1:10 and IL-5 and IL-13 concentrations were measured by MSD immunoassays (V-PLEX Meso Dale Discovery, Rockville, MD, USA) following manufacturer’s protocol. Raw values are depicted in Additional file 1: Figs. S2 and S3.

Statistics

GraphPad Prism 9.0.1 was used for analysis. Four parameter non-linear agonist or inhibitor regression models were used fit curves and to calculate EC₅₀/IC₅₀ values, respectively. Fitting curves constraints are indicated in respective figure legends.

Maximal PGD₂-induced responses were compared with respective maximal responses of metabolites using paired One-Way ANOVA with post-hoc Dunn’s multiple comparisons tests.

Results

PGD₂ metabolites induce eosinophil shape change with similar potencies

PGD₂ and seven PGD₂ metabolites were evaluated for their ability to induce activation of eosinophils by measurement of cellular shape changes. All prostaglandins induced a concentration dependent shape change with nanomolar potency except for 9α,11β-PGF₂ which was less potent (Additional file 1: Fig. S1). EC₅₀ values of prostaglandins, described as mean ± standard error of the mean (SEM), fell into following rank order: PGD₂: 0.7 ± 0.2 nM < Δ¹²-PGD₂: 1.2 ± 1.8 nM < 15-deoxy-Δ^12,14-PGD₂: 1.5 ± 1.6 nM < PGJ₂: 1.6 ± 3.8 nM < DK-PGD₂: 2.7 ± 2.3 nM < Δ¹²-PGJ₂: 5.6 ± 1.0 nM < 15-deoxy-Δ^12,14-PGJ₂: 12.0 ± 0.7 nM < 9α,11β-PGF₂: > 1000. Similar EC₅₀ values were reported by Sandig et al. but with different rank order (Table 1) [19].

Table 1.

EC₅₀ values of PG-induced eosinophil shape change were calculated and are given as mean ± standard error of the mean (SEM), n = 2–3

Eosinophil shape change EC50 [nM]
Metabolite	Measured values	Sandig et al. [19]
PGD2	0.7 ± 0.2	0.4
DK-PGD2	2.7 ± 2.3	1.1
∆12-PGD2	1.2 ± 1.8	7.3
15-deoxy-∆12,14-PGD2	1.5 ± 1.6	2.4
9α,11β-PGF2	> 1000	156.0
PGJ2	1.6 ± 3.8	2.2
∆12-PGJ2	5.6 ± 1.0	3.7
15-deoxy-∆12,14-PGJ2	12.0 ± 0.7	8.4

Published reference values are given in the right column

Metabolites induced shape changes with similar maximal responses as compared to PGD₂, which was observed by a comparison of FSC_max values normalized on 100% PGD₂ shape change. Activation responses of PGD₂ metabolites ranged between 98.0 and 102.9%, compared to the parent PGD₂ (Fig. 2). Sandig et al. did not compare maximal responses but their results indicate similar responses for most metabolites as well while responses of Δ¹²-PGD₂ and Δ¹²-PGJ₂ seemed to have higher FCS values.

Fig. 2.

Comparison of maximal eosinophil shape change in the presence of PGD₂ and selected metabolites. The mean forward scatter values in the presence of the highest PGD₂ concentration was normalized to 100% shape change, n = 6–8. Values are given as mean ± SD

PGD₂ metabolites induce ILC2 migration and IL-5 and IL-13 cytokine secretion

Next, PGD₂ metabolites were evaluated for their potential to induce ILC2 migration and Type 2 cytokine secretion. The metabolites induced cell migration concentration dependently and with nanomolar potency except for 9α,11β-PGF₂. 9α,11β-PGF₂ increased migration in comparison to the medium control but dose independently (Fig. 3A). Migratory potencies were increased for the D-series of metabolites in comparison to the J-series with mean EC₅₀ ± SEM values of PGD₂ 17.4 ± 3.9 nM, DK-PGD₂ 14.2 ± 3.4 nM, ∆¹²-PGD₂ 19.3 ± 3.2 nM, 15-deoxy-∆^12,14-PGD₂ 21.8 ± 6.3 nM, PGJ₂ 66.3 ± 7.0 nM, ∆¹²-PGJ₂ 91.7 ± 9.2 nM and 15-deoxy-∆^12,14-PGJ₂ 38.1 ± 5.4 nM (Table 2). Micromolar metabolite concentrations abolished cell migration.

