Opposing roles of Prostaglandin D₂ receptors in ulcerative colitis

Abstract

Pro-resolution functions were reported for Prostaglandin D₂ (PGD₂) in colitis, but the role of its two receptors, DP and in particular CRTH2 are less well defined. We investigated DP and CRTH2 expression and function during human and murine ulcerative colitis (UC). Expression of receptors was measured by flow cytometry on peripheral blood leukocytes, and by immunohistochemistry and immunoblotting in colon biopsies of patients with active UC and healthy individuals. Receptor involvement in UC was evaluated in a mouse model of DSS colitis. DP and CRTH2 expression changed in leukocytes of patients with active UC in a differential manner. In UC patients, DP showed higher expression in neutrophils but lower in monocytes as compared to control subjects. In contrast, CRTH2 was decreased in eosinophils, NK and CD3⁺ T cells but not in monocytes and CD3⁺/CD4⁺ T cells. The decrease of CRTH2 on blood eosinophils clearly correlated with disease activity. DP correlated positively with disease activity in eosinophils but inversely in neutrophils. CRTH2 internalized upon treatment with PGD₂ and 11-dehydroTXB₂ in eosinophils of controls. Biopsies of UC patients revealed an increase of CRTH2-positive cells in the colonic mucosa and high CRTH2 protein content. The CRTH2 antagonist CAY10595 improved while the DP antagonist MK0524 worsened inflammation in murine colitis. DP and CRTH2 play differential roles in UC. Although expression of CRTH2 on blood leukocytes is downregulated in UC, CRTH2 is present in colon tissue where it may contribute to inflammation whereas DP likely promotes anti-inflammatory actions.

Introduction

Inflammatory bowel diseases (IBD), such as ulcerative colitis (UC), are chronic inflammatory conditions of the intestine resulting from an inappropriate immune response (1). Prostaglandins and their receptors have been shown to play important roles in the pathogenesis of IBD (2). However, their actions are divers and complex. Part of this complexity lies in the expression of different receptors for each prostaglandin and in the fact that prostaglandins are rapidly metabolized, a process in which the metabolite can gain new specificity as a ligand (3-5). The exact task of prostaglandins during inflammation (pro- and/or anti-inflammatory) is sometimes hard to define as their actions are cell specific and depend on activation of multiple receptors, cross-reactivity and coupling to different signal transduction pathways (6).

Prostaglandins and thromboxanes are formed by cyclooxygenase (COX) enzymes from arachidonic acid. COX-derived mediators are thought to play a beneficial role in IBD, because inhibition of COX-2 by NSAIDs has been shown to dampen healing and resolution of colitis (7). Prostaglandin D₂ (PGD₂) has been categorized as an anti-inflammatory mediator because it caused reduction of granulocyte infiltration during experimental colitis (8). Levels of PGD₂ in colon biopsies of patients during remission of UC are increased (9). The ameliorating effect of PGD₂ in experimental colitis is thought to be mediated by activation of the D-type prostanoid (DP) receptor (now named DP1 receptor) (8); however, the role of the other PGD₂ receptor, CRTH2 (chemoattractant receptor homologous molecule expressed on T-helper type 2 cells; now also termed DP2) in UC is still unclear. Interestingly, while PGD₂ is regarded as anti-inflammatory in colitis it promotes carcinogenesis after resolution of colitis (10). A recent study reported increased expression of the lipocalin-type prostaglandin D synthase (L-PGDS), the PGD₂-producing enzyme, in UC (11). The expression correlated with disease activity challenging the role of PGD₂ as a mediator with purely anti-inflammatory action (11).

The therapeutic potential of the two PGD₂ receptors in inflammatory processes, especially that of CRTH2, has been only recently discovered (12). CRTH2 may not only be activated by PGD₂, but also by its stable metabolites and even by a thromboxane metabolite (5, 13). CRTH2 has been found in leukocytes, such as human and mouse Th2 cells, eosinophils, basophils and monocytes (14-16), and it selectively induces chemotaxis in these cells (3, 17, 18). Human peripheral blood neutrophils lack expression of the CRTH2 receptor (19, 20) despite the detection of PGD₂ binding sites (21) which are most likely DP receptors through which PGD₂ inhibits their activity and chemotaxis (19, 22). Human monocytes and eosinophils both express DP and CRTH2 receptors and have shown chemotaxis upon activation with PGD₂ (15, 22).

It is known that the development of IBD is associated with dense infiltration of leukocytes, such as neutrophils, T cells, eosinophils and macrophages in the colon (23). Eosinophils, in particular, show prominent expression of DP and CRTH2 (22) and these two PGD₂ receptors likely cooperate in eosinophil responses such as Ca²⁺signaling, chemotaxis or mobilization from bone marrow (22, 24). Eosinophils were therefore investigated in more detail in the present study. Although the role of eosinophils in UC is not yet clarified, there is some evidence that they may play a pro-inflammatory role (25). Several reports describe that the gut mucosa of IBD patients is densely infiltrated with eosinophils (26, 27) and more recent studies show that eosinophil-selective chemokines and eosinophil peroxidase promote inflammation in human and experimental IBD (28, 29).

Because of its pro-inflammatory role in allergic inflammation we intended to elucidate whether CRTH2 would act similarly in human and experimental UC thus behaving differently to the anti-inflammatory DP. Expression of DP and CRTH2 were investigated in blood leukocytes, especially in eosinophils, and also in colon biopsies of human UC patients and healthy individuals. We further wanted to assess the role of these receptors in a mouse model of experimental IBD by applying selective antagonists against each receptor. We observed that the two receptors are in fact differentially expressed in human UC and that their expression in eosinophils correlates inversely with disease severity. Contrary to DP, our results in the experimental model suggest that CRTH2 plays a different, pro-inflammatory, role in IBD.

