Interleukin-4 inhibits cyclo-oxygenase-2 expression and prostaglandin E₂ production by human mature dendritic cells

Abstract

Interleukin-4 (IL-4) is considered the key cytokine for inducing T helper type 2 (Th2) cell differentiation, while interferon-γ and IL-12 are pivotal cytokines for Th1 immune responses. Paradoxically, IL-4 has also been demonstrated to enhance IL-12 production by dendritic cells, suggesting an IL-4-dependent regulatory feedback of the Th1/Th2 system. In addition, prostaglandin E₂ (PGE₂), a lipid mediator of inflammation, has been implicated in the enhancement of Th2-type responses acting directly on T and B lymphocytes. PGE₂ synthesis is dependent on the serial engagement of various enzymes, among which the inducible cyclo-oxygenase-2 (COX-2) exerts a critical role in monocytes and dendritic cells. In this study we demonstrate that IL-4 inhibits COX-2 gene expression and consequently prevents secretion of PGE₂ by mature human dendritic cells. We also show that PGE₂ does not regulate IL-12 and IL-10 production by dendritic cells in an autocrine fashion. Hence, we suggest that IL-4 may exploit an IL-12-independent regulatory feedback of the Th1/Th2 system through PGE₂ inhibition.

Introduction

Prostaglandins are potent lipid mediators that regulate several physiological processes in both homeostatic and inflammatory conditions.^1–3 The production of prostaglandins starts with the release of arachidonic acid from membrane phospholipids by phosholipase A₂. Arachidonic acid is then converted into prostaglandin H₂ (PGH₂) by two isoenzymes, cyclo-oxygenase-1 (COX-1) and COX-2.⁴ Lastly, cell-specific prostaglandin synthases convert PGH₂ into a series of prostaglandins, including PGI₂, PGF_2α, PGD₂ and PGE₂. COX-1 is expressed constitutively in most tissues of the body and is primarily responsible for cellular homeostasis, whereas COX-2 is the early inducible isoform, which is up-regulated in ‘inflammatory cells’ such as macrophages, vascular endothelial cells and fibroblasts, in response to cytokines, hormones, and mitogenic or inflammatory stimuli.^5–7 COX-2 induction preferentially shifts arachidonic acid metabolism to the selective synthesis of two prostanoids, PGI₂ and PGE₂.⁸

Several studies have described PGE₂ as a ‘regulatory’ prostanoid in a variety of cellular processes during the immune response.^3,9,10 In particular, it has been demonstrated that it controls many functions of professional antigen-presenting cells.^11–16 Indeed, when added exogenously, it inhibits IL-12 production by dendritic cells (DC) in response to inflammatory stimuli^17–19 but it improves their maturation and migration in vitro.^20,21 Moreover, it inhibits interferon-γ (IFN-γ) production by T lymphocytes while enhancing T helper type 2 (Th2) release of cytokines such as interleukin-4 (IL-4) and IL-5.^17,22–25 Also, B cells are influenced by PGE₂, which can increase immunoglobulin G1 (IgG1) synthesis after lipopolysaccharide (LPS) stimulation or promote the immunoglobulin isotype switch to IgE.^26,27

Several cytokines, such as IL-1α, transforming growth factor-β, IL-4 and IL-10 have been described as regulating COX-2 induction and PGE₂ production.^7,28 In particular, IL-4 and Il-10 inhibit COX-2 expression in human neutrophils and monocytes by means of transcriptional and post-transcriptional regulation.^29,30

IL-4 is a multifunctional cytokine produced by eosinophils, mast cells, Th2 cells and natural killer T cells. In spite of its pivotal role in promoting Th2 differentiation of naive T cells, IL-4 has also been shown to enhance the production of IL-12p70 by DC, instructing them to induce a Th1 polarization.^31,32 This latter IL-4 ability would allow a regulatory feedback mechanism of the Th1/Th2 system during a developing immune response.

DC are the central target of immune regulation and many cytokines act on their function to manipulate T-cell priming and polarization.³³ In this work we asked whether IL-4 would accomplish its regulatory feedback mechanism of the Th1/Th2 system by inhibiting COX-2 expression and function in DC, thus affecting the production of a ‘Th2-oriented prostanoid’ such as PGE₂.

