HDL3 Induces Cyclooxygenase-2 Expression and Prostacyclin Release in Human Endothelial Cells Via a p38 MAPK/CRE-Dependent Pathway: Effects o
http://www.100md.com
《动脉硬化血栓血管生物学》
From the Department of Pharmacological Sciences (G.D.N., E.C., A.L.C.), University of Milan, Italy; the Department of Pharmacology (H.I.), National Cardiovascular Centre Research Institute, Osaka, Japan; and the Centro per lo Studio e la Prevenzione delle Vasculopatie Periferiche (A.L.C.), Ospedale Bassini, Cinisello Balsamo, Italy.
ABSTRACT
Objective— In endothelial cells, cyclooxygenase-1 (COX-1) and COX-2 both contribute to prostacyclin production. Recent findings suggest that COX-2 contributes significantly to systemic prostacyclin synthesis in humans; whether COX-2 inhibition is related to an increased cardiovascular risk is undergoing debate. HDLs have been shown to increase prostacyclin synthesis, thus in the present study we investigated the molecular mechanisms involved in this effect in endothelial cells.
Methods and Results— HDL3 (30 μg/mL) induced COX-2 expression in a time- and dose-dependent manner. COX-2 was found mainly in the perinuclear area where it co-localizes with PGI synthase. Transient transfection experiments showed that CRE is required for HDL-induced COX-2 transcription, and we demonstrated that p38 MAPK activation by HDL3 is involved in COX-2 mRNA transcription and stabilization. As a consequence of COX-2-induction by HDL3 prostacyclin production increased, incubation with a COX-2 selective inhibitor blocked this effect. Moreover, HDL3 increased caveolin-1 phosphorylation, thus promoting PGI-synthase shuttling from the membrane to the perinuclear area.
Conclusion— We conclude that in endothelial cells, HDL modulates COX-2/PGI-S activity via both p38 MAPK-dependent COX-2 mRNA stability and transcription and both caveolin-1–dependent PGI-synthase shuttling and COX-2 coupling. The understanding of these mechanisms may provide new insights into the antiatherogenic role of HDL.
Key Words: HDL ? cyclooxygenase-2 ? p38 MAPK ? prostacyclin ? caveolin-1
Introduction
High-density lipoprotein (HDL) protects from atherosclerotic vascular disease.1 Beyond reverse cholesterol transport, HDL particles posses several anti-atherosclerotic effects,2 including the induction of prostacyclin (PGI2), a strong vasorelaxant3 that acts also as an inhibitor of platelet and leukocyte activation.4 The stimulatory effect on PGI2 depends mainly on the supply by HDL of endothelial cells with arachidonic acid.3 The rate-limiting step in the conversion of the arachidonic acid to eicosanoids is the activity of cyclooxygenase (COX).4 Two major forms of COX, COX-1 and COX-2, have been identified.5 Although COX-1 is constitutively expressed in most cell types, COX-2 is induced by various growth factors and cytokines.6,7 Recent findings suggest that COX-2 contributes significantly for PGI2 synthesis in endothelial cells,8,9 whereas COX-1 is mainly involved in TXA2 synthesis by platelets.8,9 Whether COX-2 inhibition is related to an increase of cardiovascular risk is uncertain.10 HDL induces COX-2 expression in rabbit smooth muscle cells11 and cooperates with TNF-alpha to elicit this effect,12 the molecular mechanisms involved, however, are unclear. COX-2 expression is modulated by growth factors and cytokines via mitogen-activated protein kinase (MAPKs) cascade.13,14 Once activated, the MAPKs may modulate the activity of several transcription factors such as CREB, NFAT, AP-1, and NF-KB,15–17 which are involved in COX-2 expression.18–21
In the present study, we investigated the molecular mechanisms involved in the effect of HDL3 on COX-2 expression and eicosanoid production in cultured human endothelial cells.
Methods
HDL3 (density 1.125 to 1.21 g/mL) was isolated from human plasma and protein content was determined as described.22 HDL3 was used within 6 hours from isolation. No LPS contamination was detected as assessed by the endotoxin kit (Sigma, Italy).
HUVECs were isolated and cultured as described.23 In all experiments, cells were preincubated with serum-free medium for 6 hours,22–24 then HDL3 was added.
The antibodies to phospho-p38 MAPK, phospho-p44/42 MAPK, phospho-IkB-alpha, phospho-CREB, and phospho-caveolin-1 were from New England Biolabs (Germany). COX-1, COX-2, and PGI and PGE synthase (PGIS, mPGES-1) monoclonal antibodies were from Cayman (USA). ?-Actin antibody was from Sigma. Secondary antibodies were from Biorad (Italy). Western blotting analysis was performed as described;23 all antibodies were diluted 1:1000, except ?-actin (1:10000).
The MEK inhibitor, U0126 (New England Biolabs), and the p38 MAPK inhibitor SB203580 (Sigma) were used at a final concentration of 10 μmol/L and 1 μmol/L, respectively. Indomethacin heptyl ester (Cayman), a selective COX-2 inhibitor,25 was used at 0.1 μmol/L.
Immunocytochemistry
Cells were cultured on coverslips in 24-well plates. Fixed cells23 were incubated with a monoclonal antibody for COX-1 or COX-2 (1:50) overnight at 4°C, followed by incubation with anti-mouse IgG FITC-conjugated (1:100, RD, Italy) for 30 minutes, then propidium iodide (2,5 μg/mL) was added for 30 minutes. For the studies of COX-2 co-localization with PGI-S, mPGES-1 and phospho-caveolin-1 fixed cells were incubated overnight with the antibody, followed by incubation with anti IgG FITC-conjugated (30 minutes), anti-COX-2 phycoerythrin-labeled for 1 hour, and TOPRO 3 (Molecular Probes) (1:500) for 15 minutes. The coverslips were analyzed with a confocal microscope (Nikon Eclipse TE 2000-S; Radiance 2100 Biorad) at 600x magnifications. Sixty sections were captured (0.01 μm each) and a three-dimensional reconstruction was obtained using the software Image ProPlace 4.5 (Media Cybernetics, USA).
Real-Time Quantitative Reverse-Transcriptase Polymerase Chain Reaction
Total RNA was extracted and underwent reverse transcription as described.22,24 Three μL of cDNA were amplified by real-time quantitative polymerase chain reaction (PCR) with 1x Syber green universal PCR mastermix (Biorad). The specificity of the Syber green fluorescence was tested by plotting fluorescence as a function of temperature to generate a melting curve of the amplicon. The melting peaks of the amplicons were as expected (not shown). The primers used, the amplicon size, and the melting temperature are indicated in online Table I (available online at http://atvb.ahajournals.org). Each sample was analyzed in duplicate using the IQTM-Cycler (Biorad). The PCR amplification was related to a standard curve ranging from 10–11 M to 10–14 M.
