Prostaglandin D₂ and leukotriene E₄ synergize to stimulate diverse T_H2 functions and T_H2 cell/neutrophil crosstalk

T_H2 lymphocytes

T_H2 cells were isolated from buffy coats (National Blood Service, Bristol, United Kingdom), as described in the Methods section in this article's Online Repository at www.jacionline.org.¹¹ They are memory cells showing a CD4⁺CRTH2⁺CD45RO⁺GATA3⁺CCR6⁻CD45RA⁻ retinoic acid–related orphan receptor γt (RORγt)⁻ phenotype with relatively high purity (see Fig E1, A, in this article's Online Repository at www.jacionline.org).

For analysis of gene regulation and cytokine production, T_H2 cells were treated with PGD₂ or LTE₄ alone or their combination in X-VIVO 15 medium (Lonza, Basel, Switzerland) in the presence or absence of antagonist compounds for 2.5 hours (microarray, PCR array, and quantitative PCR [qPCR]) or 4 hours (Luminex).

Cells were treated with the same compounds in serum-free RPMI medium for 4 hours to prepare T_H2 cell–conditioned media for neutrophil assays.

Neutrophils

Human neutrophils were isolated from fresh whole blood. Briefly, the red blood cell pellet was collected after Ficoll-Paque Plus density gradient, suspended in HBSS, and mixed with 3% dextran. Neutrophil-rich supernatant was collected and treated in a 0.2% NaCl solution for red blood cell lysis and resuspended in RPMI medium.

Microarrays

Total cellular RNA was extracted with RNeasy Mini kits (Qiagen, Hilden, Germany). Microarrays were performed by Cambridge Genomic Services (Cambridge, United Kingdom) using a HumanHT-12 v4 chip. Genes significant at a P value of less than .05 were analyzed by using the Venn Diagram module within GenePattern.²⁰ Pathway analyses were conducted with IPA (Ingenuity Systems, www.ingenuity.com). Heat maps were generated by using GENE-E software (http://www.broadinstitute.org/cancer/software/GENE-E/index.html).

PCR arrays

PCR arrays were performed with an RT² Profiler PCR Array Human Common Cytokines kits (SABiosciences, Frederick, Md) in a LightCycler 480 Real-Time PCR System (Roche, Mannheim, Germany).

Luminex

Cytokines were measured with a Procarta Human Cytokine Immunoassay kit (Affymetrix, Santa Clara, Calif). The results were obtained with a Bio-Plex 200 System (Bio-Rad Laboratories, Hercules, Calif).

qPCR

qPCR was performed, as described previously.¹⁷ Primers and probes (Roche) used are listed in Table E1 in this article's Online Repository at www.jacionline.org.

ELISA

Cytokines were assayed with ELISA kits (R&D Systems, Minneapolis, Minn). The results were measured in a FLUOstar OPTIMA luminescence plate reader (BMG LabTech, Ortenberg, Germany).

Flow cytometric analysis

Cells were labeled with antibody to CD16–fluorescein isothiocyanate or Annexin V–allophycocyanin and then acquired with an LSR II Flow Cytometer (BD Biosciences, San Jose, Calif).

Cell aggregation analysis

Cell aggregation was photographed with a Nikon Eclipse TS100 microscope (Nikon, Tokyo, Japan). Images were analyzed with CellProfiler 2 software (Broad Institute, Cambridge, Mass; also see the Methods section in this article's Online Repository).²¹

Chemotaxis assays

Chemotaxis assays were conducted, as described previously.⁹

Statistics

Data were analyzed by using 1-way ANOVA, followed by the Newman-Keuls test. P values of less than .05 were considered statistically significant.