Materials and methods

Reagents

Human recombinant IL-4 was produced by polymerase chain reaction (PCR) cloning and expression in the myeloma expression system.³⁴ Granulocyte–monocyte colony-stimulating factor (GM-CSF) (Leucomax) was purchased from Sandoz (Basel, Switzerland). RPMI-1640 (Euroclone Ltd, Wetherby, UK) was used supplemented with 100 U/ml kanamycin, 1 mm glutamine, 1 mm sodium pyruvate, 1% non-essential amino acids, and 10% fetal bovine serum (Hyclone, Logan, UT) (complete medium). NS398 and indomethacin were from Cayman Chemical Co (Ann Arbor, MI). LPS from Escherichia coli 055:B5 was purchased from Sigma Chemical Co. (St Louis, MO).

Monocyte isolation and DC generation

Peripheral blood mononuclear cells were purified from heparinized blood on a density gradient (Lymphoprep, Nycomed Pharma AS, Oslo, Norway). Monocytes were sorted using anti-CD14-labelled magnetic beads (magnetic antibody cell sorting, Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions and were cultured for 5 days in complete medium containing GM-CSF (50 ng/ml) and IL-4 (1000 U/ml) at a concentration of 4 × 10⁵ cells/ml. DC were harvested and stimulated with 0·1 μg/ml LPS for 4 hr or overnight in complete medium (IL-4 DC). Alternatively, DC were extensively washed and immediately stimulated with LPS, or were stimulated after an additional 6 hr of culture in complete medium deprived of IL-4 and GM-CSF (depDC).

Fluorescence-acitvated cell sorting analysis

For surface staining the following monoclonal antibodies were used: fluorescein isothiocyanate (FITC)-conjugated mouse anti-human CD1a, CD86, CD80, phycoerythrin-conjugated mouse anti-human CD14 and appropriate isotype controls for background determination (all from BD Biosciences, Pharmingen San Diego, CA). DC were harvested and washed in phosphate-buffered saline containing 1% fetal bovine serum and 0·1% NaN₃ (staining buffer) and then stained using the antibodies described above.

Staining of intracellular COX-2 was performed using FITC-conjugated mouse anti-human COX-2 (Cayman Chemical, Ann Arbor, MI), after fixation and permeabilization using Cytofix/Cytoperm™ (BD Biosciences, Pharmingen San Diego, CA), according to the manufacturer's instructions. A FITC-conjugated mouse IgG1 for intracellular staining was used as isotype control.

Stained cells were analysed using a FACScan cytometer (Becton Dickinson, Mountain View, CA) equipped with Cellquest software (Becton Dickinson). Fluorescence intensity was evaluated by computerized analysis of dot plots or histograms generated by 5000 viable cells.

Reverse transcription and real-time polymerase chain reaction (PCR)

Total ribonucleic acid (RNA) was extracted from 10⁶ DC, DNased and reverse transcribed as described in Gagliardi etal.³⁵ Quantitative real-time PCR of DC cDNA was performed in 20 μl with iQ SYBR Green Supermix (Bio-Rad, Hercules, CA) in an iCycler iQ (Bio-Rad) for 40 cycles as follows: 40 min at 95°, 40 min at 57° and 60 min at 72°. COX-2 and β-actin mRNAs were amplified in separate tubes using human COX-2 sense (5′-GCTGCTGAATTTAACACCCTCTATC-3′) and antisense (5′-CTGCCTGCTCTGGTCAATGG-3′) primers and human β-actin sense (5′-TCCTGTGGCATCCACGAAACT-3′) and antisense (5′-GAAGCATTTGCGGTGGACGAT-3′) primers. Primers were designed to flank introns to discriminate between cDNA and genomic DNA products. Each RNA sample was tested by real-time PCR before reverse trascription to exclude genomic DNA contamination. A final melting curve was generated with temperature increases from 55° to 95° to analyse the amplification products. Each cDNA was tested in triplicate. A non-template control was run in every assay. Threshold cycle data were collected, ΔC_t was calculated as the C_t difference between COX-2 and β-actin amplification curves and comparative gene expression was established by the ΔΔC_t method. Finally, data were expressed as the fold increase using the 2^−ΔΔC_t formula. P-values were determined by uncoupled two-tailed t-test.