Transcription Assay
The construction of various reporter vectors for the human COX-2 gene has been described previously.20,21 Transfection experiments were first performed using HUVECs and EAhy 926 cells; however, the efficiencies reached were very low, with a high degree of cytotoxicity (data not shown). Because human COX-2 promoter regulation is similar in a wide number of cell types,26–28 we performed transfection experiments in CHO cells, a cell line widely used for studies involving the effects of HDL in vitro.29,30 CHO cells were transiently transfected with COX-2 (nucleotide –327/+59), the NF-kB mutated site (KBM), or the CRE mutated site (CRM) luciferase reporter vectors using lipofectamine (Invitrogen, Italy) according to the manufacturer instructions. Luciferase activity was determined and normalized for the cellular protein concentration.21
Detection of Prostaglandin Release by Competitive Enzyme Immunoassay
Competitive enzyme immunoassay kits for 6-keto PGF1, TXB2, and PGE2 were from Cayman. HUVECs were exposed to HDL3 (30 μg/mL) for 6 hours, washed twice with PBS, and then incubated for 30 minutes with exogenous AA (10 μmol/L); 50 μL for each sample were processed for prostaglandin release according to the manufacturer instructions.
Statistical Analysis
Statistical analysis was performed by ANOVA with the use of Statsoft Statistica Package.
Results
HDL3 Induces COX-2 Expression in HUVECs
COX-2 protein was expressed at low levels in unstimulated cells and was strongly induced 2 hours after exposure to HDL3 (30 μg/mL). In preliminary experiments, this concentration maximally induced COX-2 expression and no further increase was observed up to 600 μg/mL of HDL3. The induction was maximal after 4 hours and begun to decrease after 8 hours (Figure 1a). In unstimulated cells, COX-2 expression remained low at all time points (data not shown). Under the same experimental conditions, HDL3 did not affect COX-1 expression (Figure 1a). These findings were confirmed by immunocytochemistry. COX-2 expression increased after 4 hours in cells incubated with HDL3 without changes of COX-1 expression (Figure 1b). On three-dimensional reconstruction, COX-2 localized in the perinuclear area and in the cytoplasm31 (Figure 1c).
Figure 1. Time dependency of COX-1 and COX-2 expression as detected by Western immunoblotting (a) and indirect immunofluorescence (b) in HUVEC incubated with HDL3 for 4 hours (30 μg/mL). (The green signal represents COX-1 or COX-2, whereas the red one is the nuclear staining with propidium iodide). A three-dimensional reconstruction showing that COX-2 localizes mainly in the perinuclear cytoplasm is shown (c). Three different projections are shown. The results are representative of 4 separate experiments.
Effects of HDL3 on Intracellular Kinase Pathways and on COX-2 Promoter Activity
HDL3 activated ERK1/2 and p38 MAPK, with a the peak of phosphorylation reached after 5 to 10 minutes of incubation (Figure 2). Several transcription factors are activated through MAPK-dependent pathways.17–19 HDL3 activated CREB, with a peak of activity at 10 to 20 minutes (Figure 2), in agreement with the observation that both ERK1/2 and p38 MAPK activate CREB via p90RSK or via MSK-1, respectively. Ik-B alpha phosphorylation results in the release and nuclear translocation of active NF-kB.17 Under our experimental conditions, a basal level of phosphorylation of Ik-B alpha was present, and only a minimal effect on phosphorylation was observed after 5 and 10 minutes of incubation with HDL3 (Figure 2). The human COX-2 promoter region (–327/+59) contains the NF-kB, the NF-IL6, and the CRE sites.20,21 Transient transfection assay showed that HDL3 induced promoter activity by 2.96±0.03-fold, whereas LPS (1 μg/mL), a positive control, induced promoter activity by 4.24±0.02-fold (P<0.01 for both versus control) (Figure 3). The promoter activity of the plasmid carrying the mutation at the NF-kB site was 1.87±0.12 fold in HDL3 incubated cells (P<0,01) and 0.93±0.09 fold in LPS-treated cells, whereas that of the mutant carrying the mutation at the CRE site was 1.15±0.03-fold in HDL3-treated cells and 1.26±0.16-fold in LPS treated cells (Figure 3; P=NS versus control).
Figure 2. Time-dependent phosphorylation of ERK1/2, P38 MAPK, CREB, and Ik-B alpha after incubation of endothelial cells with HDL3. HUVEC were incubated from 5 minutes up to 40 minutes with HDL3 (30 μg/mL). The results are representative of 4 experiments.
Figure 3. Identification of the regions responsible for HDL3-induced promoter activity of the human COX-2 gene. The 5'-flanking region of the human COX-2 gene with site-specific mutations are represented schematically on the left. After transfection, CHO cells were incubated for 6 hours with LPS (1 μg/mL), used as positive control, and with HDL3 (30 μg/mL). The results are presented as relative luciferase activity normalized to cellular protein content. Each experiment was performed in triplicate. *P<0.01 versus control.
Involvement of p38-MAPK in HDL3-Induced COX-2 Protein and mRNA Expression and Stabilization
Cells were preincubated with the MEK1 inhibitor U 0126 (25 μmol/L) or the p38 MAPK inhibitor SB 203580 (1 μmol/L) for 1 hour. HDL3 (30 μg/mL) were added for 2 and 4 hours to evaluate COX-2 mRNA and protein expression. U0126 did not affect HDL3-induced COX-2 expression. SB203580 strongly inhibited HDL3-mediated COX-2 mRNA and protein expression (Figure 4). Because p38 MAPK stabilizes COX-2 mRNA,32 we investigated whether HDL3 possesses this effect. To asses the stability of COX-2 mRNA in HUVEC, actinomycin D (2 μg/mL) was added to cells after 2 hours of HDL3 incubation and COX-2 mRNA levels were measured up to 60 minutes (Figure 4). Simultaneous addition of SB203580 (1 μmol/L) and actinomycin D to the cells after a 2-hour stimulation with HDL3 resulted in a more rapid decrease in COX-2 mRNA levels, suggesting mRNA stabilizing effect by p38 MAPK activity.
Figure 4. Involvement of p38 MAPK pathway in COX-2 mRNA (a and b) and protein expression induced by HDL3 (c and d). HUVECs were incubated for 2 hours (for mRNA detection) and 4 hours (for protein detection) with HDL3 (30 μg/mL) alone or in the presence of U0126 (25 μmol/L) or of SB203580 (1 μmol/L). a, COX-2 mRNA expression was assessed with real-time quantitative PCR; the target sequence was normalized to the beta actin content. b, Inhibition of p38 MAPK destabilizes COX-2 mRNA expression induced by HDL3. HUVECs were incubated for 2 hours with HDL3 (30 μg/mL), then actinomycin D alone or actinomycin D plus SB203580 were added (time 0). COX-2 mRNA levels were measured using real-time quantitative PCR from time 0 minutes up to 60 minutes. Data are mean±SD of 4 separate experiments. *P<0.01. c and d, COX-2 protein expression was assessed by immunoblotting. c, The results are mean±SD of 4 separate experiments and are normalized for beta actin content. *P<0.01.