Effect of PGD₂ and LTE₄ on the gene expression profile of T_H2 cells

PGD₂ and LTE₄ synergistically evoke type 2 cytokine production from human T_H2 cells.¹⁷ To understand their broader synergistic effects on T_H2 cell function, we investigated the transcriptional responses to LTE₄ or PGD₂ added either alone or in combination by using RNA microarrays. Three experimental replicates were prepared for each of the 4 groups (control, LTE₄, PGD₂, and their combination). The concentrations of LTE₄ (50 nmol/L) and PGD₂ (100 nmol/L) used for the treatments were close to their relative median effective concentration values for type 2 cytokine production in T_H2 cells.¹⁷

The data showed broad transcriptional changes after treatment. The mRNA levels of 1344, 4750, and 5868 genes were significantly (P < .05) modulated (including upregulation and downregulation) by LTE₄, PGD₂, or their combination, respectively (Fig 1, A). The effect of PGD₂ was much broader than that of LTE₄. Although some (approximately 675) of the gene responses overlapped, most of them were regulated distinctly: 669 only by LTE₄ and 4075 only by PGD₂. The combination of LTE₄ and PGD₂ amplified significantly the range of the transcriptional response. Expression of a group of 1885 genes was altered only by combination treatment, indicating the combinatorial effect of PGD₂ and LTE₄ on gene expression. Among the modulated genes, about half were upregulated and half were downregulated (Fig 1, B).

The genes regulated included those involved in the pathways critical for T-cell intrinsic functions and interactions with other cell types (through cell-surface receptors and secreted mediators). To define the significance of the gene expression changes specifically in relation to T_H2-mediated allergic inflammation, we analyzed first the effect of lipid mediators on T-cell intrinsic functions and then the effect on cytokine-driven crosstalk with downstream effector cells.

Effects of PGD₂ and LTE₄ on the apoptosis and migration of T_H2 cells

Ingenuity pathway analysis of the microarray data suggested that PGD₂ and LTE₄ treatment altered the expression of clusters of genes associated with distinct cell-signaling pathways in T_H2 cells, including the phosphoinositide 3-kinase (PI3K) and apoptosis pathways (see Fig E2 in this article's Online Repository at www.jacionline.org). Western blotting for phospho-Akt also confirmed that activation of the PI3K pathway by PGD₂ and LTE₄ was inhibited by TM30089 and montelukast (see Fig E3 in this article's Online Repository at www.jacionline.org). PI3K pathway signaling is critical in mediating the antiapoptotic and chemotactic roles of PGD₂/CRTH2.^12,22 Therefore we further addressed the effect of PGD₂, LTE₄, and their combination on these functions (Fig 2).

In the case of apoptosis, both PGD₂ and LTE₄ markedly reduced upregulation of Annexin V in T_H2 cells after 16 hours of IL-2 withdrawal (Fig 2, A). The combination of 2 mediators additively enhanced this effect.

In chemotaxis assays we first compared the chemotactic effect of PGD₂ with that of cysLTs, including LTD₄ and LTE₄ (Fig 2, B, left panel). These induced migration in a dose-dependent manner, peaking around 30 nmol/L for PGD₂ and LTE₄ and 20 nmol/L for LTD₄. The maximum response achieved by LTE₄ was higher (approximately 2-fold) than that elicited by LTD₄ but only approximately 27% of that elicited by PGD₂. The combination of PGD₂ and LTE₄ at concentrations close to their median effective concentration synergistically enhanced cell migration (Fig 2, B, right panel).

In both apoptosis and chemotaxis assays the contribution of PGD₂ and LTE₄ was ablated by TM30089 and montelukast, respectively, and the combination of TM30089 and montelukast inhibited almost all cell responses induced by the combination of PGD₂ and LTE₄.