PGE₂ and cytokine production analysis

IL-4 DC and depDC were harvested on day 5 of culture, adjusted to 4 × 10⁵ cells/ml and stimulated with 0·1 μg/ml LPS for 24 h in the presence or absence of 10 μg/ml NS398 or indomethacin. Supernatants were examined for IL-10 and IL-12 by enzyme linked immunosorbent assay using commercially available kits (R&D Systems Europe, Ltd, Abingdon, UK) according to the manufacturer's instructions; the detection limit of the assay was 15 pg/ml. PGE₂ was detected using a competitive enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI); the detection limit of the assay was 31 pg/ml.

Results

IL-4 inhibits the synthesis of COX-2 in DC

Previous studies have reported that IL-4 inhibits COX-2 expression in human monocytes and neutrophils.^29,30

In view of the fact that IL-4 is used to generate human monocyte-derived DC, we asked whether it could affect COX-2 expression also in these cells. Intracellular staining in Fig. 1 shows that immature DC (no stimulus) cultured in medium containing IL-4 and GM-CSF (IL-4 DC) do not express COX-2. Moreover, LPS addition during the last 4 hr of culture, while inducing DC to undergo functional and phenotypical maturation (Table 1), does not induce COX-2 synthesis.

Figure 1.

Detection of intracellular COX-2 in human DC by flow cytometry. DC on day 5 of culture were harvested and immediately stimulated with LPS for 4 hr (IL-4 DC), or after three washes (wDC), or after 6 hr of culture in complete medium deprived of IL-4 and GM-CSF (depDC). No stimulus: immature DC. Staining of intracellular COX-2 was performed using FITC-conjugated mouse anti-human COX-2 after DC fixation and permeabilization. The dotted histograms represent staining with FITC-conjugated isotype control monoclonal antibody. Numbers in the histogram plots indicate the mean fluorescene intensity. One representative experiment out of seven is shown.

Table 1.

Flow cytometric analysis of DC maturation markers

	CD1a Mean1 (%)2	CD83 Mean (%)	CD86 Mean (%)
IL-4 DC
Alone	302 (82)	3 (6)	19 (28)
+ LPS	280 (80)	15 (71)	149 (95)
depDC
Alone	220 (79)	9 (10)	30 (32)
+ LPS	186 (70)	18 (63)	180 (90)

¹Values indicate the mean fluorescence intensity (mean) of one representative experiment out of seven.

²Values indicate the percentage of positive cells.

To test whether the expression of COX-2 would be increased by removing IL-4 and GM-CSF from the culture, DC were extensively washed before their stimulation with LPS (washed DC: wDC). As shown in Fig. 1, wDC up-regulated COX-2 in response to LPS, but only weakly. Interestingly COX-2 expression dramatically increased when DC were stimulated with LPS after the extensive washing and additional 6-hr culture (IL-4/GM-CSF-deprived DC: depDC). None of the markers of DC differentiation and maturation, i.e. expression of CD1a as well as CD83 induction and CD86 up-regulation after LPS stimulation, were affected by the 6-hr deprivation of IL-4 and GM-CSF as shown in Table 1.

These results suggest that growth factors' deprivation is required for the switching off in DC of a putative repressor mechanism induced by IL-4.

In another series of experiments, DC were extensively washed and cultured for 6 hr in the absence of IL-4, but in the presence of GM-CSF. When only IL-4 deprivation was performed, LPS-matured DC showed an expression of COX-2 comparable to that of IL-4 and GM-CSF-deprived DC (data not shown), indicating that GM-CSF is not involved in COX-2 inhibition.