Materials and Methods

Patients

Blood was collected from adult patients with confirmed active UC (n = 13; mean age: 37.3 ± 14.4; 10 males/3 females), from healthy controls subjects (n = 12; mean age: 37.4 ± 21.2; 6 males/6 females; Table I) and from adult UC patients in remission (n = 8; mean age: 38.6 ± 14.2; 5 males/3 females; Table II). UC patients were recruited from the IBD clinic of the Department of Internal Medicine at the Medical University of Graz, control subjects were recruited from healthy volunteers. Diagnosis of UC was established by standard clinical, endoscopic and histologic criteria (30). All UC patients with active disease also underwent contemporaneous colonoscopy to assess endoscopic disease activity. For UC patients, disease activity was assessed by using different clinical activity scores, i.e. the Clinical Activity Index (CAI, also termed Rachmilewitz Index) and the total Mayo score which combines clinical and endoscopic activity scores (partial and endoscopy Mayo subscore) (31, 32). All subjects suffering from UC, except for one, were on some sort of medication (Table I). Blood was collected in Vacuette® sodium citrate blood tubes (Greiner-Bio-One, Austria) and immediately processed for flow cytometric experiments. For the measurement of prostaglandins, blood was collected in Vacuette ® serum tubes (Greiner-Bio-One, Austria), and frozen at −80°C until use.

Table I. Characteristics of patients with UC.

UC patients	Mayo score	CAI	Ongoing treatment, response and duration of treatment	Previous treatment	Disease duration (in years)	Montreal classification
1	9 (moderate)	4	5-ASA (PR, 2y) Corticosteroids (PR, 5y)	AZA (NR), MTX (NR), Adalimumab (NR), Infliximab (NR, I)	8	E2
2	4 (mild)	2	AZA (PR, 10y) E. coli Nissle 1917 (PR, 1y)	5-ASA (I), Corticosteroids (R)	11	E3
3	3 (mild)	2	5-ASA (PR, 3y) MTX (R, 1y)	Corticosteroids (R), AZA (I)	4	E3
4	5 (mild)	1	5-ASA (PR, 1.5y) AZA (R, 6y)	Corticosteroids (R)	12	E2
5	4 (mild)	2	—	5-ASA (PR), Corticosteroids (NR), AZA (R), Tacrolimus (R)	10	E3
6	10 (moderate)	8	5-ASA (PR >10y) Adalimumab# (PR, 5y) Antibiotics (R, 3days)	Corticosteroids (R), Infliximab (I)	21	E3
7	6 (moderate)	4	Corticosteroids (2mo)	5-ASA (NR), Corticosteroids (R), AZA (I), Adalimumab (NR)	2	E2
8	10 (moderate)	5	5-ASA (NR, 4mo) AZA (NR, 4mo) Tacrolimus (PR, 4mo)	Corticosteroids (NR)	2	E3
9	10 (moderate)	8	5-ASA (PR, >10y)	Corticosteroids (R)	16	E2
10	8 (moderate)	11	5-ASA (R, 1.5y)	-	2	E3
11	9 (moderate)	6	5-ASA (NR 0,5y)	Corticosteroids (R)	1	E2
12	10 (moderate)	12	5-ASA (PR, 7days)	-	0	E3
13	9 (moderate)	6	5-ASA (PR, 1y)	-	2	E3
14*	12 (severe)	9	5-ASA (NR 1mo) Corticosteroids (NR 14 days) Antibiotics (NR, 3 days)	-	4	E3
15*	9 (moderate)	5	5-ASA (PR, 2y) AZA (PR, 2y)	Corticosteroids (R)	2	E2
16*	6 (moderate)	4	Corticosteroids (PR, 2mo)	AZA (I), Infliximab (NR)	2	E2
17*	7 (moderate)	6	5-ASA (PR, 21 days) Corticosteroids (R 21 days)	-	0	E2
18*	5 (mild)	3	5-ASA (R, > 10y)	Corticosteroids (R)	14	E2

UC, ulcerative colitis; 5-ASA, 5-aminosalicylic acid; AZA, azathioprine; MTX, methotrexate; CAI, Clinical activity index (also termed Rachmilewitz index)

^*Samples 14-18 were used for Western Blots only;

^#Indication for therapy: ankylosing spondylitis

R (response), PR (partial response), NR (no response), I (intolerance), y (years), mo (months)

Table II. Characteristics of UC patients in remission.

UC patient in remission	Mayo score	Partial Mayo Score	CAI	Ongoing treatment, response and duration of treatment	Previous treatment	Disease duration (in years)	Montreal classification
1	0	0	0	5-ASA (R, 5y)	Corticosteroids (R)	12	E3
2	0	0	3	5-ASA (R, >10y) Corticosteroids (R, 2y)	AZA (I) Infliximab (I) Adalimumab (I) MTX (I)	20	E3
3	2	1	0	5-ASA (PR, 7y) Adalimumab (R, 4y)	Corticosteroids (R) AZA (I)	18	E2
4	1	0	1	5-ASA (R, 0.5y)	none	1	E1
5	*	0	0	5-ASA (R, 1yr)	Corticosteroids (R)	1	E1
6	*	0	0	5-ASA (R, >10y )	Corticosteroids (R) AZA (PR)	17	E2
7	*	0	0	5-ASA (PR, 7y) Infliximab (R, 1y) AZA (PR, 6y)	-	8	E2
8	0	1	1	5-ASA (PR, >10y) AZA (R, 6y)	Corticosteroids (R)	20	E3

UC, ulcerative colitis; 5-ASA, 5-aminosalicylic acid; AZA, azathioprine; MTX, methotrexate; CAI, Clinical activity index (also termed Rachmilewitz index)

R (response), PR (partial response), NR (no response), I (intolerance), y (years), mo (months)

^*only partial Mayo score available (no colonoscopy performed)

For colonic tissue samples, UC patients and control subjects included in the study were recruited from the endoscopy unit of the Department of Internal Medicine, Medical University of Graz. Colonoscopy in UC patients was performed as part of the clinical workup because of active disease (patients 14-18; Table I). Biopsies from control subjects were obtained from individuals undergoing colonoscopy as part of the colon cancer screening program and from patients with diagnostic workup of occult or overt gastrointestinal bleeding, with a normal colonoscopy. Subjects with significant comorbidities and individuals with an intercurrent illness as infections or pregnant women were excluded from the study. Control subjects that were taking NSAIDs including acetylsalicylic acid were also excluded from the study. Biopsies were taken from the sigmoid colon, 30 cm proximal to the anal canal and were either immediately fixed in 10% phosphate buffered formalin for immunohistochemistry or immediately frozen for Western blot experiments. The study was approved by the Ethics Committee of the Medical University of Graz (Protocol number: 23-002 ex 10/11 and 24-281 ex 11/12) and all participants provided written, informed consent.