Effects of HDL3 on Eicosanoid Production
The effects of HDL3 on eicosanoids production were assessed in HUVECs exposed to 30 μg/mL of lipoproteins for 6 hours, followed by 30 minutes of incubation with exogenous AA (10 μmol/L).11 In control cells, the production of 6-keto PGF1 (PGI2 main metabolite) was 73.14±6.79 pg/mg of cellular protein. Incubation of endothelial cells with HDL3 increased 6-keto PGF1 production to 113.38±2.54 pg/mg of cellular protein (P<0,01) (Table II, available online at http://atvb.ahajournals.org). In the presence of 0.1 μmol/L indomethacin eptyl ester, a selective COX-2 inhibitor,25 HDL3-induced 6-keto PGF1 production was reduced to 77.95±10.19 pg/mg of cellular protein and PGE2 resulted in 74.10±3.45 pg/mg of cellular protein and was not affected by HDL3 incubation.
Effects of HDL3 on PGI Synthase Expression and Cellular Localization
As HDL3 induces COX-2 expression and increases PGI2 release, we investigated whether HDL3 can affect PGI-S or mPGES-1 expression. HDL incubation did not change PGI-S or mPGES-1 expression (1.10±0.2-fold and 1.07±0.5-fold versus control cells, respectively) (Figure 5a). Furthermore, in HDL-treated cells, PGI-S co-localized with COX-2 while mPGES-1 showed a different subcellular distribution (Figure 5b).
Figure 5. Effects of HDL3 on PGI-S and PGE-S mRNA expression and cellular localization. a, HUVECs were incubated for 2 hours with HDL3 (30 μg/mL) and COX-2 mRNA levels were measured using quantitative real-time PCR. Data are mean±SD of 4 separate experiments. b, cellular localization of COX-2 and PGI-S or PGE-S. The green signal is for PGI-S or PGE-S, the red signal is for COX-2, and the blue signal is the staining for the nucleus. The results are representative of 4 experiments.
PGI-S resides in caveolae in resting cells.33 Caveolin-1 is the main protein of caveolae, and when phosphorylated34 it moves into the cytoplasm,34,35 shuttling PGI-S in the perinuclear area where it couples to COX-2,35 thus increasing prostacyclin synthesis. We investigated, therefore, whether HDL can influence caveolin-1 phosphorylation and shuttling in the perinuclear space. After 4 hours of incubation, HDL increased caveolin-1 phosphorylation (Figure 6a), mainly in the area surrounding the nucleus (Figure 6b). Moreover, a three-dimensional reconstruction shows that phosphorylated caveolin-1 localizes near COX-2 in the perinuclear area of HDL-treated cells (Figure 6c), where PGI-S is also located (Figure 6d).
Figure 6. Effects of HDL3 on caveolin-1 phosphorylation and cellular localization. a, HUVECs were incubated for 4 hours with HDL3 (30 μg/mL) and phospho-caveolin-1 levels were determined by Western blotting. b, cellular localization of phospho-caveolin-1. The green signal is for phospho-caveolin-1, the blue is staining for the nucleus. c, Cellular localization of phospho-caveolin-1 and COX-2 (transversal projection). The green signal is for phospho-caveolin-1, the red signal is for COX-2, and the blue signal is the nucleus. d, Cellular localization of phospho-caveolin-1 and PGI-S (transversal projection). The green signal is for phospho-caveolin-1, the red signal is for PGI-S, and the blue signal is the nucleus. The results are representative of 4 experiments.
Discussion
The major finding of this study is that HDL3 induces COX-2 expression and PGI2 release in human endothelial cells via p38 MAPK activation. The activation of this signaling pathway promotes COX-2 mRNA transcription and stabilization.
On incubation of cells with HDL3, COX-2 protein localized mainly in the perinuclear area, in agreement with previous findings showing that COX-2 accumulation near the nuclear envelope and in the cytoplasm is required for the increase in COX-2–mediated prostanoid synthesis in vascular endothelial cells.35,36 This effect is specific for COX-2; in fact, COX-1 was mainly localized in the cytoplasm and was not modulated by HDL3; moreover, PGI2 synthesis was downregulated by a specific COX-2 inhibitor.
The molecular mechanisms by which HDL3 induces COX-2 are unknown. Here we show that HDL3 activates 2 of the major kinases pathways involved in COX-2 gene transcription: ERK1/2 and p38 MAPK.13,14 HDL can activate ERK1/2 via cell surface S1P receptor in astroglial cells.37 However, the possibility that MAPK activation results from plasma membrane cholesterol depletion cannot be excluded.38 In support of this hypothesis, Smith et al39 showed that increasing concentration of LDL or free cholesterol decreases COX-2 expression and PGI2 synthesis. As HDL triggers the release of cholesterol from cells,2 our observation suggests that cellular cholesterol balance plays an important role in determining COX-2 levels.
HDL3 also activates CREB in a time-dependent fashion, CREB binds to CRE, which serves as an anchor for P300 interaction with upstream transactivators and downstream transcription machinery,40 thus suggesting that CRE plays a relevant role in COX-2 induction by a number of stimuli.40 Using transient transfection experiments, we demonstrated that mutation in CRE abrogated the luciferase activity induced by HDL3, confirming the role of CRE in HDL3-induced COX-2 gene transcription.
NF-kB has also been suggested to be involved in determining COX-2 gene transcription.20,21 We show that a mutation in the NF-kB response element abrogates luciferase activity induced by LPS, used as a positive control, while it slightly decreases HDL3-induced luciferase activity, suggesting a minor role of this pathway in COX-2 induction by HDL3.
As transcriptional regulation of the COX-2 gene occurs via activation of MAPKs,13,14 we investigated whether inhibition of ERK1/2 or p38 MAPK pathway affected HDL3-induced COX-2 mRNA and protein expression. We show that the p38 MAPK pathway is responsible for the induction of COX-2 by HDL3.
P38 MAPK plays a housekeeping role in maintaining COX-2 mRNA stability32 via the recognition of the AUUUA motifs present in the 3' untranslated region of COX-2.41 We therefore studied COX-2 mRNA stability in cells stimulated with HDL3. Simultaneous addition of actinomycin D and SB203580 to the cells resulted in a more rapid decrease in COX-2 mRNA compared with actinomycin D alone. This represents a new mechanism by which HDL can influence gene expression at a posttranscriptional level and is likely to contribute to the increase of COX-2 protein levels in endothelial cells.