IL-4 inhibits PGE₂ production by DC

To investigate the outcome of COX-2 inhibition by IL-4 we measured PGE₂ production by DC. As shown in Fig. 2, IL-4 DC do not synthesize PGE₂ in response to LPS, confirming at a functional level the cytofluorimetric results. Only LPS-matured depDC produced high amounts of PGE₂, which was drastically reduced when both NS398, a selective COX-2 inhibitor, or indomethacin, a non-selective COX inhibitor, were added together with LPS, suggesting that PGE₂ synthesis occurs principally through COX-2. COX-inhibitors alone do not have any effect on immature DC (data not shown). The absolute amounts of PGE₂ produced by depDC from different donors varied considerably, but the difference due to IL-4 deprivation was always superimposable.

Figure 2.

Analysis of PGE₂ production by DC in response to LPS. IL-4 DC and depDC were harvested, adjusted to 4 × 10⁵/ml and left unstimulated (no stimulus) or stimulated overnight with 0·1 μg/ml LPS, in the presence or absence of 10 μg/ml indomethacin (INDO) or NS398. Supernatants were examined for PGE₂ by a competitive enzyme immunoassay. The detection limit of the assay was 31 pg/ml. Values indicate the mean ± SD of four independent experiments.

IL-4 inhibits COX-2 mRNA expression by DC

To determine whether the IL-4-dependent inhibition of COX-2 expression was paralleled by a decreased mRNA abundance, COX-2 mRNA steady-state levels were measured by real-time reverse transcription PCR. COX-2 mRNA was barely detectable in immature IL-4 DC and depDC and, as shown in Fig. 3, it was up-regulated after LPS stimulation in the two populations. But strikingly, IL-4 deprivation dramatically increased COX-2 mRNA levels in LPS-matured depDC. These data prove that IL-4-induced suppression of COX-2 gene expression is mainly acting at the level of transcription or mRNA stability. However, the increase of COX-2 mRNA steady-state levels in LPS-matured IL-4 DC was not sufficient to detect COX-2 protein and PGE₂ production, suggesting that additional mechanisms of translation impairment or decreased protein stability may be involved.

Figure 3.

Analysis of COX-2 mRNA steady-state levels in human DC by SYBR Green quantitative real-time RT-PCR. The cDNAs were synthesized from DC total RNA and aliquots were tested with oligonucleotides specific for human COX-2 and β-actin genes as described in the Materials and methods. Each COX-2 C_t value was normalized to β-actin by calculating the ΔC_t. Differential COX-2 gene expression in different cDNAs was calculated by the ΔΔC_t method and values were expressed as fold increase compared to baseline levels (IL-4 DC). Each sample was tested in triplicate. Data shown are representative of one of three independent experiments with similar results. P < 0·03 compared to IL-4 DC + LPS.

Autocrine PGE₂ does not regulate cytokine production by DC

It has been previously demonstrated that PGE₂ regulates the production of cytokines by DC and macrophages, thus shaping the immune response.^{2,10,18,19,36}

For this reason we asked whether depDC would display an altered cytokine profile by means of an autocrine PGE₂-mediated feedback regulation. Interestingly, depDC barely produced IL-12 but synthesized significant amounts of IL-10 in response to LPS stimulation as compared to IL-4 DC (Fig. 4). However, the treatment with NS398 or indomethacin, which blocked the synthesis of PGE₂, did not cause a decrease in IL-10 or a restoration of IL-12 production by depDC. These data rule out the possibility that endogenous PGE₂ could per se modulate the DC cytokine profile, or that the reduction of IL-12 is a consequence of a PGE₂-dependent increase of IL-10.

Figure 4.

Analysis of cytokine production by IL-4 DC and depDC. IL-4 DC and depDC were harvested, adjusted to 4 × 10⁵/ml and left unstimulated (no stimulus) or stimulated overnight with 0·1 μg/ml LPS, in the presence or absence of 10 μg/ml NS398. Supernatants were examined for IL-10 and IL-12 by enzyme-linked immunosorbent assay. The detection limit of the assay was 15 pg/ml. Values indicate the mean ± SD of six independent experiments.