Human peripheral blood cell populations

Citrated whole blood from patients and healthy individuals was lysed with 1x BD FACS Lysing solution. Samples were washed once in PBS and resuspended in Antibody Diluent (Dako; Glostrup, Denmark). Cells were stained with FITC-conjugated CD4 Ab (1:25; Miltenyi Biotec; Bergisch Gladbach, Germany), PE-conjugated CD56 Ab (1:25) PerCP-conjugated CD14 Ab (1:25) and PE-Cy5-conjugated CD3 Ab (1:25) (all antibodies from BD Biosciences; San Jose, CA, USA) for 30 min at 4°C. Samples were washed once in PBS, fixative solution was added and samples were kept on ice until analyzed on a FACSCalibur flow cytometer. Peripheral blood cell populations were identified on the basis of forward versus side scatter parameters and specific Ab binding. Eosinophils were distinguished from neutrophils in an unstained sample according to granularity (side scatter) and by their autofluorescence. Monocytes were identified as CD14⁺, NK-cells as CD56⁺/CD3⁻ and helper T-cells as CD3⁺/CD4⁺ population. Cell populations were quantified as percent of total peripheral blood cells. Fixative solution was prepared by adding 9 ml of distilled water and 30 ml of FACS-Flow to 1 ml of CellFix (BD Biosciences; San Jose, CA, USA).

Flow cytometric analysis of CRTH2 and DP receptor expression in whole blood

Citrated whole blood from patients with active UC or healthy individuals was lysed with 1x BD FACS Lysing solution to remove erythrocytes. Samples were washed once in PBS, resuspended in fixative solution and kept on ice for 10 min. After washing with PBS, cells were blocked with Ultra-V Block (Lonza; Allendale, NJ, USA) for 30 min at 4°C. For flow cytometric visualization of the CRTH2 receptor, cells were stained with Alexa Fluor 647-conjugated rat anti-human CRTH2 Ab or Alexa Fluor 647-conjugated rat IgG2a isotype control (10 μg/ml; BD Biosciences; San Jose, CA, USA) for 30 min at 4°C. For flow cytometric visualization of the DP receptor, cells were incubated with goat polyclonal Ab against DP1 (20 μg/ml; Santa Cruz Biotechnology, Dallas, TX, USA), or normal goat IgG (20 μg/ml; Sigma Aldrich, St. Louis, MO, USA) for 30 min at 4°C. Cells were washed in PBS, resuspended in Antibody Diluent (Dako, Glostrup, Denmark) and incubated for 30 min at 4°C with the secondary Ab (1:2000; Alexa Fluor 647-conjungated rabbit anti-goat secondary Ab; Life Technologies, Carlsbad, CA, USA). Finally, all samples were washed with PBS, fixative solution was added and samples were kept on ice until analyzed on a FACSCalibur flow cytometer. Peripheral blood cell populations were identified on the basis of forward versus side scatter parameters and specific Ab binding. CRTH2 and DP receptor expression was recorded as mean fluorescence intensity and expressed as fold increase over isotype control signals. Representative flow plots and gating criteria are shown in supplemental material S1.

Eosinophil CD11b expression

Citrated whole blood from patients or healthy individuals was stained with PE-conjugated anti-CD11b Ab (1:25; Biolegend, San Diego, CA, USA) and PE-Cy5-conjugated CD16 Ab (1:50; BD Biosciences; San Jose, CA, USA). Samples were mixed with agonists and incubated for 20 min at 37°C in a water bath. After incubation, fixative solution was added and samples were kept on ice for 10 min. Erythrocytes were lysed with 1x BD FACS Lysing solution; cells were washed once in PBS and resuspended in fixative solution. CD11b expression on CD16-negative eosinophils was quantified by flow cytometry and expressed as percen of the vehicle response (in the absence of an agonist).

CRTH2 receptor internalization assay in eosinophils

Changes in CRTH2 surface expression in eosinophils were recorded by means of flow cytometry as described previously (24). Citrated whole blood was mixed with agonists and stimulated for 1 h at 37°C in a water bath. To stop the reaction, samples were transferred onto ice for 10 min. Erythrocytes were lysed with 1x BD FACS Lysing solution; cells were washed in PBS and resuspended in Antibody Diluent (Dako, Glostrup, Denmark). Cells were stained with Alexa Fluor 647-conjugated rat anti-human CRTH2 Ab or Alexa Fluor 647-conjugated rat IgG2a isotype control (10 μg/ml; BD Biosciences; San Jose, CA, USA) for 30 min at 4°C. After washing with PBS, fixative solution was added and samples were kept on ice until analyzed on a FACSCalibur flow cytometer. Eosinophils were detected as CD16-negative cells in a higher side scatter region. Data were expressed as percent of vehicle control (unstimulated sample).

Immunofluorescence microscopy of peripheral blood leukocytes

The DP and CRTH2 receptors were visualized in polymorphonuclear leukocyte (PMNL) and peripheral blood mononuclear cell (PBMC) fractions from healthy individuals by immunofluorescence staining. PMNL and PBMC (3× 10⁵) were fixed with 3.7% formaldehyde on ice for 10 min prior to the staining.

For visualization of CRTH2, cells were washed once in PBS and blocked with Ultra-V Block (Thermo Fisher Scientific, Waltham, MA, USA) for 30 min at 4°C. Thereafter, cells were washed once with PBS, resuspended in Antibody Diluent (Dako, Glostrup, Denmark) and incubated with rabbit polyclonal Ab against CRTH2 (10 μg/ml; Acris Antibodies; Herford, Germany), or normal rabbit IgG (10 μg/ml; Santa Cruz Biotechnology, Dallas, TX, USA) for 1 h at 4°C. After washing with PBS, cells were resuspended in Antibody Diluent (Dako) and incubated for 1 h at 4°C with the secondary Ab (1:1000; Cy3-conjungated donkey anti-rabbit secondary antibody). For visualization of the DP receptor, cells were washed once in PBS and blocked with 2% donkey serum for 2 hours at 4°C. After washing with PBS, cells were resuspended in Antibody Diluent (Dako) and incubated with goat polyclonal Ab against DP (20 μg/ml; Santa Cruz Biotechnology), or normal goat IgG (20 μg/ml; Sigma Aldrich, St. Louis, MO, USA) for 1 h at 4°C. Cells were washed in PBS, resuspended in Antibody Diluent (Dako) and incubated for 1 h at 4°C with the secondary antibody (1:2000; Cy5-conjungated donkey anti-goat secondary Ab; Life Technologies, Carlsbad, CA, USA). All samples were finally washed twice with PBS, resuspended in fixative solution and cyto-spun on microscopy slides. After being mounted with Vectashield/DAPI medium (Vector Laboratories; Burlingname, CA, USA) the samples were analyzed on an Olympus IX70 fluorescence microscope (Olympus, Vienna, Austria) and an Olympus UPlanApo–60x/1.20 lens using an Olympus DP50-CU digital camera and Olympus Cell^P software.

