Platelet-Activating Factor Induces Dual-Specificity Phosphatase 1 and 5 Gene Expression

Platelet-activating factor (PAF) is a potent inflammatory phospholipid mediator that is known to play a role in early-phase responses in asthma and other diseases. Through its high affinity receptor, PAFR, PAF is known to activate multiple signalling pathways contributing to its proinflammatory effects. Of these pathways, the mitogen-activated protein kinase (MAPK) cascade is initiated upon PAF stimulation, leading to the activation of the conventional MAPKs ERK1/2, p38 and JNK. Since dual-specificity phosphatases (DUSP) downregulated MAPK activity, we postulated that PAF could also enhance DUSP expression and thus induced an autoregulatory loop. In this report, we studied the effect of PAF on DUSP mRNA expression in human monocytes. Our results demonstrate that PAF induces DUSP1 and DUSP5 gene expression in a timeand concentration-dependent manner, with maximal effects at PAF 100 nM and at 20 30 min of stimulation. In contrast, DUSP2 and DUSP6 gene expression was not enhanced by PAF. Moreover, leukotriene D4, another lipid mediator of inflammation, was unable to modulate DUSP expression. PAF-induced DUSP expression was prevented by the PAFR antagonist WEB2170 and by pretreatment with the transcriptional inhibitor Actinomycin D. Moreover, enhanced DUSP5, but not DUSP1 expression was prevented by pretreatment with the ERK inhibitor PD98059 or the PI3K inhibitor Wortmannin. Taken together, our results indicate that PAF selectively enhances DUSP1 and DUSP5 gene expressions through PAFR activation, and suggest that PAF may have an active role in the resolution of inflammation by its ability to upregulate the two DUSPs and thus provide a negative auto-regulatory signalling mechanism.


Introduction
Platelet activating factor (PAF) 1 is a potent phospholipid mediator that has a role in an array of inflammatory diseases [1].It is secreted by a wide variety of cells including leukocytes and endothelial cells [2].
PAF-induced effects are mediated through activation of its high-affinity, widely expressed receptor, PAFR, a G-protein-coupled receptor.PAF is known to induce the activation of multiple signalling pathways leading to expression of a number of cytokines and chemokines, thus contributing to the proinflammatory effects of the lipid.One of these, the mitogen-activated protein kinase (MAPK) cascade is initiated early upon PAF stimulation [3]- [5] leading to the activation of the conventional MAPKs ERK1/2, p38 and JNK [6]- [8].ERK1 and ERK2 are mainly implicated in the proliferation and differentiation processes, whereas p38 is associated with pro-inflammatory cytokine production and JNK, with regulation of extracellular matrix [9].
In contrast, dual-specificity phosphatases (DUSPs) are involved in preventing MAPK signalling.In mammalian cells, DUSPs are represented by 10 catalytically active phosphatases divided in three sub-families.Of these, the nuclear mitogen kinase phosphatase 1 (MKP-1) (encoded by the DUSP1 gene), a member of the Stress-Inducible sub-family [10], is the prototypical member of the DUSP family that dephosphorylates MAPKs (JNK, p38, >ERK1/2) in a cell-type dependent manner.DUSP1 has been linked to innate immunity and inflammatory diseases, as DUSP1 -/-mice show enhanced inflammatory cytokine levels [11] [12].Moreover, subjects with severe asthma demonstrate lower DUSP1 expression compared to control subjects [13].On the other hand, VH3 (encoded by DUSP5 gene), may also be involved in resolution of inflammation through its ability to inhibit and anchor nuclear ERK1/2 [14].Furthermore, DUSP5 was shown to be upregulated in cases of sustained inflammation [15] and was presented as a key anti-apoptotic molecule in activated eosinophils [16].
We postulated that PAF could enhance DUSP expression and thus induce a negative, autoregulatory loop.Here, we demonstrate that PAF induces DUSP1 and DUSP5 mRNA upregulation in a time-and concentrationdependent manner in human monocytes through a mechanism involving receptor-mediated signalling and gene transcription.

Cells
Human monocytes were isolated from peripheral blood mononuclear leukocytes (PBMLs) obtained from healthy donors after informed consent, in accordance with an internal review board-approved protocol, according to the technique described by Böyum [17].PBMLs were firstly enriched by sedimentation using Dextran T500 (Pharmacosmos, Holbaek, Denmark) followed by a density gradient centrifugation on Ficoll-Paque™ Plus (GE Hearthcare, Mississauga, ON, Canada).Mononuclear cells were collected at the interface of the Ficoll and monocytes were then purified by petri dish adherence using dishes previously coated with defibrinized autologous serum.Monocytes were resuspended in RPMI 1640 medium (Gibco ® , Burlington, ON, Canada) supplemented with 10% fetal bovine serum (FBS) (PAA, Piscataway, NJ, USA), 100 μg/mL penicillin G (PPC inc., Richmond Hill, ON, Canada) and 100 μg/mL streptomycin (Sigma-Aldrich, St. Louis, MO, USA).Cells were rested overnight under normal conditions-37˚C + 5% CO 2 -in round bottom 15 mL polypropylene tubes at 1 × 10 6 monocytes/mL.
Before experiments, cells were centrifuged and resuspended at 2 × 10 6 monocytes/mL and starved for 3 hours in RPMI 1640 medium without serum.Cells were then pre-treated, as described, with WEB2170, Actinomycin D, Wortmannin or MAPK inhibitors for 30 minutes, and stimulated for the appropriated time with PAF or LTD 4 at indicated concentrations.