Effects of PGD₂ and LTE₄ on expression of adhesion molecules in T_H2 cells

Inspection of the microarray data indicated induction by PGD₂ and LTE₄ of a particularly large number of transcripts associated with leukocyte adhesion (Table I). Prominent among these were the integrins αV (ITGAV), α2 (ITGA2), αE (ITGAE), and α11 (ITGA11), which constitute subunits of the leukocyte receptors α2βL for intercellular adhesion molecule 1 (ICAM-1) on endothelium, αEβ7 for E-cadherins on epithelium, and α2β/αVβ3 for collagen and vitronectin in tissue extracellular matrix. In addition, PGD₂ and LTE₄ induced expression of transcripts encoding ICAM-1 (ICAM1) and ICAM-2 (ICAM2), as well as the homophilic adhesion molecule CD31 (PECAM1) and the cadherin/catenin protein family members CTNNAL1, CTNNA1, CTNND1, PCDHA1, PCDHA4, and CDH1, which also act as homophilic adhesion receptors. The majority of these genes (n = 14) were induced by PGD₂ alone, whereas only 3 (ITGAV, CTNND1, and ITGA11) were induced by LTE₄. Upregulation of some genes (IGSF3, CTNNA1, NINJ1, DCHS1, PCDHA4, and ITGA2) was amplified by the combination of PGD₂ and LTE₄.

To explore the consequences of such increased integrin expression on T_H2 cell-cell adhesion, we used an in vitro cell activation–induced aggregation (CAIA) assay.²³ Stimulation with either lipid alone caused marked CAIA that formed within 0.5 and 1 hours and persisted for 4 to 6 hours (Fig 3, A). The combination of both mediators further enhanced the aggregation. To verify the involvement of integrins, which are critically dependent on Ca²⁺ ions, we tested the effect of EDTA, an inhibitory chelating agent, and MnCl₂, an activator of integrin function, on PGD₂/LTE₄-induced CAIA. As shown in Fig 3, B, EDTA (5 mmol/L) inhibited and MnCl₂ (1 mmol/L) prolonged the CAIA. Blocking antibodies to CD54 (ICAM1) and CD31 (PECAM1) were used to further confirm the contribution of integrins to the CAIA (Fig 3, C). Both antibodies partially reduced the intensity of CAIA in a concentration-dependent manner. The inhibitory effect of anti-CD54 was slightly more potent than that of anti-CD31, and the combination showed a marginal additive effect.

Enhancement of proinflammatory protein production by PGD₂ and LTE₄ in T_H2 cells

To focus on gene regulation potentially relevant to T cell–mediated diseases, we next analyzed the genes encoding cytokines, chemokines, their receptors, and cluster of differentiation (CD) molecules detected by the microarray (Fig 4). A total of 95 of these genes were significantly modulated, most of them upregulated (Fig 4, A, and see Table E2 in this article's Online Repository at www.jacionline.org), most significantly the cytokines and chemokines. Although some of these effects were induced by PGD₂ alone (eg, IL26, IL1RL1, and CCR4) or LTE₄ alone (eg, CCL3, CCL3L1, CCL3L3, and CCL4L2), most were driven by their combination. In addition to type 2 cytokine genes, the expression of many other genes was also synergistically enhanced by the combination treatment (Fig 4, B). A number of genes were downregulated (see Table E2), notably transcription of GPR44, the gene for CRTH2 (CD294), which was downregulated by 2.4-fold by PGD₂ alone and 3.8-fold by combination treatment. Importantly, these microarray data were largely confirmed also by using a PCR array assay for human common cytokines, including 84 cytokine genes, among which approximately 30 showed significant changes (see Table E3 in this article's Online Repository at www.jacionline.org), although some effects (IL10 and IL21) were only detected by using the PCR array (Fig 4, C).

To further verify our findings, we also conducted qPCR (Fig 5, A) and Luminex (Fig 5, B) assays on selected cytokines. The qPCR data strongly supported the synergistic effects of PGD₂ and LTE₄. At the protein level, combination treatment either additively (IL-3, IL-22, and macrophage colony-stimulating factor) or synergistically (IL-8, IL-9, IL-21, and GM-CSF) enhanced the cytokine production. A similar cytokine profile was also observed after stimulation with phorbol 12-myristate 13-acetate (PMA)/ionomycin in both cultured T_H2 cells (by means of intracellular cytokine staining) and ex vivo T_H2 cells (by means of qPCR/Luminex; see Fig E1, B and C). As expected, TM30089 or montelukast only partially inhibited cytokine upregulation after PGD₂ plus LTE₄, but a combination of TM30089 and montelukast completely blocked this effect (see Fig E4 in this article's Online Repository at www.jacionline.org).