Western blots

Proteins were normalized to wet weight by giving an appropriate volume (100 μl/20 mg colon) of extraction buffer (50 mmol/l Tris, 10 mmol/l EDTA, 1% Triton X and protease inhibitor cocktail (Roche Diagnostics; Vienna, Austria). Samples were homogenized accurately and lysates were diluted in equal amounts of reducing sample buffer (15% glycerol, 5% SDS, 5% mercaptoethanol and bromphenol blue in Tris-HCl [pH 6.8]) and incubated for 5 min at 95°C. Proteins were resolved by SDS-polyacrylamide gel electrophoresis (SDS/PAGE) (Life Technologies, Carlsbad, CA, USA) and transferred to a PVDF membrane (Merck Millipore; Billerica, MA, USA). Membranes were blocked in TBST buffer (1 mM CaCl_2, 136 mM NaCl, 2.5 mM KCl, 25 mM Tris-HCl, 0.1% (v/v) Tween 20) containing 5% milk powder. Further, membranes were incubated with rabbit anti-CRTH2 (1:1000; Acris, Herford, Germany), rabbit anti-DP1 (1:1000; Cayman Chemicals, Ann Arbor, MI, USA) and mouse anti-β-actin antibodies (Sigma-Aldrich, St. Louis, MO, USA) overnight at 4°C. Membranes were washed and subsequently immunoblotted with HRP-conjugated Abs (1:4000; Jackson ImmunoResearch, West Grove, PA, USA) for 1 h at room temperature. ECL Western blotting Substrate (Thermo Fisher Scientific) was used to visualize protein bands and blots were quantified with ImageJ software (NIH, Bethesda, MD, USA). Protein expression of DP and CRTH2 were normalized to their respective actin levels.

Induction of DSS colitis and tissue treatment

C57BL/6 (males, 5-9 weeks old, 20-26g) were obtained from Charles River (Sulzfeld, Germany) and kept in house for 2 weeks prior to experiments. Mice were housed in plastic sawdust floor cages at constant temperature (22°C) and a 12:12-h light–dark cycle with free access to standard laboratory chow and tap water. Animals were matched by age and body weight. Colitis was induced in mice by adding 3% (wt/vol) of dextran sulfate sodium (DSS salt; reagent grade; 36–50,000 Da; MP Biomedicals Europe) to the drinking water (tap water) while control animals received tap water only. Mice were kept on DSS over a 7-day period. Body weights were measured daily and the DSS-containing drinking water was daily monitored to insure that the different treatment groups would consume comparable amounts of DSS. Scoring of inflammation was carried out as previously published in a blinded fashion (33). Following macroscopic scoring, segments of the distal colon were stapled flat onto cardboard with the mucosal side up and fixed for 24 h in 10% neutral-buffered formalin. Tissue was then dehydrated, embedded in paraffin and standard hematoxylin/eosin staining was performed on 5 μm thick sections. Experimental procedures were approved by the Austrian Federal Ministry of Science and Research (protocol number: BMWF-66.010/0146-II/3b/2012) and performed in accordance with international guidelines.

Isolation and flow cytometric analysis of lamina propria leukocytes

The colon was removed, rinsed in HBSS, weighed, cut into small pieces and transferred into a 50 mL falcon tube containing HBSS, Hepes and Penicillin/Streptomycin (PS). Samples were incubated at 37°C and washed 4 times for 10 mins with HBSS/Hepes/PS, followed by 4 washes with HBSS/EDTA/PS. Afterwards, samples were rinsed in complete RPMI medium for 5 min and then transferred into complete RPMI with 100 units/mL of collagenase type 2 (Life Technologies, Carlsbad, CA, USA) for 1 hr at 37°C. After collagenase digestion, the cell suspension was passed through a 40μm cell strainer and centrifuged at 400×g for 7min. The pellet was washed with PBS, centrifuged and cells were prepared for flow cytometry. Leukocytes were stained with Alexa Fluor 647-conjugated CD11b Ab (1:100), PE-conjugated Siglec F Ab (1:100), Alexa Fluor 488-conjugated F4/80 Ab (1:20; all Abs from BD Biosciences; San Jose, CA, USA), and PerCP-Cy5.5 conjugated Gr1 Ab (1:100, eBioscience, Vienna, Austria) for 30 minutes at 4°C. Samples were washed once in PBS, fixative solution was added and samples were kept on ice until analyzed on a FACSCalibur flow cytometer. Data are normalized to colon weight and expressed as % of total cells. Representative flow plots and gating criteria are shown in supplemental material S2.

Immunohistochemistry of colon tissue

Paraffin-embedded sections of human colon from UC patients and controls were cut (10 μm) and deparaffinized, microwaved for 2 × 5 min cycles in 10 mM citrate buffer and then processed by ABC method (Vectastain ABC kit; Vector Laboratories, Burlingname, CA, USA) according to the manufacturer’s protocol. Sections were incubated with rabbit anti-CRTH2 (1:200; Acris, Herford, Germany) (34) or rabbit anti-DP1 (1:100; Cayman Chemicals, Ann Arbor, MI, USA) (9) visualized with 3-3′-diaminobenzidine and counterstained with haematoxylin. The specificity of the Abs was tested by omitting the primary Ab and by incubating with the respective blocking peptide provided by the manufacturers (see supplemental material S3). Images were taken with a high-resolution digital camera (Olympus DP 50), processed and analyzed by Cell^A imaging software (Olympus, Vienna, Austria). Only contrast and brightness of images were adjusted.

Myeloperoxidase (MPO) activity

MPO activity represents an index of neutrophil accumulation in the tissue and correlates with the severity of the colitis. MPO activity has been performed as described previously (33).