Reagents
Reagents used in this paper were purchased from the following sources: PAF C-16, LTD 4 , SB203580, PD98059, SP600125 and Wortmannin were purchased from Cayman Chemical (Ann Arbor, MI, USA).WEB2170 was from Boehringer Ingelheim (Burlington, ON, Canada).Actinomycin D was purchased from Biovision, Inc. (Milpitas, CA, USA).

RNA Isolation and RT-PCR
Total RNA was purified using Trizol ® (Life Technologies™, Burlington, ON, Canada) according to the manufacturer's instructions using the classical phenol/chloroform technique.To exclude genomic DNA contamination, 1.5 μg of RNA was digested with RNasin ® (Promega, Madison, WI, USA) and DNase 1 (Thermo Scientific, Ottawa, ON, Canada).The reaction was stopped with 25 μM EDTA.First-strand cDNA synthesis was performed on this pre-treated RNA using oligo(dT) 18 primer (Thermo Scientific) and M-MLV reverse transcriptase (Promega) in a final volume of 22 μL.

Statistical Analyses
Statistical analyses were performed using Prism 6.0 software (GraphPad).Data were analyzed using one way ANOVA or one-tailed, paired Student's t test as appropriate.A p value of <0.05 was considered statistically significant.

Human Monocytes Express Mainly DUSP1 and DUSP5
Using human monocytes, we first studied the mRNA expression levels of four common DUSPs (DUSP1, DUSP2, DUSP5, DUSP6).Unstimulated human monocytes mainly expressed DUSP1 and DUSP5 as compared to DUSP2 and DUSP6 (Figure 1).Both DUSP1 and DUSP5 have been implicated in inflammation [12] [15] and DUSP1 expression was shown to be upregulated in airway smooth muscle cells and in monocytes/macrophages in response to inflammatory stimuli [19] [20].We thus proceeded to study the potential effect of PAF on DUSP1 and DUSP5 gene expression.

PAF Induces DUSP1 and DUSP5 Expression
When human monocytes were exposed to PAF 10 nM for periods of time between 10 to 120 minutes, both DUSP1 and DUSP5 gene expression was significantly enhanced, but with an earlier and greater effect on DUSP1 expression (Figure 2(a), Figure 3(a)).Maximal expression of DUSP1 was observed between 20 to 30 minutes, whereas that of DUSP5, between 30 to 45 minutes, with a slow recovery to the basal expression state from 30 to 120 minutes.However, both phosphatases showed the same expression pattern when exposed to graded concentrations of PAF (0.001 nM to 100 nM) for 30 minutes with a maximal response at PAF 100 nM (Figure 2(b), Figure 3(b)).
Moreover, when human monocytes were pre-incubated for 30 minutes with WEB2170 (10 μM), a PAFR   antagonist, the PAF-induced increase in DUSP1 and DUSP5 mRNA expression was abolished (Figure 2(c), Figure 3(c)).Pretreatment with the transcriptional inhibitor Actinomycin D (10 μM) also blocked DUSP mRNA upregulation upon PAF stimulation (Figure 2(d), Figure 3(d)).Thus, our results demonstrate that PAF, acting through its high affinity receptor PAFR, induced DUSP1 and DUSP5 mRNA expression in a time-and concentration-dependent manner, and through a transcription-dependent process.Moreover, the PAF effect was selective, as DUSP2 and DUSP6 expression was not affected upon PAF stimulation (data not shown).

LTD 4 Does Not Induce DUSP Expression
We next investigated whether another inflammatory lipid mediator could also modulate DUSP expression.Since cysteinyl-leukotrienes are also involved in asthmatic symptoms [21], we exposed human monocytes to LTD 4 (100 nM) from 30 minutes to 24 hours, but could observe no significant DUSP mRNA modulation (data not shown).