COX-2 has been proposed to exert both an antiatherogenic or a proatherogenic role depending on the eicosanoids produced and the arterial wall cells where it is expressed.9 Eicosanoids are involved in a variety of physiological processes in atherosclerosis and thrombosis, including leukocyte–endothelial cell adhesion, vasorelaxation, and platelet aggregation.9 The dominant prostaglandin produced by endothelial cells is PGI2.4 PGI2 is believed to play a protective role in atherothrombosis.4 COX-2 contributes significantly to systemic PGI2 synthesis in humans;42 therefore, it is possible that COX-2 induced in endothelial cells at lesion-protected areas catalyzes the formation of the anti-atherogenic molecule prostacyclin. This may be the case in the presence of HDL3 that increases PGI2 release mediated by AA in endothelial cells. This effect is dependent mainly on COX-2 as indomethacin eptyl ester, a specific COX-2 inhibitor, abolished PGI2 release induced by HDL3. This observation may also be relevant to the recent observation that COX-2 inhibitors may increase CHD risk.10 In vitro 30 μg/mL of HDL3 induces maximally COX-2 expression, and no further increase is observed up to 600 μg/mL (a physiological concentration that constantly bathes arteries in vivo), thus suggesting that low concentrations of HDL are enough to support COX-2 expression, and higher levels may only provide the substrate. Alternatively, the in vitro conditions allow for a better interaction of HDL with cultured endothelial cells as compared with in vivo settings, in which proteoglycans may trap lipoproteins and reduce their availability for interactions with the endothelial cells.
The observation that COX-2 induced by HDL3 does not increase PGE2, a proatherogenic eicosanoid, synthesized mainly via COX-2,12 confirms that COX-2 expression in the arterial wall could play both a proatherogenic or anti-atherogenic role, but it is the final eicosanoid produced that is responsible for its proatherogenic or anti-atherogenic properties.
Moreover, HDL3-induced COX-2 protein co-localizes with PGI-S in endothelial cells, thus suggesting that in this model, once induced, COX-2 can drive prostacyclin synthesis. PGI-S is associated with caveolae33 and is activated when shuttled from the plasma membrane in the perinuclear area;35 moreover, disruption of caveolae organization downregulates prostacyclin production and impairs angiogenesis.43,44 Here we demonstrate that HDL3 induces caveolin-1 phosphorylation, which shuttles with PGI-S from the plasma membrane to the perinuclear area where it co-localizes with COX-2. Furthermore, the possibility that the abundant increase in COX-2 observed can be related to an increase of prostanoids synthesis other than prostacyclin cannot be excluded.
Also, endothelial nitric oxide synthase, the enzyme responsible for nitric oxide synthesis in the endothelium, localizes in the caveolae.33,35 Nitric oxide is responsible for several beneficial effects of HDL on endothelial cells,45 such as helping to maintain endothelial integrity, facilitating vascular relaxation, inhibiting cell adhesion to vascular endothelium, decreasing radical oxygen production, and inhibiting apoptosis.45 Even if we have not addressed the role of HDL in modulating endothelial nitric oxide synthase shuttling through caveolin-1 phosphorylation, it is conceivable that some of the effects of HDL are mediated via this pathway46
In summary, our data suggest that in human endothelial cells, HDL can modulate COX-2 expression via p38 MAPK-dependent COX-2 mRNA transcription and stabilization. Moreover, the HDL-dependent caveolin-1 phosphorylation favors PGI-S shuttling and COX-2 coupling. These data add further insights into the molecular mechanisms involved in the anti-atherogenic activity of HDL.
Acknowledgments
We are grateful to Roberto Zecca for software assistance, to Giulio Simonutti for technical assistance with confocal microscopy, to Fabio Pellegatta, who provided the HUVECs, and to Francesco Cipollone, who provided the anti-mPGES antibody.
References
Assmann G, Schulte H, von Eckardstein A, Huang Y. High-density lipoprotein cholesterol as a predictor of coronary heart disease risk. The PROCAM experience and pathophysiological implications for reverse cholesterol transport. Atherosclerosis. 1996; 124: S11–S20.
Nofer JR, Kehrel B, Fobker M, Levkau B, Assmann G, von Eckardstein A. HDL and arteriosclerosis: beyond reverse cholesterol transport. Atherosclerosis. 2002; 161: 1–16.
Fleisher LN, Tall AR, Witte LD, Miller RW, Cannon PJ. Stimulation of arterial endothelial cell prostacyclin synthesis by high density lipoproteins. J Biol Chem. 1982; 257: 6653–6655.
Thiemermann C. Biosynthesis and interaction of endothelium-derived vasoactive mediators. Eicosanoids. 1991; 4: 187–202.
Vane JR, Bakhle Y, Botting R. Cyclooxygenase 1 and 2. Ann Rev Pharmacol Toxicol. 1998; 38: 97–120.
Bartlett SR, Sawdy R, Mann GE. Induction of cyclooxygenase-2 expression in human myometrial smooth muscle cells by interleukin-1 beta: involvement of p38 mitogen-activated protein kinase. J Physiol. 1999; 520: 399–406.
Lee SH, Soyoola E, Chanmugam P, Hart S, Sun W, Zhong H, Liou S, Simmons D, Hwang D. Selective expression of mitogen-inducible cyclooxygenase in macrophages stimulated with lipopolysaccharide. J Biol Chem. 1992; 267: 25934–25938.
Caughey GE, Cleland LG, Penglis PS, Gamble JR, James MJ. Roles of cyclooxygenase (COX)-1 and COX-2 in prostanoid production by human endothelial cells: selective up-regulation of prostacyclin synthesis by COX-2. J Immunol. 2001; 167: 2831–2838.
Linton MF, Fazio S. Cyclooxygenase-2 and atherosclerosis. Curr Opin Lipidol. 2002; 13: 497–504.
Ray WA, Stein CM, Daugherty JR Hall K, Arbogast PG, Griffin MR. COX-2 selective non-steroidal anti-inflammatory drugs and risk of serious coronary heart disease. Lancet. 2002; 360: 1071–1073.
Vinals M, Martinez-Gonzalez J, Badimon L. Regulatory effects of HDL on smooth muscle cell prostacyclin release. Arterioscler Thromb Vasc Biol. 1999; 19: 2405–2411.
Cockerill GW, Saklatvala J, Ridley SH, Yarwood H, Miller NE, Oral B, Nithyanathan S, Taylor G, Haskard DO. High-density lipoproteins differentially modulate cytokine-induced expression of E-selectin and cyclooxygenase-2. Arterioscler Thromb Vasc Biol. 1999; 19: 910–917.
Guan Z, Buckman SY, Miller BW, Springer LD, Morrison AR. Interleukin-1?-induced cyclooxygenase-2 expression requires activation of both c-Jun NH2-terminal kinase and p38 MAPK signal pathway in rat mesangial cells. J Biol Chem. 1998; 273: 28670–28676.
LaPointe MC, Isenovic E. Interleukin-1? regulation of inducible nitric oxide synthase and cyclooxygenase-2 involves the p44/44 and p38 MAPK signalling pathways in cardiac myocytes. Hypertension. 1999; 33: 276–282.