Liquid chromatography (LC)-mass spectrometry (MS)

LC-MS/MS was employed to detect PGD₂, prostaglandin E₂ (PGE₂ ) and thromboxane B₂ (TXB₂) in serum of human and mouse blood. Sample analysis was performed using liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS). The LC-MS/MS system consisted of a hybrid triple quadrupole-ion trap QTrap 5500 mass spectrometer (AB Sciex, Darmstadt, Germany) equipped with a Turbo-V-source operating in negative ESI mode, an Agilent 1200 binary pump and degasser (Agilent, Waldbron, Germany) and an HTC Pal autosampler (Chromtech, Idstein, Germany). Serum samples were spiked with the isotopically labeled internal standards and twice extracted using ethyl acetate. After evaporation of the organic extracts and reconstitution in mobile phase, the chromatographic separation of the analytes was carried out in a Synergi Hydro-RP column (150 × 2 mm I.D., 4-μm, Phenomenex, Aschaffenburg, Germany) under gradient conditions (300 μl/min) using water/formic acid (100:0.0025, v/v) and acetonitrile/formic acid (100:0.0025, v/v) as mobile phases. Sample run time was 16 minutes and injection volume of samples, 45 μl. Retention times of TXB₂, PGE₂ and PGD₂ were 8.0 min, 8.7 min and 9.2 min, respectively.

The mass spectrometer was operated in the negative ion mode with an electrospray voltage of −4500 V at 450°C. Multiple Reaction Monitoring (MRM) was used for quantification. The mass transitions used were m/z 351.1 → m/z 315.0 for PGE₂ and PGD₂, m/z 369.1 → m/z 169.1 for TXB₂, m/z 355.1 → m/z 275.1 for [²H₄]-PGE₂ and [²H₄]-PGD₂ and m/z 373.1 → m/z 173.1 for [²H₄]-TXB₂, all with a dwell time of 50 ms. Both quadrupoles were working at unit resolution. Quantitation was performed with Analyst Software V1.5 (Applied Biosystems, Darmstadt, Germany) using the internal standard method. Ratios of analyte peak area and internal standard peak area (y-axis) were plotted against concentration (x-axis) and calibration curves for each analyte were calculated by least square regression using 1/concentration² weighting.

Statistical analysis

Data were analyzed either by Student’s t-test or One-way ANOVA followed by Tukey’s posthoc test using GraphPad Prism (GraphPad Software). P values <0.05 were considered significant.

Results

Peripheral blood leukocytes show differential expression of DP and CRTH2 in UC

We first investigated the surface expression of the two PGD₂ receptors in peripheral blood leukocytes using flow cytometry. The DP receptor expression was increased in neutrophils, but decreased in CD14⁺ monocytes of UC patients in comparison to healthy individuals (controls) whereas in NK cells, CD3⁺ T cells and CD3⁺ /CD4⁺ T cells, expression was similar between the two cohorts (Fig. 1 A). Although the means of the DP receptor expression in eosinophils did not differ significantly between controls and UC patients, the expression was increased in correlation to disease activity in UC (see Figs. 3 C and D). Contrary to the DP receptor, expression of CRTH2 was downregulated in eosinophils, NK cells and CD3⁺ T cells whereas in monocytes and CD3⁺ /CD4⁺ T cells, no significant differences in receptor expression were seen between UC patients and healthy controls (Fig. 1 B). CRTH2 was practically undetectable in neutrophils from both cohorts (Fig. 1 B). Immunofluorescence stainings of DP and CRTH2 on polymorphonuclear leukocytes (PMNL) and peripheral blood mononuclear cells (PBMC) from healthy individuals are shown in Fig. 2. DP receptor expression on PMNLs and PBMCs appeared generally moderate in comparison to the expression of the CRTH2 receptor. See supplemental material S3 for isotype control.

Fig. 1. Expression of DP and CRTH2 on peripheral blood leukocytes.

DP and CRTH2 receptor expression were evaluated in peripheral blood leukocytes of patients with active ulcerative colitis (UC) and healthy (control) subjects. (A) Differences in DP expression between UC patients and control subjects were observed for neutrophils and monocytes while receptor expression in the other leukocyte populations did not differ between the two cohorts. (B) Differences in CRTH2 expression between UC patients and control subjects were observed for eosinophils, NK cells and CD3⁺ T cells while expression on monocytes and CD3⁺/CD4⁺ T cells did not differ between the two cohorts. Changes in DP and CRTH2 receptor expression were recorded as mean fluorescence intensity and data were expressed as fold increase of fluorescence over isotype control. Student’s t-test; n = 8-13. p values <0.05 were considered significant. (C) Comparison of CRTH2 expression on eosinophils between UC patients on 5-ASA treatment (n=10) and UC patients in remission also on 5-ASA treatment (n=8; Student’s t-test). Remission patients highly expressed CRTH2 similar to control subjects indicating that 5-ASA treatment was not responsible for the decrease in CRTH2 expression seen in eosinophils from UC patients.

Fig. 3. Correlation between UC scores and CRTH2/DP receptor expression in peripheral blood leukocytes.

In eosinophils, indices of disease activity in UC patients (Mayo score and CAI) correlate inversely with CRTH2 (A, B) but positively with DP expression (C, D) while in neutrophils, DP expression correlates inversely with the Mayo score (E) and CAI (F). X-axes in graphs depict fold increase of receptor expression (fluorescence) over isotype control. Linear regression; r², goodness of fit; n=7-10. p values <0.05 were considered significant.

Fig. 2. Immunofluorescence for CRTH2 and DP in polymorphonuclear (PMNL) and mononuclear (PBMC) leukocytes.

Stainings were performed in PMNLs (polymorphonuclear leukocytes) and PBMCs (peripheral blood mononuclear cells) from healthy donors. DAPI was used to counterstain cell nuclei. Immunofluorescence images demonstrate that CRTH2 is highly expressed on the cell surface of peripheral blood eosinophils (PMNL; upper left image; cell with bi-lobed nucleus), but hardly expressed on neutrophils (PMNL; upper left image; cell with polymorph nucleus). A moderate to high CRTH2 expression is found on the cell surface of mononuclear cells (PBMC). Immunofluorescence staining of DP displays a low to moderate expression on peripheral blood eosinophils and neutrophils (PMNL) and a moderate surface expression on mononuclear cells (PBMC). Calibration bar: 20 μm. The respective isotype controls for CRTH2 (rabbit IgG) and DP (goat IgG) are shown in supplemental material S3.

Since nearly all patients were on 5-ASA treatment we compared eosinophils from active UC patients on 5-ASA therapy with remission patients on 5-ASA therapy to investigate whether the decreased expression of CRTH2 could have been due to effects of the drug. CRTH2 expression was increased in UC patients in remission as compared to patients with acute UC (Fig. 1 C) excluding the possibility that 5-ASA is responsible for the decrease in CRTH2 in active UC.