Fig. 3.

A ILC2 migration in the presence of ascending concentrations of PGD₂ and selected metabolites, n = 3. Values are given as mean ± SD. Nonlinear curve fit with constraints of bottom constant equal to zero and top constant equal to 100 was applied to derive EC₅₀ and EC₇₀. B Maximal cell migration of PGD₂ and selected metabolites. Absolute numbers of migrated cells were compared and normalized to 100% PGD₂, n = 7. Values are given as mean ± SD. DK-PGD₂: ns, ∆¹²-PGD₂: ns, 15-deoxy-∆^12,14-PGD₂: p = 0.05, 9α,11β-PGF₂: ns, PGJ₂: ns, ∆¹²-PGJ₂: p = 0.03 and 15-deoxy-∆^12,14-PGJ₂: p = 0.01. *p < 0.05

Table 2.

EC₅₀ values of PG-induced ILC2 migration and IL-5 and IL-13 cytokine secretion were calculated and are given as mean ± standard error of the mean (SEM), n = 3

ILC2 migration and cytokine secretion EC50 [nM]
Metabolite	Migration	IL-5 secretion	IL-13 secretion
PGD2	17.4 ± 3.9	139.2 ± 15.5	131.9 ± 10.8
DK-PGD2	14.2 ± 3.4	146.0 ± 18.0	159.2 ± 28.8
∆12-PGD2	19.3 ± 3.2	108.1 ± 11.8	125.2 ± 8.3
15-deoxy-∆12,14-PGD2	21.8 ± 6.3	175.4 ± 26.4	159.1 ± 13.9
9α,11β-PGF2	–a	526.9 ± 12,926.7	788.3 ± 173.7
PGJ2	66.3 ± 7.0	244.4 ± 26.3	321.9 ± 43.8
∆12-PGJ2	91.7 ± 9.2	234.1 ± 30.6	343.2 ± 52.4
15-deoxy-∆12,14-PGJ2	38.1 ± 5.4	185.3 ± 21.5	226.2 ± 35.2

^aCurve fit was ambiguous, EC₅₀ value could not be determined

The maximal migration response measured by total number of migrated cells compared to PGD₂ was increased for ∆¹²-PGD₂, 15-deoxy-∆^12,14-PGD₂, PGJ₂, ∆¹²-PGJ₂ and 15-deoxy-∆^12,14-PGJ₂. In contrast DK-PGD₂ and 9α,11β-PGF₂ showed a non-significant trend to reduced maximal migration responses, (Fig. 3B).

IL-5 and IL-13 cytokine secretion of ILC2s was induced by PGD₂ metabolites in a concentration dependent manner. ILC2 responses were comparable between IL-5 and IL-13 (Fig. 4A). Similar to cell migration, higher potencies for cytokine secretion were found for the D-series of metabolites compared to the J-series (Table 2). 9α,11β-PGF₂ was able to induce cytokine secretion although less potent than the other metabolites and with high variability. The maximal cytokine secretion response was significantly lower in the presence of 9α,11β-PGF₂ and PGJ₂ compared to PGD₂ while the maximal response was similar for DK-PGD₂, ∆¹²-PGJ₂ and 15-deoxy-∆^12,14-PGJ₂. Cytokine secretion was non-significantly enhanced for ∆¹²-PGD₂ and 15-deoxy-∆^12,14-PGD₂ compared to PGD₂ (Fig. 4B).

Fig. 4.

