Endogenous prostaglandin D 2 synthesis inhibits e-selectin generation in human umbilical vein endothelial cells

We examined the role of prostaglandin D2 (PGD2) in the formation of E-selectin following interleukin-1 (IL-1) stimulation in human umbilical vein endothelial cells (HUVEC) transfected with lipocaline-type PGD2 synthase (L-PGDS) genes. HUVEC were isolated from human umbilical vein and incubated with 20 U/mL IL-1 and various concentrations of authentic PGD2. The isolated HUVEC were also transfected with L-PGDS genes by electroporation. The L-PGDS-transfected HUVEC were used to investigate the role of endogenous PGD2 in IL-1-stimulated E-selectin biosynthesis. We also used an anti-PGD2 antibody to examine whether an intracrine mechanism was involved in E-selectin production. PGD2 and E-selectin levels were determined by radio-immunoassay and enzyme-immunoassay, respectively. E-selectin mRNA was assessed by real-time RT-PCR. IL-1-stimulated E-selectin production by HUVEC was dose-dependently inhibited by authentic PGD2 at concentrations greater than 106 mol/L. L-PGDS gene-transfected HUVEC produced more PGD2 than HUVEC transfected with the reporter gene alone. IL-1 induced increases in E-selectin production in HUVEC transfected with the reporter genes alone. However, this effect was significantly attenuated in the case of IL-1 stimulation of HUVEC transfected with L-PGDS genes, and accompanied by an apparent suppression of E-selectin mRNA expression. Neutralization of extracellular PGD2 by anti-PGD2-specific antibody influenced neither E-selectin mRNA expression nor E-selectin biosynthesis. HUVEC transfected with L-PGDS genes showed increased PGD2 synthesis. This increase was associated with attenuation of both E-selectin generation and E-selectin mRNA expression. The results suggest that endogenous PGD2 decreases E-selectin synthesis and E-selectin mRNA expression, probably through an intracrine mechanism.


INTRODUCTION
Adhesion molecules play an important role in the development and progression of atherosclerosis.The hypothesis proposed by Ross and its modifications have been generally accepted as a mechanism of atherosclerosis where adhesion of circulating monocytes and lymphocytes to vascular endothelium presumably initiates a series of events toward atherosclerosis [1].Cellular adhesion molecules mediate the adhesion, margination, and transendothelial migration of circulating mononuclear cells from the blood stream to the extravascular compartment to have an important part in the progression of atherosclerotic plaque [2].Recent studies have elucidated further that to anchor leukocytes onto the endothelial cells, the adhesion molecules expressed on the surface of endothelial cells necessitated to bind to their ligands expressed on leukocytes.Endothelial cells are stimulated by inflammatory agents to express selectins, such as endothelial-leukocyte adhesion molecule-1 (E-selectin), which interact with carbohydrate ligands on leukocytes, and to express immunoglobulin superfamily proteins, such as vascular cell adhesion molecule-1 (VCAM-1).The adhesion molecules expressed on the endothelial cells include VCAM-1, intracellular adhesion molecule (ICAM)-1, P-selectin and E-selectin [3].Selectins, including E-selectin and P-selectin, are in-volved in the first step of leukocyte adhesion at sites of inflammation or injury.Selectins are characterized by rolling and tethering of leukocytes to the endothelial surface, to platelets or to other leukocytes [4].In fact, E-selectin is demonstrated to occur in atherosclerotic lesions in the coronary artery of humans [5].Therefore, E-selectin is believed to be a key factor for the development of immune-mediated cardiovascular injury.
Intrestngly, we have recently reported that PGD 2 attenuates inducible nitric oxide generation in vascular smooth muscle cells [6].Endogenous prostaglandin D 2 synthesis reduces plasminogen activator inhibitor-1 generation following cytokines stimulation in bovine endothelial cells [7].Especially, PGD 2 is synthesized in vascular components of atheromatous lesions including endothelial cells, macrophages, platelets, and mast cells [8] and lipocalin-type PGD 2 synthase (L-PGDS) is demonstrated to occur in atheromatous lesions in the cardiovascular system [9].These data strongly suggest that L-PGDS/PGD 2 is upregulated in response to immune-related vascular lesions and in turn, the increase of L-PGDS/PGD 2 is exerted to attenuate the progression of the arterial remodeling.
Taken together, we proposed the hypothesis that PGD 2 regulates E-selectin expression in endothelial cells, thereby contributing to leukocyte adhesion, an integral component of the development of vascular injury.However, there had been few data investigating the crosstalk between endogenous L-PGDS/PGD 2 system and the adhesion molecule expression by cytokines.In the present study, in order to test our hypothesis that L-PGDS/ PGD 2 protects the vascular wall against immune-related vascular injury, we examined the relationship between endogenous PGD 2 and E-selectin expression by endothelial cells and attempted to reveal its intracellular mechanism mediated by PGD 2 using L-PGDS genetransfected endothelial cells in culture.We also examined whether the increases in intracellular PGD 2 synthesis influenced E-selectin mRNA expression and E-selectin biosynthesis observed following interleukin-1b (IL-1) stimulation.