Effect of cytokines induced by PGD₂ and LTE₄ from T_H2 cells on neutrophil function

Neutrophilia is detected in the majority of patients with severe asthma,¹⁹ and the cytokines induced by activation of T_H2 cells included several that could potentially interact with neutrophils. To explore this possibility, we tested the ability of IL-8 and GM-CSF produced by T_H2 cells to elicit relevant changes in human neutrophil behavior (Fig 6). As indicated by the data in Fig 6, A, treatment with PGD₂ and LTE₄ promoted secretion of IL-8 and GM-CSF from resting levels (<90 and <70 pg/mL, respectively, in supernatant 1 to more than 560 and 650 pg/mL, respectively, in supernatant 2; Fig 6, A). This was inhibited by coincubation with TM30089 and montelukast (approximately 94 pg/mL IL-8 and approximately 50 pg/mL GM-CSF in supernatant 3).

First, we addressed the effect of endogenous IL-8 on neutrophil chemotaxis. Recombinant IL-8 induced neutrophil migration with a typical chemotaxis dose curve (see Fig E5, A, in this article's Online Repository at www.jacionline.org). This effect could be inhibited by a neutralizing antibody against IL-8 in a dose-dependent manner (see Fig E5, B). The T_H2 supernatant containing high levels of IL-8 (supernatant 2) had a strong capacity to induce neutrophil migration (Fig 6, B, left panel). Inhibition of IL-8 production by TM30089 and montelukast caused substantial (approximately 43%) reduction of the chemotactic activity of supernatant 3. The neutrophil migration to supernatant 2 was mostly blocked by IL-8 neutralizing antibody. To rule out the possibility that cell migration was induced directly by PGD₂ or LTE₄ (used to prepare supernatant 2), we examined the effects of these mediators (Fig 6, B, right panel). Neither was chemotactic for neutrophils.

Next, we studied the influence of endogenous GM-CSF on neutrophil behavior by measuring its ability to rescue cells from serum starvation–induced apoptosis using CD16 (FcγRIII) expression as a biomarker of cell integrity (see Fig E6, A, in this article's Online Repository at www.jacionline.org).²⁴ Confirming the validity of the assay, the numbers of CD16^high neutrophils decreased after serum withdrawal, which was inhibited by recombinant GM-CSF in a dose-dependent manner (see Fig E6, B). The inhibitory effect of GM-CSF was reversed by a neutralizing antibody against human GM-CSF (see Fig E6, C). Importantly, similar protection against apoptosis was observed when T_H2 cell supernatants, particularly supernatant 2, were substituted for recombinant GM-CSF (Fig 6, C, left panel). The protective capacity depended on the level of GM-CSF in the supernatant, which was reduced either by the inhibition of GM-CSF production in T_H2 culture (supernatant 3) or GM-CSF neutralizing antibody. To further confirm the antiapoptotic activity of the endogenous GM-CSF induced by PGD₂ and LTE₄, we also examined the expression of Annexin V in neutrophils (Fig 6, C, right panel). As expected, T_H2 supernatants, especially supernatant 2, reduced the numbers of Annexin V–positive cells markedly, which was partially inhibited by the GM-CSF antibody.

Reagents

PGD₂ and LTE₄ were purchased from Enzo Life Science (Farmingdale, NY). TM30089 was supplied by ChemieTek (Indianapolis, Ind). The human CD4⁺ T Cell Isolation Kit II, anti-human CRTH2 MicroBead Kits, T-cell activation/expansion kits, and anti-human CRTH2 antibody were from Miltenyi Biotec (Bergisch Gladbach, Germany). X-VIVO 15 medium was purchased from Lonza. AIM V medium was purchased from Invitrogen (Carlsbad, Calif). HBSS was from Gibco (Carlsbad, Calif). Ficoll-Paque Plus was supplied by GE Healthcare (Pittsburgh, Pa). Lymphoprep was purchased from Axis-Shield UK (Dundee, United Kingdom). The RNeasy Mini kit and Omniscript Reverse Transcription kit were supplied from Qiagen. Real-time qPCR Master Mix and probes were from Roche. Primers were synthesized by Eurofins MWG Operon (Ebersberg, Germany). Human rIL-8, anti–IL-8, and anti-CD54 (ICAM-1) antibodies were purchased from R&D Systems. Anti-CD31 (PECAM-1) antibody was from Thermo Fisher Scientific (Waltham, Mass). Anti-human GM-CSF, CD3, CD4, CCR6, CCR7, CD45RO, IL-4, IL-5, IL-8, IL-13, and IFN-γ antibodies and Annexin V–allophycocyanin were obtained from BioLegend (San Diego, Calif). Anti-human CD16, IL-22, and GATA3 antibodies and the viability dye eFluor 780 were from eBioscience (San Diego, Calif). Anti-human RORγt antibody was from BD Biosciences. Human rIL-2, human rIL-4, and human rGM-CSF were from PeproTech (Rocky Hill, NJ). Other chemicals were from Sigma-Aldrich (St Louis, Mo).