PAF-Induced DUSP5, But Not DUSP1, Expression Is Dependent on PI3K and MAPK Activation
We next investigated the PAFR signalling pathways leading to DUSP1 and DUSP5 mRNA upregulation.We first determined whether DUSP1 and DUSP5 mRNA upregulation was dependent on the activation of the phosphatidylinositol 3-kinase (PI3K).To this end, Wortmannin (0.1 μM), a PI3K inhibitor, was added 30 minutes before stimulation with PAF (10 nM).As shown in Figure 4, inhibition of PI3K activation abolished PAF-in-duced DUSP5, but not DUSP1 mRNA upregulation.Therefore, DUSP1 and DUSP5 expression appears not to be upregulated via the same pathway.Subsequently, the potential role for the MAPK signalling cascade was investigated.As shown in Figure 5(a)-(c), the use of PD58059 (10 μM), a MEK1/2 inhibitor, or the use of SP600125 (5 μM) or SB203580 (10 μM), inhibitors of the JNK and p38 pathways, respectively, failed to affect DUSP1 mRNA expression upon PAF (10 nM) stimulation.Similarly, combinations of MAPK inhibitors had no effect on DUSP1 mRNA expression (data not shown).
In contrast, our results demonstrate that the MEK/ERK and JNK pathways may be involved in the modulation of DUSP5 mRNA expression as the use of either PD58059 (10 μM) or SP600125 (5 μM) inhibited PAF-induced DUSP5 mRNA expression (Figure 5(d)-(e)).Inhibition of p38, however, was ineffective (Figure 5(f)).Combinations of MAPK inhibitors revealed no further effects on DUSP5 mRNA expression (data not shown).

Discussion
PAF, a potent inflammatory lipid, is known to be implicated in a wide variety of inflammatory diseases.Through its high affinity receptor PAFR, PAF can induce multiple signalling pathways leading to expression of a number of cytokines and chemokines, thus contributing to the proinflammatory effects of the lipid.Of these pathways, the mitogen-activated protein kinase (MAPK) cascade initiated upon PAF stimulation [3]- [5] activates the conventional MAPKs ERK1/2, p38 and JNK [6]- [8] that can subsequently be inactivated with selective DUSPs.
Here, we show that human monocytes mainly express DUSP1 and that PAF enhances DUSP1 and DUSP5 gene expression in a time-, concentration-and transcription-dependent manner.Our results show that PAF (10 nM) induces maximal DUSP1 mRNA expression as early as 20 to 30 minutes, whereas maximal DUSP5 expression is observed at 30 to 45 minutes.Indeed, DUSP1 has been characterized as an early response gene [22], whereas DUSP5 was described as a late gene [23].However, despite differences in time-dependent mRNA modulation, the expression of both phosphatases was similar in a concentration-response to PAF (0.001 nM to 100 nM).We also demonstrated that both DUSP1 and DUSP5 mRNA upregulation was dependent on PAFR activation and on gene transcription.
Human monocytes show a higher expression level of DUSP1 than DUSP2, DUSP5 or DUSP6.This could explain why DUSP1 gene expression is more strongly upregulated when cells are stimulated with PAF.DUSP1 may thus be a leading candidate for resolution of PAF-induced inflammation as it is rapidly and significantly induced by PAF.
In concordance with previous data linking DUSP1 and DUSP5 with inflammatory diseases, PAF-induced DUSP1 and DUSP5 gene upregulation may be relevant to the mechanisms of asthma.Notably, DUSP1 was first linked to innate immunity in DUSP1 -/-mice in which elevated cytokine levels were quantified [12] [24].It was also shown that Dexamethasone, an anti-inflammatory glucocorticoid, enhances DUSP1 expression in order to transiently repress inflammatory genes implicated in human asthma [25].The phosphatase was further characterized as an immune modulator by its capacity to mediate the anti-inflammatory effect of glucocorticoids in airways smooth muscle cells by inhibition of TNFα-induced pro-inflammatory cytokines [19] [26].Moreover, a DUSP1 polymorphism (rs881152) was linked to a better response of asthmatic patients to inhaled corticosteroids [27].In contrast, subjects with severe asthma were found to express significantly less DUSP1, resulting in an increased p38 activation in alveolar macrophages, and enhanced asthmatic symptoms [13].On the other hand, the less common DUSP5 may play a different role in inflammatory pathologies such as human asthma.DUSP5 may be involved in eosinophil survival, as an anti-apoptotic signalling component [16], and its expression was reported to be elevated in sustained inflammation [15].PAF-induced expression of DUSP1 and DUSP5 appears to be selective, as DUSP2 and DUSP6 mRNA levels were not modulated upon PAF stimulation.DUSP2 is a member of the Stress-Inducible MAPK phosphatases, along with DUSP1.However, in contrast to DUSP1, DUSP2 -/-mice were found to have less inflammation in an autoimmune model of rheumatoid arthritis, suggesting that DUSP2 may have a positive activating effect on immune cells and inflammatory responses [28].
As for DUSP6, it is an ERK-specific member of the second MAPK phosphatase sub-family [10] and appears not to be involved in the immune system.The evidence supplied by a DUSP6 -/-mouse model indicates that DUSP6 may be implicated in somatic development as dwarfism was found in these mice as well as a premature fusion of the cranial suture and a defect in ear bones [29].Also, DUSP6 is likely to be implicated in mood and bipolar disorders [11].

Conclusion
In conclusion, our results indicate that PAF selectively enhances DUSP1 and DUSP5 gene expression through PAFR activation, and suggest that PAF may have an active role in the resolution of inflammation by its ability