A IL-5 (white circles) and IL-13 (black squares) cytokine secretion of ILC2s in the presence of ascending concentrations of PGD₂ and selected metabolites, n = 3. Values are given as mean ± SD. Nonlinear curve fit with constraints of bottom constant equal to zero and top constant equal to 100 was applied to derive EC₅₀ and EC₇₀. B Maximal cytokine secretion of selected metabolites compared to PGD₂. The absolute concentrations of secreted cytokines were compared and normalized to 100% PGD₂, n = 6. Values are given as mean ± SD. 9α,11β-PGF₂: IL-5 p = 0.006, IL-13 p = 0.02, PGJ₂: IL-5 p = 0.03, IL-13 p = 0.007. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Selective DP₂ inhibition of eosinophils and ILC2s

Isolated whole blood eosinophils were incubated with the selective DP₂ antagonist fevipiprant prior to activation of cells with EC₇₀ concentrations of PGD₂ or respective metabolites (supplementary table 2). Fevipiprant inhibited eosinophil shape changes in dose dependent manners with sub-nanomolar potencies (Fig. 5A and B). IC₅₀ values against metabolites were comparable and ranged between 0.1 and 0.9 nM for PGD₂ and six metabolites (Table 3). Eosinophil shape changes induced by 15-deoxy-∆^12,14-PGJ₂ could be inhibited by fevipiprant, however, the dose–response was not sigmoidal.

Fig. 5.

Inhibition of PGD₂ and metabolite-induced cell activation with DP₂ inhibitor fevipiprant. Cells were incubated with ascending concentrations of fevipiprant and stimulated with EC₇₀ concentrations of PGD₂, DK-PGD₂, ∆¹²-PGD₂, 15-deoxy-∆^12,14-PGD₂, PGJ₂, 9a,11b-PGF₂, ∆¹²-PGJ₂ and 15-deoxy-∆^12,14-PGJ₂, respectively. A, B Inhibition of eosinophil shape change, n = 3–5. Top constraints equal 100. C, D Inhibition of ILC2 migration, n = 4. Top constraints equal 100. E, F Inhibition of ILC2 IL-5 secretion, n = 4. Bottom constraints equal zero. (G,H) Inhibition of ILC2 IL-13 secretion, n = 4. Bottom constraints equal zero. Values are given as mean ± SEM

Table 3.

IC₅₀ values of fevipiprant-inhibited eosinophil shape change, ILC2 migration and IL-5 and IL-13 cytokine secretion of ILC2s

IC50 fevipiprant [nM]
Metabolite	shape change (n = 3–5)	Migration (n = 4)	IL-5 (n = 4)	IL-13 (n = 4)
PGD2	0.9 ± 0.2	2.4 ± 0.6	4.2 ± 0.8	5.1 ± 1.1
DK-PGD2	0.1 ± 0.0	1.9 ± 0.7	6.5 ± 1.6	8.5 ± 2.1
∆12-PGD2	0.1 ± 0.0	4.9 ± 1.3	4.8 ± 1.2	4.8 ± 1.3
15-deoxy-∆12,14-PGD2	0.5 ± 0.3	4.6 ± 2.0	4.5 ± 1.7	4.6 ± 1.4
9α,11β-PGF2	0.9 ± 0.1	10.9 ± 7.8	–a	–a
PGJ2	0.5 ± 0.2	3.7 ± 0.6	4.0 ± 1.0	5.9 ± 3.6
∆12-PGJ2	0.6 ± 0.3	3.1 ± 0.5	3.7 ± 0.9	4.6 ± 1.6
15-deoxy-∆12,14-PGJ2	–a	3.8 ± 0.8	2.3 ± 0.9	2.6 ± 0.9

Values were calculated from 3 to 5 experiments and are given as mean ± standard error of the mean (SEM)

^aCurve fit was ambiguous, IC₅₀ values could not be determined

Similar to eosinophil activation, ILC2 migration (Fig. 5C and D) and cytokine secretion (Fig. 5E–H) could be inhibited by fevipiprant in a concentration dependent manner. In comparison to eosinophils, ILC2 related IC₅₀ values were approximately one order of magnitude higher. Migration IC₅₀ values ranged between 1.9 and 10.9 nM, IL-5 IC₅₀ values ranged between 2.3 and 6.5 nM, and IL-13 IC₅₀ values ranged between 2.6 and 8.5 nM (Table 3). 9a,11b-PGF₂ induced cytokine secretion results were highly variable among donors and IC₅₀ values could not be determined.