DP and CRTH2 receptor expression correlate with the activity of UC

We also analyzed the correlations of DP/CRTH2 expression for eosinophils and neutrophils with the clinical disease scores of UC patients. CRTH2 expression on eosinophils correlated inversely with Mayo scores and CAI (Fig. 3 A and B) as opposed to DP expression which correlated positively with the scores (Fig. 3 C and D). In contrast, DP expression showed inverse correlation with the disease scores on neutrophils (Fig. 3 E and F).

Effect of PGD₂ ligands on CD11b expression and CRTH2 internalization in blood eosinophils

The β2 integrin CD11b is rapidly activated and upregulated on the leukocyte membrane upon activation of the cells. To investigate the inflammatory activity of peripheral blood eosinophils, expression of CD11b was evaluated in the presence of PGD₂ and the selective CRTH2 agonist DK-PGD₂ (Fig. 4). While eosinophils from healthy controls showed increased expression of CD11b against rising concentrations of both ligands, hardly any increase in CD11b expression was seen in UC patients (Fig. 4 A). Since agonist-induced stimulation of GPCRs can lead to receptor internalization, we tested whether the low expression of CRTH2 on eosinophils in UC patients could have been caused by internalization. Peripheral blood eosinophils isolated from healthy subjects were incubated with agonists for 60 min and CRTH2 expression on the cell surface was recorded thereafter by flow cytometry. The internalization assay showed that CRTH2 is profoundly internalized in the presence of PGD₂, DK-PGD₂ and the CRTH2-activating metabolite 11-dehydro TXB₂ (Fig. 4B). To document that specific agonist binding with PGD₂ or the CRTH2-activating metabolite 11-dehydro TXB₂ is required for CRTH2 internalization, we tested another important chemokine involved in UC, i.e. eotaxin (29), in the internalization assay. Eotaxin did not lead to internalization of CRTH2 (Fig. 4 B). In a previous study, we have already demonstrated that the DP receptor does not internalize in human eosinophils in the presence of DP agonists (24).

Fig. 4. CD11b expression and CRTH2 internalization in human peripheral blood eosinophils.

(A) CD11b expression was evaluated in peripheral blood eosinophils in response to the CRTH2 ligands PGD₂ and DK-PGD₂. UC patients showed less CD11b upregulation than healthy individuals (controls). Changes in CD11b surface expression were recorded as percent of the vehicle control and data are shown as means ± SEM, n = 6; *p < 0.05 vs. controls; ANOVA. (B) PGD₂, DK-PGD₂ and 11-dehydro TXB₂ induced internalization of CRTH2. Assays were performed in citrated whole blood from healthy individuals by means of flow cytometry. CRTH2 surface expression in CD16-negative eosinophils was expressed as percent of the vehicle control (ETOH only) and data are shown as means, n = 6-12; ANOVA; *p < 0.05, ***p < 0.001 vs. vehicle.

Increased expression of CRTH2 and DP in colon biopsies from UC patients vs. control subjects

Immunohistochemistry revealed increased infiltration of CRTH2-positive cells into the lamina propria of the colonic mucosa in biopsies of UC patients as compared to control subjects (Fig. 5 A). CRTH2-positive staining was also observed in epithelial cells of the colon (Fig. 5 A). To show the antibody’s specificity, preabsorption experiments with a blocking peptide were carried out (supplemental material S3). In previous studies, Western blots for CRTH2 revealed two bands (approximately 41 and 55 KDa) (14, 34). The 41 kDa band represents the predicted molecular weight of CRTH2 while the 55 kDa band represents the glycosylated form of CRTH2 as previously shown (14). Evaluation of Western blots from colon biopsies of UC patients showed a strong and significant increase of CRTH2 protein at 55 kDa as well as for total CRTH2 protein (41 kDa and 55 kDa bands together) when compared to healthy subjects (Fig. 5 A), suggesting a shift of native CRTH2 to its glycosylated form in UC. Immunohistochemical staining with a DP antibody in sections of colon biopsies showed moderate expression in lamina propria and epithelial cells from UC patients and controls (preabsorption control shown in supplementary material S3). Similar to CRTH2, Western blots detected two bands for DP, a strong band at approximately 41 kDa and a weak band at 55 kDa (34). Density of the 55 kDa band was increased in samples from UC patients whereas there was no change in the band density at 41 kDa (Fig. 5 B).

Fig. 5. Immunohistochemistry and Western blots of colon biopsies from UC patients and control subjects.

Immunohistochemical staining was performed for CRTH2 and DP in colon biopsies of UC patients and healthy (control) subjects. (A) Brown DAB staining of CRTH2 was visible in lamina propria cells and epithelial cells of colon sections (scale bar: 50 μm). The density of lamina propria cells was strongly increased in UC patients. Western blots revealed significant increase in CRTH2 55 kDa protein in biopsies from UC patients. (B) With DP immunohistochemistry, some DAB-stained cells were visible in lamina propria, however, epithelial cells only lightly stained in the colon (scale bar: 50 μm). In contrast to the CRTH2 staining, there was no prominent increase in DP-positive lamina propria cells. Density of the 55 kDa band of DP was stronger in UC patients whereas for the 41 kDa band, no increase was noticed. UC values were set at 100%. Other data are relative to these values. (n=5; Students t-test; * p<0.05; n.s., not significant);

The increased expression of CRTH2 in colon biopsies from UC patients correlates with disease activity

We correlated disease activity of the UC patients with the CRTH2 protein expression measured in the Western blots and found that both Mayo score and CAI strongly correlated with the level of 41 KDa CRTH2 protein (Figs. 6 A and B). We also performed a histologic score of CRTH2 positive cells in the lamina propria of colon biopsies from UC patients and compared severely affected mucosa from UC patients that exhibited a Mayo score consistent with moderate-severe disease activity with mucosa from UC patients that exhibited a Mayo score consistent with mild activity. We observed a significant increase in CRTH2 positive cells in the biopsy sections of UC patients with a moderate-severe disease (Figs. 6 E and F).

Fig. 6. CRTH2 in colon biopsies of UC patients correlates with disease activity.