Cell Culture
Human umbilical vein endothelial cells (HUVEC) were harvested enzymatically as described previously [12].They were maintained in medium 199 (GIBCO BRL, Gaithersburg, MD), containing Hepes, heparin (1%), endothelial cell growth factor (50 mg/ml), L-glutamin (1%), antibiotics, and 5% fetal bovine serum (FBS).When the cells reached confluence, they were replanted onto low pyrogen fibronectin at 20,000 cells/ cm 2 .HUVEC which were isolated from a confluent monolayer of polygonal cells.The cells expressed von Willbrand factor as determined by their content of specific mRNA and immunoreactive protein.Cellular viability was assessed by Trypan blue exclusion.

Effect of Exogenous PGD 2 on E-Selectin Expression in Endothelial Cells
Cultures of HUVEC were treated with 20 U/ml IL-1, according to the previous study, in the presence of various concentrations of PGD 2 to be incubated for 18 h.Thereafter, the HUVEC were washed three times with FBS-free Dulbecco's phosphate-buffered saline (D-PBS; Gibco) and the cells were re-incubated in 1 ml of fresh D-PBS for 2 h.Subsequently, the cells and culture supernatants were used in various assays.

Transfection of L-PGDS Genes into HUVEC
HUVEC were transfected with L-PGDS genes using the Shimadzu GTE-10 electroporation device (Gene Transfer Equipment-10, Shimadzu Co., Ltd., Kyoto, Japan).This equipment transiently increases the permeability of plasma membranes of HUVEC, thereby facilitating translocation of genes into the cytoplasm.Briefly, the cells were washed three times with FBS-free D-PBS and 10 mg of pcD2-rat PGDS in 0.5 ml of fresh D-PBS were added to each well [13].A transient electrical current was applied onto HUVEC growing in the culture dishes, using a 35-mm round electrode (Model FTC-33D3, Shimadzu Co., Ltd., Kyoto, Japan), after which the cells were incubated for 30 minutes.
Following transfection, 2 ml of DMEM containing 10% FBS was added to the HUVEC and the dishes were incubated for an additional 24 hours.Endogenous PGD 2 production was stimulated by adding 10 -6 mol/l arachidonic acid for 24 hours.Fresh medium containing 10 -6 mol/l arachidonic acid was then added and E-selectin mRNA expression and E-selectin expression were stimulated for 18 hours by the addition of 20 U/ml IL-1.At the end of this incubation period, the HUVEC were washed three times with FBS-free D-PBS and incubated in 1 ml of fresh D-PBS for 2 hours.Finally, the cells and culture supernatants were collected for analysis.
For comparison, HUVEC were transfected with bgalactosidase genes (b-gal) by electroporation to determine the efficacy of gene transfection.Three days after the transfection, HUVEC were stained with X-gal [14], and b-galactasidase expression was measured.Transfection efficacy was estimated as the ratio of the X-gal stained area to the sectional area of the HUVEC.