Seger R, Krebs EG. The MAPK signalling cascade. FASEB J. 1995; 9: 726–735.
Hazzalin CA, Mahadevan LC. MAPK-regulated transcription: a continuously variable gene switch? Nat Rev Mol Cell Biol. 2002; 3: 30–40.
Schulze-Osthoff K, Ferrari D, Riehemann K, Wesselborg S. Regulation of NF-kappa B activation by MAP kinase cascades. Immunobiology. 1997; 198: 35–49.
Tan Y, Rouse J, Zhang A, Cariati S, Cohen P, Comb MJ. FGF and stress regulate CREB and ATF-1 via a pathway involving p38 MAP kinase and MAPKAP kinase-2. EMBO J. 1996; 15: 4629–4642.
Xie W, Herschman HR. v-src induces prostaglandin synthase 2 gene expression by activation of the c-Jun N-terminal kinase and the c-Jun transcription factor. J Biol Chem. 1995; 270: 27622–27628.
Inoue H, Yokoyama C, Hara S, Tone Y, Tanabe T. Transcriptional regulation of human prostaglandin-endoperoxide synthase-2 gene by lipopolysaccharide and phorbol ester in vascular endothelial cells. Involvement of both nuclear factor for interleukin-6 expression site and cAMP response element. J Biol Chem. 1995; 270: 24965–24971.
Inoue H, Umesono K, Nishimori T, Hirata Y, Tanabe T. Glucocorticoid-mediated suppression of the promoter activity on the cyclooxygenase-2 gene is modulated by expression of its receptor in vascular endothelial cells. Biochem Biophys Res Comm. 1999; 254: 292–298.
Norata GD, Pellegatta F, Hamsten A, Catapano AL, Eriksson P. Effects of high density lipoprotein subfraction 3 on the expression of matrix-degrading proteases in human endothelial cells. Int J Mol Med. 2003; 12: 73–78.
Norata GD, Pirillo A, Callegari E, Hamsten A, Catapano AL, Eriksson P. Gene expression and intracellular pathways involved in endothelial dysfunction induced by VLDL and oxidised VLDL. Cardiovasc Res. 2003; 59: 169–180.
Norata GD, Bjork H, Hamsten A, Catapano AL, Eriksson P. High density lipoprotein decrease ADAMTs1 expression induced by LPS and TNF in human endothelial cells. Matrix Biology. 2004; 22: 557–560.
Kalgutkar AS, Marnett AB, Crews BC, Remmel RP, Marnett LJ. Ester and amide derivatives of the non-steroidal anti-inflammatory drug, indomethacin, as selective cyclooxygenase-2 inhibitors. J Med Chem. 2000; 43: 2860–2870.
Singer CA, Baker KJ, McCaffrey A, AuCoin DP, Dechert MA, Gerthoffer WT. P38 MAPK and NF-{kappa}B Mediate COX-2 Expression in Human Airway Myocytes. Am J Physiol. 2003; 285: L1087–L1098.
Loudon JA, Elliott CL, Hills F, Bennett PR. Progesterone represses interleukin-8 and cyclo-oxygenase-2 in human lower segment fibroblast cells and amnion epithelial cells. Biol Reprod. 2003; 69: 331–337.
Tamura M, Sebastian S, Yang S, Gurates B, Fang Z, Okamura K, Bulun SE. Induction of cyclooxygenase-2 in human endometrial stromal cells by malignant endometrial epithelial cells: evidence for the involvement of extracellularly regulated kinases and CCAAT/enhancer binding proteins. J Mol Endocrinol. 2003; 31: 95–104.
Baez JM, Barbour SE, Cohen DE. Phosphatidylcholine transfer protein promotes apolipoprotein A-I-mediated lipid efflux in Chinese hamster ovary cells. J Biol Chem. 2002; 277: 6198–6206.
Ioka RX, Kang MJ, Kamiyama S, Kim DH, Magoori K, Kamataki A, Ito Y, Takei YA, Sasaki M, Suzuki T, Sasano H, Takahashi S, Sakai J, Fujino T, Yamamoto TT. Expression cloning and characterization of a novel glycosylphosphatidylinositol-anchored high density lipoprotein-binding protein, GPI-HBP1. J Biol Chem. 2003; 278: 7344–7349.
Parfenova H, Parfenov VN, Shlopov BV, Levine V, Falkos S, Pourcyrous M, Leffler CW. Dynamics of nuclear localization sites for COX-2 in vascular endothelial cells. Am J Physiol. 2001; 281: C166–C178.
Dean JLE, Brook M, Clark AR, Saklatvala J. P38 MAPK regulates cyclooxygenase-2 mRNA stability and transcription in lipopolysaccharide-treated human monocytes. J Biol Chem. 1999; 274: 264–269.
Frank PG, Woodman SE, Park DS, Lisanti MP. Caveolin, caveolae, and endothelial cell function. Arterioscler Thromb Vasc Biol. 2003; 23: 1161–1168.
Nomura R, Fujimoto T. Tyrosine-phosphorylated caveolin-1: immunolocalization and molecular characterization. Mol Biol Cell. 1999; 10: 975–986.
Spisni E, Bianco MC, Griffoni C, Toni M, D’Angelo R, Santi S, Riccio M, Tomasi V. Mechanosensing role of caveolae and caveolar constituents in human endothelial cells. J Cell Physiol. 2003; 197: 198–204.
Lim H, Dey SK. A Novel Pathway of Prostacyclin Signalling–Hanging Out with Nuclear Receptors. Endocrinology. 2002; 143: 3207–3210.
Kimura T, Sato K, Malchinkhuu E, Tomura H, Tamama K, Kuwabara A, Murakami M, Okajima F. High-density lipoprotein stimulates endothelial cell migration and survival through sphingosine 1-phosphate and its receptors. Arterioscler Thromb Vasc Biol. 2003; 23: 1283–1288.
Chen X, Resh MD. Activation of mitogen-activated protein kinase by membrane-targeted Raf chimeras is independent of raft localization. J Biol Chem. 2001; 34617–34623.
Smith LH, Boutaud O, Breyer M Morrow JD, Oates JA, Vaughan DE. Cyclooxygenase-2-Dependent Prostacyclin Formation Is Regulated by Low Density Lipoprotein Cholesterol In Vitro. Arterioscler Thromb Vasc Biol. 2002; 22: 983–988.
Schroer K, Zhu Y, Saunders MA, Deng WG, Xu XM, Meyer-Kirchrath J, Wu KK. Obligatory role of cyclic adenosine monophosphate response element in cyclooxygenase-2 promoter induction and feedback regulation by inflammatory mediators. Circulation. 2002; 105: 2760–2765.