Levels of CRTH2 protein, as measured by Western blot (41 KDa and total CRTH2), were correlated with UC disease activity (Mayo score and Clinical activity index; CAI) by linear regression (r², goodness of fit; n=5; p <0.05). (A, B) The increase in CRTH2 levels (41 KDa) correlated significantly with the activity scores. (C, D) Total levels of CRTH2 (41 and 55 KDa) also correlated positively with the activity scores, however, not significantly. (E, F) Counting of CRTH2 positive cells in sections of colon biopsies from UC patients (3 different biopsies from 3 UC patients consistent with Mayo score severe-moderate and from 3 UC patients consistent with Mayo score mild; Student’s t-test) revealed a significant increase in CRTH2 positive cells in severely affected mucosa in comparison to mildly affected mucosa.

The selective CRTH2 antagonist CAY 10595 improved while the DP antagonist MK0524 worsened DSS-induced colitis in mice

In an attempt to clarify the functional role of DP and CRTH2 in UC we used a mouse model of DSS-induced colitis and treated mice with antagonists against each of the PGD₂ receptors. Daily treatment of DSS colitic mice with the selective CRTH2 antagonist CAY10595 (5 mg/kg; s.c.) (35) reduced weight loss, inflammation score and MPO activity indicating pro-inflammatory actions of CRTH2 in this mouse model (Fig. 7 A). Hematoxylin/eosin stainings of histologic sections (10 μm) show reduced mucosal damage and less infiltration of immunocytes into the submucosa of the mouse colon after treatment with CAY10595 (Fig. 7 A). As most populations of mouse leukocytes express CRTH2 (5, 16-18), we were interested whether CAY10595 may alter leukocyte influx into the lamina propria of the mouse colon in the DSS colitis. Treatment with CAY10595 significantly lowered the lymphocyte (comprising T cells, B cells and NK cells) and neutrophil population thus corroborating the findings of reduced MPO activity (Fig. 7 A). Contrary to the CRTH2 antagonist, daily treatment with the DP antagonist MK0524 (1 mg/kg; s.c.) worsened colitis and increased MPO activity in the colon confirming the anti-inflammatory action for DP in colitis (Fig. 7 B) (8).

Fig. 7. Effects of CRTH2 and DP receptor antagonists in DSS-induced colitis.

(A) The CRTH2 antagonist CAY10595 (5 mg/kg; s.c.) lowered inflammation scores, MPO activity and weight loss in C57BL/6 mice with DSS colitis. n=8-12; means ± SEM; ANOVA; Tukey’s post hoc; * p<0.05. Histological images (H/E staining) show representative colon sections (5 μm) from control mice, mice with DSS colitis (+ vehicle treatment) and DSS-mice treated with CAY10595 (scale bar: 200 μm). MPO assay demonstrates relative change to DSS + vehicle (100%). Treatment with CAY10595 reduced infiltration of lymphocytes and neutrophils into the lamina propria of the colon (n=4; Student’s t-test). (B) Effect of daily treatment with DP antagonist MK0524 (1mg/kg s.c.) on severity of DSS colitis. The DP antagonist worsened inflammation score and increased MPO activity, thereby showing opposite effects to the CRTH2 antagonist (n=6-8; means ± SEM; ANOVA; Tukey’s post hoc). MPO assays shows relative change to DSS + vehicle (100%). p values <0.05 were considered significant. n.s., not significant.

Increase of prostaglandin levels in serum from human UC patients and mice with DSS colitis

Levels of the prostaglandins PGD₂, PGE₂ and TXB₂ were measured in serum from UC patients by LC-MS and were found increased as compared to healthy controls (Fig. 8 A). PGE₂ and PGD₂ were also measured in sera of mice with DSS colitis and were found to have increased during experimental UC. The CRTH2 antagonist CAY10595 decreased PGE₂ levels at a daily dose of 5 mg/kg s.c. in mice with DSS colitis but had no effect on PGD₂ levels (Fig. 8 B).

Fig. 8. Increased levels of prostaglandins in serum from UC patients and mice with DSS colitis.

(A) PGD₂, PGE₂ and TXB₂ were measured by LC-MS in serum from UC patients and healthy (control) subjects and were found increased in the UC cohort. n= 9-13; Student’s t-test. (B) An increase in PGE₂ and PGD₂ levels was also observed in mice with DSS colitis. The CRTH2 antagonist CAY10595 (5 mg/kg s.c.) decreased PGE₂ levels whereas PGD₂ remained unaltered after treatment with the antagonist. DSS + vehicle was set at 100%. Data of groups denote relative change to DSS + vehicle group. n=8-15; ANOVA; Tukey’s post hoc; p values <0.05 were considered significant.

Discussion

IBD can be described as a multifactorial disease with an uncontrolled immune response causing inflammation in genetically predisposed individuals; however, much of the etiology of IBD still remains unresolved (23). Despite significant progress in the pharmacotherapy of IBD, such as with the introduction of anti-TNF agents, side effects of drugs and treatment failure are still major drawbacks (36, 37). In the current study we focused on the role of the two PGD₂ receptors in human and experimental UC and provide several lines of evidence that the CRTH2 receptor may play a different role than the “anti-inflammatory” DP receptor and that blockade of CRTH2 in fact improves inflammation in mice with DSS colitis. The results of the present study raise the possibility that “pro-inflammatory” CRTH2 receptors could represent a new drug target for IBD. It should be noted that the study may have certain limitations as only a low patient sample number was investigated and only UC patients that were already on medication, except for one, were included in the study.

First, we demonstrated that DP and CRTH2 expression is regulated in a differential manner in peripheral blood leukocytes, indicating that the two receptors may follow different tasks in UC. While during UC, CRTH2 demonstrates low surface expression on eosinophils and other types of leukocytes in the blood, DP is upregulated on neutrophils, but not on eosinophils as compared to healthy subjects. Expression of CRTH2, however, is restored on eosinophils of UC patients in remission. On the other hand, we could also show that a high amount of CRTH2-positive leukocytes infiltrated the colon in UC patients and were largely present in the lamina propria of the colonic mucosa. Since the major task of CRTH2 is to facilitate chemotaxis (12), the high presence of CRTH2 in the colon of UC patients may reflect increased chemotaxis of CRTH2-positive leukocytes to the inflamed colon tissue.