Discussion

In vivo and in vitro, PGD₂ is rapidly degraded to various metabolites which presumably contribute to DP₂ immune cell activation in large part due to their DP₂ selectivity [15, 38]. Here, we show for the first time that ILC2s respond to PGD₂ metabolites by migration and IL-5 and IL-13 secretion. We determined respective EC₅₀ values, quantified cellular responses of eosinophils and ILC2s and compared the activation maximal responses with respective PGD₂ responses. DP₂ dependency was demonstrated using fevipiprant which abolished cell activities with nanomolar potencies.

The influence of prostaglandin metabolites on eosinophil cell activation and shape change is well known. Sandig et al. first reported EC₅₀ values, which we could largely confirm [19]. EC₅₀ concentrations resulted in different rank orders of potencies compared to ours, however, those differences were negligibly small. The potency of 9α,11β-PGF₂ was clearly reduced in both studies compared to other metabolites. Since 9α,11β-PGF₂ has an approximately 130 fold reduced binding affinity to DP₂ compared to PGD₂ [38], it can be assumed that the concentrations used were too low to sufficiently induce a shape changes in eosinophils. Merely concentrations of 1 µM demonstrated a shift in FCS properties (Additional file 1: Fig. S1). Comparing the maximal changes in the forward scatters indicated that all metabolites induced shape changes with similar maximal responses. This may imply their comparable agonistic function on eosinophil activation. This finding could be supported by further eosinophil activation data addressing e.g. cell surface CD11b expression or degranulation.

The selective relevance of DP₂ for eosinophil activation has been reported for PGD₂ and DK-PGD₂ using various DP₂ antagonists [19, 39, 40]. Nevertheless, blocking the activation of other metabolites was described only using the dual DP₂ and thromboxane receptor (TP) antagonist ramatroban [19]. We complemented data affirming that metabolites induced eosinophil shape changes via DP₂ selectively using fevipiprant [41].

Less is known about the role of PGD₂ metabolites on ILC2 activation. Recently it was shown that ILC2s produce endogenous PGD₂ upon alarmin activation which was metabolized to PGJ₂, ∆¹²-PGJ_2, DK-PGD₂ and 15-deoxy-∆^12,14-PGJ₂ in culture supernatants [28]. However, so far only exogeneous and endogeneous PGD₂ was reported to induce migration and IL-5 and IL-13 secretion in ILC2s [26–28]. Here, we show that ILC2s also migrate and release Type 2 cytokines in presence of DK-PGD₂, PGJ₂, Δ¹²-PGJ₂, Δ¹²-PGD₂, 15-deoxy-Δ^12,14-PGJ₂, and 15-deoxy-Δ^12,14-PGD₂. D-series metabolites were more potent than those of the J-series. 9α,11β-PGF₂ was not able to induce ILC2 migration in the concentration range used. The comparably low binding affinity of 9α,11β-PGF₂ to DP₂ might have resulted in the low response. Varying activating potencies among the other metabolites might be related to differences in respective DP₂ binding affinities as well [15, 38], however, EC₅₀ rank orders of shape change, migration and cytokine release experiments were different. Moreover, quantifying the maximal assay responses among metabolites revealed that ILC2 cell migration was enhanced in presence of most metabolites compared to PGD₂ while cytokine production was not enhanced. Interestingly, cell migration was strongest for the J-series metabolites Δ¹²-PGJ₂, 15-deoxy-Δ^12,14-PGJ₂ and cytokine secretion was strongest for the D-series metabolites Δ¹²-PGD₂, 15-deoxy-Δ^12,14-PGD₂. Our data may indicate that PGD₂ metabolites contribute predominantly to ILC2 migration rather than cytokine secretion which needs to be confirmed in vivo. Although it is known that numbers of eosinophils and ILC2s are elevated in the airways of severe asthmatic patients [24] little is known about local distributions of PGD₂ metabolite concentrations in human tissues. ILC2 numbers and PGD₂ concentrations in the bronchoalveolar lavage (BAL) fluid were reported to correlate after allergen challenge of mild asthmatics [25]. Moreover, metabolic products of PGD₂ were found in human serum [42], plasma, BAL [43] and urine [44, 45] samples, however, their contribution to the asthmatic disease is poorly understood [20]. Studies on eosinophils in mice showed that ∆¹²-PGJ₂ mobilizes eosinophils from the bone marrow and facilitates cell extravasation [18] which fits with the observation that PGD₂ is preferably converted to ∆¹²-PGD₂ and ∆¹²-PGJ₂ in blood plasma [12]. The contribution of metabolites to cell recruitment might be of biological relevance but remains speculative without further experiments. 9α,11β-PGF₂ is another major degradation products found in human and showed low agonistic potency on eosinophils and ILC2s [42, 44, 46]. At micromolar concentrations, 9α,11β-PGF₂ was able to activate cells, which probably do not represent physiological concentrations of the metabolite. Further, quantification of ILC2 responses indicated a reduced maximal response compared to PGD₂. Therefore, it seems more likely that 9α,11β-PGF₂ does not contribute to DP₂ signaling in vivo but rather serves as an signal-inactivating degradation product.