Inoue H, Taba Y, Miwa Y, Yokota C, Miyagi M, Sasaguri T. Transcriptional and posttranscriptional regulation of cyclooxygenase-2 expression by fluid shear stress in vascular endothelial cells. Arterioscler Thromb Vasc Biol. 2002; 22: 1415–1420.
McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX)-2: The human pharmacology of selective inhibitors of COX-2. Proc Natl Acad Sci U S A. 1999; 96: 272–277.
Griffoni C, Spisni E, Santi S, Riccio M, Guarnieri T, Tomasi V. Knockdown of caveolin-1 by antisense oligonucleotides impairs angiogenesis in vitro and in vivo. Biochem Biophys Res Commun. 2000; 276: 756–761.
Spisni E, Griffoni C, Santi S, Riccio M, Marulli R, Bartolini G, Toni M, Ullrich V, Tomasi V. Colocalization of PGI-S and caveolin-1 in endothelial cells and new role for prostacyclin in angiogenesis. Exp Cell Res. 2001; 266: 31–43.
Calabresi L, Gomaraschi M, Franceschini G. Endothelial protection by HDL. From bench to beside. Arterioscler Thromb Vasc Biol. 2003; 23: 1724–1731.
Uittenbogaard A, Shaul PW, Yuhanna IS, Blair A, Smart EJ. HDL prevents Ox-LDL-induced inhibition of eNOS localization and activation in caveolae. J Biol Chem. 2000; 275: 11278–11283.(G.D. Norata; E. Callegari)
ABSTRACT
Objective— In endothelial cells, cyclooxygenase-1 (COX-1) and COX-2 both contribute to prostacyclin production. Recent findings suggest that COX-2 contributes significantly to systemic prostacyclin synthesis in humans; whether COX-2 inhibition is related to an increased cardiovascular risk is undergoing debate. HDLs have been shown to increase prostacyclin synthesis, thus in the present study we investigated the molecular mechanisms involved in this effect in endothelial cells.
Methods and Results— HDL3 (30 μg/mL) induced COX-2 expression in a time- and dose-dependent manner. COX-2 was found mainly in the perinuclear area where it co-localizes with PGI synthase. Transient transfection experiments showed that CRE is required for HDL-induced COX-2 transcription, and we demonstrated that p38 MAPK activation by HDL3 is involved in COX-2 mRNA transcription and stabilization. As a consequence of COX-2-induction by HDL3 prostacyclin production increased, incubation with a COX-2 selective inhibitor blocked this effect. Moreover, HDL3 increased caveolin-1 phosphorylation, thus promoting PGI-synthase shuttling from the membrane to the perinuclear area.
Conclusion— We conclude that in endothelial cells, HDL modulates COX-2/PGI-S activity via both p38 MAPK-dependent COX-2 mRNA stability and transcription and both caveolin-1–dependent PGI-synthase shuttling and COX-2 coupling. The understanding of these mechanisms may provide new insights into the antiatherogenic role of HDL.
Key Words: HDL ? cyclooxygenase-2 ? p38 MAPK ? prostacyclin ? caveolin-1
Introduction
High-density lipoprotein (HDL) protects from atherosclerotic vascular disease.1 Beyond reverse cholesterol transport, HDL particles posses several anti-atherosclerotic effects,2 including the induction of prostacyclin (PGI2), a strong vasorelaxant3 that acts also as an inhibitor of platelet and leukocyte activation.4 The stimulatory effect on PGI2 depends mainly on the supply by HDL of endothelial cells with arachidonic acid.3 The rate-limiting step in the conversion of the arachidonic acid to eicosanoids is the activity of cyclooxygenase (COX).4 Two major forms of COX, COX-1 and COX-2, have been identified.5 Although COX-1 is constitutively expressed in most cell types, COX-2 is induced by various growth factors and cytokines.6,7 Recent findings suggest that COX-2 contributes significantly for PGI2 synthesis in endothelial cells,8,9 whereas COX-1 is mainly involved in TXA2 synthesis by platelets.8,9 Whether COX-2 inhibition is related to an increase of cardiovascular risk is uncertain.10 HDL induces COX-2 expression in rabbit smooth muscle cells11 and cooperates with TNF-alpha to elicit this effect,12 the molecular mechanisms involved, however, are unclear. COX-2 expression is modulated by growth factors and cytokines via mitogen-activated protein kinase (MAPKs) cascade.13,14 Once activated, the MAPKs may modulate the activity of several transcription factors such as CREB, NFAT, AP-1, and NF-KB,15–17 which are involved in COX-2 expression.18–21
In the present study, we investigated the molecular mechanisms involved in the effect of HDL3 on COX-2 expression and eicosanoid production in cultured human endothelial cells.
Methods
HDL3 (density 1.125 to 1.21 g/mL) was isolated from human plasma and protein content was determined as described.22 HDL3 was used within 6 hours from isolation. No LPS contamination was detected as assessed by the endotoxin kit (Sigma, Italy).
HUVECs were isolated and cultured as described.23 In all experiments, cells were preincubated with serum-free medium for 6 hours,22–24 then HDL3 was added.
The antibodies to phospho-p38 MAPK, phospho-p44/42 MAPK, phospho-IkB-alpha, phospho-CREB, and phospho-caveolin-1 were from New England Biolabs (Germany). COX-1, COX-2, and PGI and PGE synthase (PGIS, mPGES-1) monoclonal antibodies were from Cayman (USA). ?-Actin antibody was from Sigma. Secondary antibodies were from Biorad (Italy). Western blotting analysis was performed as described;23 all antibodies were diluted 1:1000, except ?-actin (1:10000).
The MEK inhibitor, U0126 (New England Biolabs), and the p38 MAPK inhibitor SB203580 (Sigma) were used at a final concentration of 10 μmol/L and 1 μmol/L, respectively. Indomethacin heptyl ester (Cayman), a selective COX-2 inhibitor,25 was used at 0.1 μmol/L.
Immunocytochemistry
Cells were cultured on coverslips in 24-well plates. Fixed cells23 were incubated with a monoclonal antibody for COX-1 or COX-2 (1:50) overnight at 4°C, followed by incubation with anti-mouse IgG FITC-conjugated (1:100, RD, Italy) for 30 minutes, then propidium iodide (2,5 μg/mL) was added for 30 minutes. For the studies of COX-2 co-localization with PGI-S, mPGES-1 and phospho-caveolin-1 fixed cells were incubated overnight with the antibody, followed by incubation with anti IgG FITC-conjugated (30 minutes), anti-COX-2 phycoerythrin-labeled for 1 hour, and TOPRO 3 (Molecular Probes) (1:500) for 15 minutes. The coverslips were analyzed with a confocal microscope (Nikon Eclipse TE 2000-S; Radiance 2100 Biorad) at 600x magnifications. Sixty sections were captured (0.01 μm each) and a three-dimensional reconstruction was obtained using the software Image ProPlace 4.5 (Media Cybernetics, USA).