The low expression of CRTH2 on peripheral blood leukocytes seemed contradictory to the strong influx of CRTH2 positive cells into the colon of UC patients. Interestingly, however, recent studies demonstrated that IBD patients had significant decreases in Treg/Th17 ratios and Treg frequency in peripheral blood but increased expression of Foxp3, IL-17a and higher amounts of Tregs in the colon mucosa suggesting a disparity between inflammatory activity of leukocytes in the blood and the inflamed tissue during IBD (38, 39). We therefore studied CRTH2 expression in more detail in eosinophils, a cell type that colocalizes both receptors (22) and promotes inflammation in IBD (25). In accordance with the low expression of CRTH2 on blood eosinophils, activation of CRTH2 with PGD₂ and its specific ligand DK-PGD₂ induced little expression of CD11b in eosinophils from UC patients as compared to eosinophils from healthy controls, indicating low inflammatory activity of blood eosinophils from UC patients. A different study also described low expression of CD11b, CD18 and the chemokine receptor CCR3 in eosinophils of UC patients which confirms our observations (40). 62% of UC patients in that study were on no treatment indicating that medication most likely does not interfere with CD11b expression on leukocytes in our UC patients. In our study we could rule out that 5-ASA treatment resulted in different CRTH2 receptor expression, since UC patients in remission taking 5-ASA were similar to controls. It is also unlikely that the low expression of CRTH2 in UC patients is due to some immature form of leukocytes since CRTH2 is largely present in eosinophils and precursor eosinophils of human bone marrow (22). Other groups described that eosinophils from IBD patients are activated since they noticed higher eosinophil peroxidase baseline release (41). We conducted additional experiments to investigate whether the low expression of CRTH2 may have been caused by receptor internalization due to long term activation and exposure to CRTH2 ligands, such as PGD₂ and other metabolites. We have previously described internalization for CRTH2 in HEK293 cells (24). Indeed, presence of CRTH2 ligands promoted internalization of CRTH2 in assays with human peripheral blood eosinophils. Notably, internalization was not observed for DP in human eosinophils (24). Internalization of CRTH2 could therefore occur in the serum of UC patients in which higher PGD₂ levels are present than in healthy subjects. Although PGD₂ has been shown to be rapidly degraded in plasma, the generated PGD₂ metabolites, such as Δ¹²-PGD₂ and Δ¹²-PGJ₂ are known to act as stable CRTH2 agonists (42). Additionally, a stable metabolite of TXB₂, 11-dehydro TXB₂, which is a major product of TXB₂ in plasma (43), also activates CRTH2 and causes internalization in eosinophils (Fig. 4 B) (44). We found highly increased levels of TXB₂ in sera from UC patients raising the possibility that CRTH2 internalization may not only be driven by PGD₂ but also by metabolites like 11-dehydro TXB₂ in vivo.

Expression of DP and CRTH2 in blood eosinophils clearly correlated with the clinical and endoscopic activity scores displaying high DP but low CRTH2 expression in severe cases of UC. This suggests that CRTH2 receptors are widely internalized in severe cases. While the DP receptor correlates positively with activity scores in eosinophils it correlates inversely with activity scores in neutrophils indicating low expression on neutrophils in severe cases. DP has previously been linked with the amelioration of experimental colitis by inhibition of granulocyte - mostly neutrophil - influx to the colon following application of the DP agonist BW-245C (8). The opposing correlation of DP and CRTH2 on eosinophils and its functional significance are not clear at the moment but DP and CRTH2 are coexpressed in many leukocytes and down- or upregulation of one receptor may modify the other’s signaling behavior, as recently demonstrated for DP and/or CRTH2 overexpressing HEK293 cells (24).

In contrast to peripheral blood leukocytes, we found high expression of CRTH2 protein and influx of CRTH2-positive cells into the lamina propria of the inflamed colon, a finding also described by others (45). This study also used UC patients that were on standard medication for UC (45). Additionally, we noticed high expression of CRTH2 in epithelial cells. The CRTH2 cell influx to the colon in UC is in line with a recent study that described increased levels of the PGD₂-producing enzyme L-PGDS in the colon of UC patients and in mice with DSS colitis (11), suggesting that a prominent PGD₂-CRTH2 axis may contribute to inflammation during UC. We therefore investigated the functional role of DP and CRTH2 in UC and applied selective antagonists against each of the receptors in a model of experimental colitis. While the DP antagonist MK0524 produced an exacerbation of colitis, confirming the anti-inflammatory role of DP from previous work (8), the CRTH2 antagonist CAY10595 improved disease severity and lowered MPO activity and PGE₂ levels, suggesting a different, i.e. pro-inflammatory, role for CRTH2 in experimental UC. CRTH2 may therefore exert its pro-inflammatory role in the colon itself since much higher levels of CRTH2 and CRTH2 positive cells were found in biopsies from UC patients than in healthy individuals. The CRTH2 antagonist was also able to reduce infiltration of lymphocytes and neutrophils into the lamina propria. It should be noted that unlike in humans, CRTH2 is expressed in mouse Th1 cells and neutrophils next to other leukocytes (16, 46). Thus a direct inhibitory effect of the drug on the migration of these cells seems possible.

Serum samples of UC patients revealed a small increase in PGD₂ but more prominent increases in TXB₂ and PGE₂, as previously described (47). Similarly, PGD₂ and PGE₂ levels were elevated in serum of mice with DSS colitis. While the CRTH2 antagonist lowered PGE₂ levels, it showed no effect on PGD₂. The decrease in PGE₂, however, may have contributed to improvement of colitis. PGE₂ is widely seen as a dual lipid mediator with pro- as well as anti-inflammatory effects. PGE₂ can actually act, depending on its level, as a pro-inflammatory mediator in IBD (48). High levels of PGE₂ may worsen IBD by shifting the IL-12/IL-23 balance in dendritic cells in favor of IL-23 (48) while a certain level might be necessary to protect the epithelium from damage via EP4 receptors (49, 50). PGE₂ may also impair wound healing in IBD (51).

In conclusion, CRTH2 displays expression levels contrary to DP in blood leukocytes during UC. While CRTH2 decreases in several peripheral blood leukocyte populations, expression of DP increases in neutrophils. The decrease in CRTH2 is most likely due to enhanced internalization of the receptor depending on the severity of UC. The situation is different in colon tissue that shows high expression of CRTH2 during UC. Our data also suggest that actions of PGD₂ very likely depend on the expression of its receptors with regard to an “anti- and/or pro-inflammatory” effect of PGD₂. In light of the clinically already available CRTH2 receptor antagonists, blocking of CRTH2 in UC may provide a worthwhile therapeutic approach.