T_H2 cell preparation

Human CD4⁺CRTH2⁺ T_H2 cells were prepared, as described in our previous reports.^E1 Briefly, PBMCs were isolated from buffy coats by using Ficoll-Paque Plus density gradient centrifugation, followed by CD4⁺ cell purification with the MACS CD4⁺ T Cell Isolation Kit II. After a 7-day culture in AIM V medium containing 10% human serum, 50 U/mL rhIL-2, and 100 ng/mL rhIL-4, CRTH2⁺ cells were isolated from the CD4⁺ culture by means of positive selection with the anti-human CRTH2 MicroBead Kit. The harvested CD4⁺/CRTH2⁺ cells were treated as T_H2 cells and further amplified in X-VIVO 15 medium containing 10% human serum and 50 U/mL rhIL-2 before use. For ex vivo T_H2 cells, CRTH2⁺ cell isolation was directly conducted after CD4⁺ cell purification without culture.

Flow cytometric analysis for T_H2 cell phenotype

For surface marker staining, T_H2 cells or PBMCs from fresh blood were fluorescently labeled with antibodies to CD3, CD4, CRTH2, CCR6, CCR7, CD45RA, and CD45RO in PBS containing 0.2% BSA and 2 mmol/L EDTA and then fixed with 2% formaldehyde. Transcription factor staining was conducted in permeabilization buffer with 0.05% saponin-containing antibodies to GATA3 and RORγt. For analysis of intracellular cytokines, T_H2 cells were stimulated with PMA (5 ng/mL) and ionomycin (500 ng/mL) for 4 to 5 hours. Brefeldin A (5 μg/mL) was added at 30 minutes after starting stimulation. Then the cells were stained in a permeabilization buffer containing antibodies to IFN-γ, IL-4, IL-5, IL-8, IL-13, and IL-22. Nonstimulated cells were used as a negative control, and dead cells were excluded by using the viability dye eFluor 780. After staining, the cells were acquired with an LSR II Flow Cytometer (BD Biosciences).

Western blotting

T_H2 cells were treated for 15 minutes. Cell lysis and Western blotting were performed, as described previously.^E2 Briefly, the cells were solubilized in lysis buffer (20 mmol/L Tris-HCl [pH 7.4], 250 mmol/L sucrose, 1 mmol/L EDTA, 1 mmol/L ethyleneglycol-bis-[β-aminoethylether]-N,N,N′,N′-tetraacetic acid, 1 mmol/L sodium orthovanadate, 10 mmol/L sodium glycerophosphate, 50 mmol/L sodium fluoride, 5 mmol/L sodium pyrophosphate, 0.1% 2-mercaptoethanol, protease inhibitor mixture, and 1% Triton X-100). The samples were fractionated by using SDS-PAGE and then electrophoretically transferred to a nitrocellulose membrane and probed with antibodies, as indicated in the Results section. The intensity of immunopositive bands was quantified by using ImageJ software (National Institutes of Health, Bethesda, Md).

Quantification of cell aggregates

Briefly, cell clumps were identified by using the Otsu method of thresholding. The intensity and size of identified objects were measured. Results are reported as integrated intensity.