Neutralizing Extrinsic PGD 2 Released from HUVEC
We attempted to neutralize PGD 2 using an anti-PGD 2specific antibody in order to investigate the effects of PGD 2 released from L-PGDS-transfected HUVEC.We estimated the amount of anti-PGD 2 antibody required to completely neutralize the secreted PGD 2 using the Scatchard analysis.The binding affinity was 0.0051 ml/pg, and the B max (maximal binding capacity) was 25 pg/l [15].Therefore, 1 liter of antibody had the capacity to bind 25 pg PGD 2 .Taking into account of these results, we used 200 ml of the antibody to inhibit the receptormediated actions of PGD 2 in HUVEC under our culture conditions.The anti-PGD 2 antibody was raised in our laboratories using PGD 2 -conjugated thyroglobulin and Freund's complete adjuvant.The antibody cross-reacted 0.003% with thromboxane B 2 , 0.01% with prostaglandin E 2 (PGE 2 ), 0.009% with prostaglandin F 2a (PGF 2a ), 0.008% with 6-keto-PGF 1a and 0.01% with arachidonate [16].Antibody activity was confirmed by suppression of intracellular cyclic AMP (cAMP) following PGD 2 stimulation via PGD 2 receptor, which acts as a second messenger for PGD 2 signal transduction [17].

Eicosanoid Radioimmunoassay
Eicosanoids were determined in culture media using the direct radioimmunoassay method described previously [16].Briefly, 0.1 ml of sample, 0.1 ml of [ 3 H] eicosanoid (5000 dpm) and 0.1 ml of the diluted antibody were mixed and incubated at 4˚C for 24 hours.To sepa-rate bound from free [ 3 H] eicosanoid, 0.1 ml of dextran-coated charcoal in a 50 mmol/l phosphate buffer at pH 7.4 containing 0.1% gelatin and 100 mmol/l NaCl was added to the ice-chilled assay mixture.The mixture was vortexed and centrifuged at 3000 rpm for 5 min at 4˚C.The supernatant was assayed for [ 3 H] eicosanoids bound to the antibody.Radioactivity was determined using an automatic liquid scintillation counter.The properties of the anti-6keto-PGF 1a antibody was described previously [15,16].The cross-reactivity and its properties of the anti-PGD 2 antibody was detailed above.The low cross-reactivity of each antibody made it feasible to measure directly the eicosanoid in media.

E-Secletin and E-Secletin mRNA Measurements
We measured E-selectin expression by commercially available cell surface enzyme immunoassay as described previously (R&D Systems, Inc., USA).
Cultured E-selectin transcripts were detected using the real-time reverse transcription-polymerase chain reaction (real-time RT-PCR) method using real-time RT-PCR machine as described previously [18].

Statistical Analysis
All values are expressed as the mean ± SE.The differences between values were assessed by an one-way ANOVA and Duncan's multiple range test using the STATISTICA program (StatSoft, Tulsa, OK, USA) on a Gateway G6-400 computer system (Gateway Inc., N Sioux City, SD, USA) running the Windows 98 operating system.P values less than 0.05 were considered statistically significant.

Effect of Exogenous PGD 2 on E-Selectin Expression in Endothelial Cells
Stimulation of endothelial cells with IL-1 significantly increased E-selectin expression.The increase was reduced in a dose-dependent manner by the addition of PGD 2 at concentrations ranging from 10 -7 to 10 -4 mol/l (Figure 1).