Real-Time Quantitative Reverse-Transcriptase Polymerase Chain Reaction
Total RNA was extracted and underwent reverse transcription as described.22,24 Three μL of cDNA were amplified by real-time quantitative polymerase chain reaction (PCR) with 1x Syber green universal PCR mastermix (Biorad). The specificity of the Syber green fluorescence was tested by plotting fluorescence as a function of temperature to generate a melting curve of the amplicon. The melting peaks of the amplicons were as expected (not shown). The primers used, the amplicon size, and the melting temperature are indicated in online Table I (available online at http://atvb.ahajournals.org). Each sample was analyzed in duplicate using the IQTM-Cycler (Biorad). The PCR amplification was related to a standard curve ranging from 10–11 M to 10–14 M.
Transcription Assay
The construction of various reporter vectors for the human COX-2 gene has been described previously.20,21 Transfection experiments were first performed using HUVECs and EAhy 926 cells; however, the efficiencies reached were very low, with a high degree of cytotoxicity (data not shown). Because human COX-2 promoter regulation is similar in a wide number of cell types,26–28 we performed transfection experiments in CHO cells, a cell line widely used for studies involving the effects of HDL in vitro.29,30 CHO cells were transiently transfected with COX-2 (nucleotide –327/+59), the NF-kB mutated site (KBM), or the CRE mutated site (CRM) luciferase reporter vectors using lipofectamine (Invitrogen, Italy) according to the manufacturer instructions. Luciferase activity was determined and normalized for the cellular protein concentration.21
Detection of Prostaglandin Release by Competitive Enzyme Immunoassay
Competitive enzyme immunoassay kits for 6-keto PGF1, TXB2, and PGE2 were from Cayman. HUVECs were exposed to HDL3 (30 μg/mL) for 6 hours, washed twice with PBS, and then incubated for 30 minutes with exogenous AA (10 μmol/L); 50 μL for each sample were processed for prostaglandin release according to the manufacturer instructions.
Statistical Analysis
Statistical analysis was performed by ANOVA with the use of Statsoft Statistica Package.
Results
HDL3 Induces COX-2 Expression in HUVECs
COX-2 protein was expressed at low levels in unstimulated cells and was strongly induced 2 hours after exposure to HDL3 (30 μg/mL). In preliminary experiments, this concentration maximally induced COX-2 expression and no further increase was observed up to 600 μg/mL of HDL3. The induction was maximal after 4 hours and begun to decrease after 8 hours (Figure 1a). In unstimulated cells, COX-2 expression remained low at all time points (data not shown). Under the same experimental conditions, HDL3 did not affect COX-1 expression (Figure 1a). These findings were confirmed by immunocytochemistry. COX-2 expression increased after 4 hours in cells incubated with HDL3 without changes of COX-1 expression (Figure 1b). On three-dimensional reconstruction, COX-2 localized in the perinuclear area and in the cytoplasm31 (Figure 1c).
Figure 1. Time dependency of COX-1 and COX-2 expression as detected by Western immunoblotting (a) and indirect immunofluorescence (b) in HUVEC incubated with HDL3 for 4 hours (30 μg/mL). (The green signal represents COX-1 or COX-2, whereas the red one is the nuclear staining with propidium iodide). A three-dimensional reconstruction showing that COX-2 localizes mainly in the perinuclear cytoplasm is shown (c). Three different projections are shown. The results are representative of 4 separate experiments.
Effects of HDL3 on Intracellular Kinase Pathways and on COX-2 Promoter Activity
HDL3 activated ERK1/2 and p38 MAPK, with a the peak of phosphorylation reached after 5 to 10 minutes of incubation (Figure 2). Several transcription factors are activated through MAPK-dependent pathways.17–19 HDL3 activated CREB, with a peak of activity at 10 to 20 minutes (Figure 2), in agreement with the observation that both ERK1/2 and p38 MAPK activate CREB via p90RSK or via MSK-1, respectively. Ik-B alpha phosphorylation results in the release and nuclear translocation of active NF-kB.17 Under our experimental conditions, a basal level of phosphorylation of Ik-B alpha was present, and only a minimal effect on phosphorylation was observed after 5 and 10 minutes of incubation with HDL3 (Figure 2). The human COX-2 promoter region (–327/+59) contains the NF-kB, the NF-IL6, and the CRE sites.20,21 Transient transfection assay showed that HDL3 induced promoter activity by 2.96±0.03-fold, whereas LPS (1 μg/mL), a positive control, induced promoter activity by 4.24±0.02-fold (P<0.01 for both versus control) (Figure 3). The promoter activity of the plasmid carrying the mutation at the NF-kB site was 1.87±0.12 fold in HDL3 incubated cells (P<0,01) and 0.93±0.09 fold in LPS-treated cells, whereas that of the mutant carrying the mutation at the CRE site was 1.15±0.03-fold in HDL3-treated cells and 1.26±0.16-fold in LPS treated cells (Figure 3; P=NS versus control).
Figure 2. Time-dependent phosphorylation of ERK1/2, P38 MAPK, CREB, and Ik-B alpha after incubation of endothelial cells with HDL3. HUVEC were incubated from 5 minutes up to 40 minutes with HDL3 (30 μg/mL). The results are representative of 4 experiments.
Figure 3. Identification of the regions responsible for HDL3-induced promoter activity of the human COX-2 gene. The 5'-flanking region of the human COX-2 gene with site-specific mutations are represented schematically on the left. After transfection, CHO cells were incubated for 6 hours with LPS (1 μg/mL), used as positive control, and with HDL3 (30 μg/mL). The results are presented as relative luciferase activity normalized to cellular protein content. Each experiment was performed in triplicate. *P<0.01 versus control.
Involvement of p38-MAPK in HDL3-Induced COX-2 Protein and mRNA Expression and Stabilization
Cells were preincubated with the MEK1 inhibitor U 0126 (25 μmol/L) or the p38 MAPK inhibitor SB 203580 (1 μmol/L) for 1 hour. HDL3 (30 μg/mL) were added for 2 and 4 hours to evaluate COX-2 mRNA and protein expression. U0126 did not affect HDL3-induced COX-2 expression. SB203580 strongly inhibited HDL3-mediated COX-2 mRNA and protein expression (Figure 4). Because p38 MAPK stabilizes COX-2 mRNA,32 we investigated whether HDL3 possesses this effect. To asses the stability of COX-2 mRNA in HUVEC, actinomycin D (2 μg/mL) was added to cells after 2 hours of HDL3 incubation and COX-2 mRNA levels were measured up to 60 minutes (Figure 4). Simultaneous addition of SB203580 (1 μmol/L) and actinomycin D to the cells after a 2-hour stimulation with HDL3 resulted in a more rapid decrease in COX-2 mRNA levels, suggesting mRNA stabilizing effect by p38 MAPK activity.