Blocking with fevipiprant confirmed the DP₂ dependency of ILC2 activities which were comparable to previous experiments performed with PGD₂ [26]. Inhibition of 15-deoxy-Δ^12,14-PGJ₂ activation in eosinophils followed a non-sigmoidal dose–response principle which should be assessed more closely in future studies. Inhibiting PGD₂ signaling has been studied in a wide range of clinical studies due to its key role in allergic inflammation. While fevipiprant itself is no longer being developed for asthma, our findings may be of relevance to other DP₂ antagonists which remain under clinical investigation [22, 47].

Conclusion

PGD₂ metabolites are effective DP₂ agonists and promote eosinophil shape change, ILC2 cell migration and Type 2 cytokine secretion in vitro. They might contribute to ILC2 recruitment in vivo since predominantly ILC2 migration but not ILC2 cytokine secretion or eosinophil shape change was enhanced compared to PGD₂.

Abbreviations

BAL

Bronchoalveolar lavage

BMI

Body mass index

BSA

Bovine serum albumin

CD

Cluster of differentiation

Conc.

Concentration

CRTH2

Chemoattractant receptor-homologous molecule expressed on Th2 cells

DK-PGD₂

13,14-Dihydro-15-keto-prostaglandin D₂

DMSO

Dimethylsulfoxide

DP

D-prostanoid receptor

DPBS

Dulbecco’s Phosphate Buffered Saline

EC₅₀

Half maximal effective concentration

EDTA

Ethylenediaminetetraacetic acid

FSC

Forward scatter

HEPES

4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid

IC₅₀

Half maximal inhibitory concentration

IL

Interleukin

ILC2

Type 2 innate lymphoid cells

MFI

Mean fluorescence intensity

PG

Prostaglandin

PHA-M

Phytohemagglutinin-M

SD

Standard deviation

SEM

Standard error of mean

SSC

Sideward scatter

Authors' contributions

All authors added and revised the manuscript. SC: Data curation, formal analysis, methodology, visualization, writing original draft, writing review and editing; CG: Data curation, formal analysis, methodology, visualization, writing review and editing; VJE: Conceptualization, project administration, writing review and editing; SDK: Conceptualization, project administration, writing review and editing; JMH: Conceptualization, writing review and editing; DAS: Conceptualization, project administration, writing review and editing; MM: Conceptualization, writing review and editing. All authors read and approved the final manuscript.

Funding

Open Access funding enabled and organized by Projekt DEAL. This research was supported by Novartis Pharma AG and was performed under a collaboration agreement between Novartis and Fraunhofer ITEM.

Availability of data and materials

The datasets supporting the conclusion of this article are available on reasonable request from MM and DAS.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Hannover Medical School (7882_BO_S_2018). All subjects gave written informed consent.

Consent for publication

Not applicable.

Competing interests

SC’s, CG’s, JMH’s and MM’s institution received funding from ALK, Allergopharma, Astellas, AstraZeneca, GlaxoSmithKline, Janssen Pharmaceutica NV, LETI, Novartis, and Sanofi-Aventis for clinical trial conduct outside the submitted work. JMH reports personal fees from Boehringer Ingelheim, CSL Behring, HAL, Merck, and Novartis for consultancy and lectures outside the submitted work. VJE and SK were employees of Novartis at the time of the study. DAS is an employee of Novartis.