Gene Transfection and Eicosanoid Generation in HUVEC
We transfected HUVEC with L-PGDS genes in order to increase endogenous PGD 2 formation.PGD 2 was assayed using radioimmunoassay.The basal levels of PGD 2 in reporter-gene-transfected HUVEC maintained in arachidonate-free media were as low as 51.5 ± 2.2 pg/10 6 cells/2 hours, and the addition of 10 -6 mol/l arachidonate had no effect on PGD 2 synthesis.In contrast, L-PGDS gene-transfected HUVEC showed an increase (178.0%) in PGD 2 generation even in arachidonate-free media, as compared with control HUVEC carrying reporter genes alone.Furthermore, 10 -6 mol/l arachidonate markedly stimulated PGD 2 biosynthesis by 640.7%, as compared with control HUVEC carrying vector genes alone (Figure 2(a)).Thereafter, PGI 2 was assayed as 6-keto-PGF 1a using radioimmunoassay.Basal levels of prostacyclin (PGI 2 ), the major eicosanoid synthesized in HUVEC, were 1488 ± 66 pg/10 6 cells/2 hours in cells having reporter genes alone under arachidonate-free conditions.The PGI 2 generation was a markedly increased by 327% when the cells were stimulated with 10 -6 mol/l arachidonate (Figure 2(b)).L-PGDS gene transfection did not influence PGI 2 synthesis in HUVEC maintained under either arachidonate-free or arachidonate-stimulated conditions, as compared to HUVEC transfected with reporter genes.These data clearly suggest that the L-PGDS genes alter the phenotype of HU-VEC so that the recombinant cells acquire the capacity to produce PGD 2 in response to arachidonate stimulation.

Effect of L-PGDS Gene Transfection on E-Selectin Expression in HUVEC
Using the HUVEC having L-PGDS genes, we invest- tigated the effects of endogenous PGD 2 on the expression of E-selectin with or without IL-1 stimulation.The E-selectin expression was significantly increased upon IL-1 stimulation in endothelial cells transfected with transporter genes.This increase in E-selectin with or without IL-1 stimulation was significantly blunted in HUVEC transfected with L-PGDS genes that produced indeed endogenous PGD 2 in response to arachidonate stimulation (Figure 3).

Intracellular Effects of PGD 2 on E-Selectin Expression
In order to assess the role of PGD 2 receptor-mediated signal transduction in E-selectin expression, we studied changes in cAMP, a second messenger of PGD 2 signal transduction in HUVEC transfected with transporter gene alone.Intracellular cAMP was unaffected by L-PGDS transfection per se.However, intracellular cAMP was increased by exogenous PGD 2 stimulation [19].This response was completely abrogated by addition of anti-PGD 2 antibody to the media, the amount of which was more than the concentrations sufficient to neutralize PGD 2 in the media (Figure 4(a)).Using such a dose of anti-PGD 2 antibody enough to inhibit the receptor-mediated cyclic AMP rising, we investigated contribution of PGD 2 receptor-mediated signal transduction to the PGD 2 -mediated E-selectin expression, and determined E-selectin expression in the supernatants following neutralization with an anti-PGD 2 -specific antibody.The addition of anti-PGD 2 antibody to the media did not influence E-selectin expression following IL-1 stimulation in the endothelial cells transfected with L-PGDS genes (Figure 4(b)).Effect of endogenous PGD 2 on E-selectin expression in endothelial cells.cAMP levels were measured to assess the antibody inhibition of PGD 2 receptor-mediated signal transduction in the cells with transporter gene alone (Graph a). 10 -5 mol/l PGD 2 greatly increased cAMP formation in the cells in the absence of anti-PGD 2 antibody (anti-PGD 2 (-)).This increase was completely abolished to the basal levels by the addition of anti-PGD 2 antibody in the media.Using such a dose of anti-PGD 2 antibody, we examined the effects of neutralization of PGD 2 in media on E-selectin expression (Graph b).E-selectin expression following IL-1 stimulation was signifycantly reduced in endothelial cells having L-PGDS genes and without anti-PGD 2 antibody (PGDS(+), anti-PGD 2 (-)), compared to endothelial cells with reporter genes alone (PGDS(-), anti-PGD 2 (-)).The reduction in E-selectin expression was unaffected when PGD 2 in the culture was neutralized with an anti-PGD 2 -specific antibody (PGDS(+), anti-PGD 2 (+)).These studies were carried out in the presence of 10 -6 mol/l arachidonate.Statistical differences were assessed by Student's t-test.*P < 0.01.N.S. represents not statistically significant.
These results clearly indicated that anti-PGD 2 -specific antibody inhibited the PGD 2 receptor-mediated signal transduction and that PGD 2 -mediated reduction in E-selectin expression was not due to the PGD 2 receptormediated events.