Discussion

In this paper we demonstrate that IL-4 inhibits COX-2 expression in human DC, affecting their capacity to produce PGE₂. We show that IL-4 deprivation from the culture dramatically increases COX-2 mRNA levels and protein expression in LPS-stimulated DC. The depDC recover the capacity to produce PGE₂, which is known to play an importantant role in many DC functions as well as in T-cell activation and differentiation.^2,9

Previous results by Zelle-Rieser etal.³⁷ have also shown that IL-4 suppresses PGE₂ biosynthesis in human DC by inhibiting the cytoplasmic form of phospholipase A2, the enzyme that specifically liberates arachidonic acid from membrane phospholipids. These data, together with ours, suggest a regulation at different steps of PGE₂ biosynthesis and a more generalized effect of IL-4 on the production of arachidonic acid derivatives by DC.

In addition, we confirmed that IL-4 increases IL-12 synthesis by DC because LPS-matured depDC produce higher amounts of IL-10 than IL-4 DC and their IL-12 production is drastically decreased. This unexpected effect of IL-4 has already been demonstrated by Hochrein etal.³¹ and recently it was described as being dependent on IL-4 ability to decrease histone acetylation of the IL-10 promoter.³² We also show that the peculiar cytokine profile of depDC is not regulated by PGE₂ synthesis in an autocrine fashion. In fact, the treatment with COX inhibitors did not cause a decrease in IL-10 or restoration of IL-12 production by depDC. These data are not in line with previous studies reporting a suppression of IL-12 production mediated directly by endogenous PGE₂³⁶ or indirectly through its effects on IL-10.³⁸ We suggest that these apparent discrepancies may be caused by differences in the stimuli used to mature DC and favour cytokine production. It could also be hypothesized that endogenous PGE₂ will accumulate to levels required for cytokine gene regulation in 24–48 hr, and by that time the effect of LPS on the cytokine could be already full-blown. Human monocytes differentiate into DC when cultured in vitro with GM-CSF and IL-4.³⁹ Beside its crucial role for monocyte differentiation into DC, IL-4 has been shown to exert additive effects on both monocytes and DC.^{16,31,32,40,41} In particular, it is tempting to propose that IL-4, i.e. the prototype Th2 cytokine, would promote a regulatory feedback mechanism, during an ongoing Th2 response, through the inhibition of another Th2-inducing factor such as PGE₂. Kalinsky etal. showed that high concentrations of PGE₂ during the development of DC profoundly suppress their IL-12 production favouring a Th2 cell priming.¹⁷ In this model, the IL-4 inhibitory effect on PGE₂ production by monocytes and neutrophils, which represent an important source of prostaglandins in inflamed tissues, could indirectly influence DC function. In addition to this regulatory mechanism, we propose that IL-4 inhibition of endogenous PGE₂ release by DC could directly affect their capacity to induce Th2 cell priming, because only mature DC reach the secondary lymphoid organs where PGE₂ would act mainly on naive T cells by inhibiting the hypomethylation of the 5′ regulatory region of the IFN-γ gene²² and increasing IL-4 and IL-5 production.²³ On the other hand, polarized Th1 or Th2 effector cells that have migrated into the inflamed tissues seem to be resistant to PGE₂ regulation, because of lower responsiveness of the cyclic adenosine-monophosphate-linked PGE₂ receptors expressed on these cells.⁴² Hence PGE₂-disabled DC, such as those stimulated by IL-4 in the inflamed tissues, would favour a Th1 cell priming, counterbalancing a developing Th2 response.

In conclusion, we suggest that IL-4 exploits two different strategies to autoregulate an ongoing Th2 immune response: the up-regulation of IL-12 and the inhibition of PGE₂ production by DC.

Glossary

Abbreviations:

cAMP

cyclic adenosine-mono-phosphate

COX

cyclo-oxygenase

FITC

fluorescein isothiocyanate

GM-CSF

granulocyte–monocyte colony-stimulating factor

IFN

interferon

Ig

immunoglobulin

IL

interleukin

LPS

lipopolysaccharide

PCR

polymerase chain reaction

PG

prostaglandin

RNA

ribonucleic acid