L-PGDS Genes and Expression of E-Selectin mRNA
We demonstrated that L-PGDS gene transfection onto HUVEC brought about increases in PGD 2 formation and decreases in E-selectin expression using real-time RT-PCR (Figure 5).The expression of E-selectin mRNA following IL-1 stimulation was significantly less in the HUVEC carrying L-PGDS genes.Expression was normalized to β-actin.Values are expressed as fold change compared to untreated controls.

DISCUSSION
Both PGD 2 and adhesion molecules play a very important part in the process of atherosclerosis.In the present study, we demonstrated that PGD 2 regulates E-selectin generation in HUVEC following IL-1stimulation.More interestingly, we demonstrated that endogenous PGD 2 production in HUVEC transfected with L-PGDS genes brought about a decrease of E-selectin expression.This effect was observed even if the PGD 2 -mediated cAMP increase was reversed to basal levels using anti-PGD 2 -specific antibody.These findings strongly suggest that PGD 2 exerts inhibitory effects on E-selectin expression via an intracellular mechanism as well as the well-known receptor-mediated mechanism [20].
In this context, recent studies have demonstrated that the orphan nuclear receptor of peroxysome proliferatoractivated receptor (PPAR)-g, a member of the nuclear receptor superfamily of ligand-dependent transcription factors, binds to PGD 2 metabolites and thereby regulates adipocyte differentiation and glucose homeostasis [21,22].
PGD 2 is converted quickly to PGJ 2 , delta 12-PGJ 2 , and 15-deoxy-delta12,14 PGJ 2 in plasma [23].15-deoxy-delta12,14 PGJ 2 inhibits inhibitor of kB kinase that phosphorylates another inhibitor of kB after activation with cytokines and also affects the DNA-binding domains of nuclear factor-kB (NF-kB) subunits [24].Because the genes involved in E-selectin expression include the NF-kB binding site in its promoter regions [25], it is presumable that PGD 2 and its metabolites inhibit cytokine induced E-selectin expression, at least in part, through inhibition of NF-kB translocation.
It is reported that E-selectin mRNA and E-selectin are much expressed in vascular lesions such as atherosclerosis [5].Different cell lines like macrophages, platelets or lymphocytes work together to secrete cytokines in response to the inflammatory events, and in turn, the secreted cytokines stimulate the endothelial cells, thereby increasing E-selectin mRNA and E-selectin expression in the atherosclerotic lesions.In fact, it is well postulated that L-PGDS is highly expressed in the stenotic lesions and in the lipid core of advanced atherosclerotic plaques in patients with stable angina [9].Moreover, PGD 2 reduces inducible nitric oxide synthase formation.These actions of PGD 2 and metabolites on vasoactive substances regulated by cytokines are in favor of vascular protection against vascular injury [6].
The endothelial cells exhibited a striking increase in PGI 2 generation, but not PGD 2 , following arachidonate stimulation, suggesting that these cells have a large capacity to synthesize PGI 2 , but not PGD 2 .In spite of this, L-PGDS gene transfection greatly enhanced PGD 2 synthesis in EC.This was particularly apparent when eicosanoid expression was stimulated with its precursor, arachidonate.The alteration of phenotype of these cells reduced E-selectin mRNA expression and consequently E-selectin expression.This genetic procedure would provide a new strategy against vascular lesions.These activities have relevance to the in vivo situation, and remain to be clarified.
In conclusion, the introduction of PGD 2 synthase genes into endothelial cells increased endogenous PGD 2 generation.This brought about a reduction in E-selectin expression and a decrease in E-selectin mRNA expression following IL-1 stimulation.The inhibitory effects of PGD 2 on E-selectin expression were due to an intracrine mechanism rather than to any receptor-mediated events.Since suppression of the E-selectin system is postulated to be protective, an increase in endogenous PGD 2 synthesis might represent a novel strategy to prevent cardiovascular injury in humans.

Figure 1 .
Figure 1.Effect of exogenous PGD 2 on E-selectin generation in endothelial cells.Stimulation of endothelial cells with IL-1 significantly increased E-selectin generation (left two columns).The increase was reduced in a dose-dependent manner by the addition of PGD 2 at concentrations ranging from 10 -7 to 10 -4 mol/l.All experiments were performed three different times with at least six replicates.Statistical differences were analyzed by one-way ANOVA and Duncan's multiple range tests.*P < 0.01 vs the value at 0 mol/L PGD 2 .

Figure 2 .
Figure 2. PGD 2 and PGI 2 generation in endothelial cells.PGD 2 was assayed using radioimmunoassay.10 -6 mol/l arachidonate did not stimulate PGD 2 generation in cells having reporter genes alone (two columns to the left in Graph a).However, cells carrying the L-PGDS genes acquired the ability to produce PGD 2 with or without arachidonate stimulation (two columns to the right in Graph a).(Graph a).PGI 2 was assayed as 6-keto-PGF 1a using radioimmunoassay.The addition of 10 -6 mol/L arachidonate to the cultures increased PGI 2 generation; however, there were no differences in PGI 2 production between PGDS(-), AA(+) and PGDS(+), AA(+) lines.(Graph b).Statistical differences were assessed by Student's t-test (n = 6).*P < 0.01.N.S. represents not statistically significant.

Figure 4 .
Figure 4.Effect of endogenous PGD 2 on E-selectin expression in endothelial cells.cAMP levels were measured to assess the antibody inhibition of PGD 2 receptor-mediated signal transduction in the cells with transporter gene alone (Graph a). 10 -5 mol/l PGD 2 greatly increased cAMP formation in the cells in the absence of anti-PGD 2 antibody (anti-PGD 2 (-)).This increase was completely abolished to the basal levels by the addition of anti-PGD 2 antibody in the media.Using such a dose of anti-PGD 2 antibody, we examined the effects of neutralization of PGD 2 in media on E-selectin expression (Graph b).E-selectin expression following IL-1 stimulation was signifycantly reduced in endothelial cells having L-PGDS genes and without anti-PGD 2 antibody (PGDS(+), anti-PGD 2 (-)), compared to endothelial cells with reporter genes alone (PGDS(-), anti-PGD 2 (-)).The reduction in E-selectin expression was unaffected when PGD 2 in the culture was neutralized with an anti-PGD 2 -specific antibody (PGDS(+), anti-PGD 2 (+)).These studies were carried out in the presence of 10 -6 mol/l arachidonate.Statistical differences were assessed by Student's t-test.*P < 0.01.N.S. represents not statistically significant.

Figure 5 .
Figure 5.Effect of PGD 2 synthase gene transfection on E-selectin mRNA.Changes in E-selectin mRNA levels were investigated using real-time RT-PCR.The expression of E-selectin mRNA following IL-1 stimulation was significantly less in the HUVEC carrying L-PGDS genes.Expression was normalized to β-actin.Values are expressed as fold change compared to untreated controls.Statistical differences were assessed by Student's t-test.*P < 0.01.