Figure 4. Involvement of p38 MAPK pathway in COX-2 mRNA (a and b) and protein expression induced by HDL3 (c and d). HUVECs were incubated for 2 hours (for mRNA detection) and 4 hours (for protein detection) with HDL3 (30 μg/mL) alone or in the presence of U0126 (25 μmol/L) or of SB203580 (1 μmol/L). a, COX-2 mRNA expression was assessed with real-time quantitative PCR; the target sequence was normalized to the beta actin content. b, Inhibition of p38 MAPK destabilizes COX-2 mRNA expression induced by HDL3. HUVECs were incubated for 2 hours with HDL3 (30 μg/mL), then actinomycin D alone or actinomycin D plus SB203580 were added (time 0). COX-2 mRNA levels were measured using real-time quantitative PCR from time 0 minutes up to 60 minutes. Data are mean±SD of 4 separate experiments. *P<0.01. c and d, COX-2 protein expression was assessed by immunoblotting. c, The results are mean±SD of 4 separate experiments and are normalized for beta actin content. *P<0.01.
Effects of HDL3 on Eicosanoid Production
The effects of HDL3 on eicosanoids production were assessed in HUVECs exposed to 30 μg/mL of lipoproteins for 6 hours, followed by 30 minutes of incubation with exogenous AA (10 μmol/L).11 In control cells, the production of 6-keto PGF1 (PGI2 main metabolite) was 73.14±6.79 pg/mg of cellular protein. Incubation of endothelial cells with HDL3 increased 6-keto PGF1 production to 113.38±2.54 pg/mg of cellular protein (P<0,01) (Table II, available online at http://atvb.ahajournals.org). In the presence of 0.1 μmol/L indomethacin eptyl ester, a selective COX-2 inhibitor,25 HDL3-induced 6-keto PGF1 production was reduced to 77.95±10.19 pg/mg of cellular protein and PGE2 resulted in 74.10±3.45 pg/mg of cellular protein and was not affected by HDL3 incubation.
Effects of HDL3 on PGI Synthase Expression and Cellular Localization
As HDL3 induces COX-2 expression and increases PGI2 release, we investigated whether HDL3 can affect PGI-S or mPGES-1 expression. HDL incubation did not change PGI-S or mPGES-1 expression (1.10±0.2-fold and 1.07±0.5-fold versus control cells, respectively) (Figure 5a). Furthermore, in HDL-treated cells, PGI-S co-localized with COX-2 while mPGES-1 showed a different subcellular distribution (Figure 5b).
Figure 5. Effects of HDL3 on PGI-S and PGE-S mRNA expression and cellular localization. a, HUVECs were incubated for 2 hours with HDL3 (30 μg/mL) and COX-2 mRNA levels were measured using quantitative real-time PCR. Data are mean±SD of 4 separate experiments. b, cellular localization of COX-2 and PGI-S or PGE-S. The green signal is for PGI-S or PGE-S, the red signal is for COX-2, and the blue signal is the staining for the nucleus. The results are representative of 4 experiments.
PGI-S resides in caveolae in resting cells.33 Caveolin-1 is the main protein of caveolae, and when phosphorylated34 it moves into the cytoplasm,34,35 shuttling PGI-S in the perinuclear area where it couples to COX-2,35 thus increasing prostacyclin synthesis. We investigated, therefore, whether HDL can influence caveolin-1 phosphorylation and shuttling in the perinuclear space. After 4 hours of incubation, HDL increased caveolin-1 phosphorylation (Figure 6a), mainly in the area surrounding the nucleus (Figure 6b). Moreover, a three-dimensional reconstruction shows that phosphorylated caveolin-1 localizes near COX-2 in the perinuclear area of HDL-treated cells (Figure 6c), where PGI-S is also located (Figure 6d).
Figure 6. Effects of HDL3 on caveolin-1 phosphorylation and cellular localization. a, HUVECs were incubated for 4 hours with HDL3 (30 μg/mL) and phospho-caveolin-1 levels were determined by Western blotting. b, cellular localization of phospho-caveolin-1. The green signal is for phospho-caveolin-1, the blue is staining for the nucleus. c, Cellular localization of phospho-caveolin-1 and COX-2 (transversal projection). The green signal is for phospho-caveolin-1, the red signal is for COX-2, and the blue signal is the nucleus. d, Cellular localization of phospho-caveolin-1 and PGI-S (transversal projection). The green signal is for phospho-caveolin-1, the red signal is for PGI-S, and the blue signal is the nucleus. The results are representative of 4 experiments.
Discussion
The major finding of this study is that HDL3 induces COX-2 expression and PGI2 release in human endothelial cells via p38 MAPK activation. The activation of this signaling pathway promotes COX-2 mRNA transcription and stabilization.
On incubation of cells with HDL3, COX-2 protein localized mainly in the perinuclear area, in agreement with previous findings showing that COX-2 accumulation near the nuclear envelope and in the cytoplasm is required for the increase in COX-2–mediated prostanoid synthesis in vascular endothelial cells.35,36 This effect is specific for COX-2; in fact, COX-1 was mainly localized in the cytoplasm and was not modulated by HDL3; moreover, PGI2 synthesis was downregulated by a specific COX-2 inhibitor.
The molecular mechanisms by which HDL3 induces COX-2 are unknown. Here we show that HDL3 activates 2 of the major kinases pathways involved in COX-2 gene transcription: ERK1/2 and p38 MAPK.13,14 HDL can activate ERK1/2 via cell surface S1P receptor in astroglial cells.37 However, the possibility that MAPK activation results from plasma membrane cholesterol depletion cannot be excluded.38 In support of this hypothesis, Smith et al39 showed that increasing concentration of LDL or free cholesterol decreases COX-2 expression and PGI2 synthesis. As HDL triggers the release of cholesterol from cells,2 our observation suggests that cellular cholesterol balance plays an important role in determining COX-2 levels.
HDL3 also activates CREB in a time-dependent fashion, CREB binds to CRE, which serves as an anchor for P300 interaction with upstream transactivators and downstream transcription machinery,40 thus suggesting that CRE plays a relevant role in COX-2 induction by a number of stimuli.40 Using transient transfection experiments, we demonstrated that mutation in CRE abrogated the luciferase activity induced by HDL3, confirming the role of CRE in HDL3-induced COX-2 gene transcription.
NF-kB has also been suggested to be involved in determining COX-2 gene transcription.20,21 We show that a mutation in the NF-kB response element abrogates luciferase activity induced by LPS, used as a positive control, while it slightly decreases HDL3-induced luciferase activity, suggesting a minor role of this pathway in COX-2 induction by HDL3.
As transcriptional regulation of the COX-2 gene occurs via activation of MAPKs,13,14 we investigated whether inhibition of ERK1/2 or p38 MAPK pathway affected HDL3-induced COX-2 mRNA and protein expression. We show that the p38 MAPK pathway is responsible for the induction of COX-2 by HDL3.
P38 MAPK plays a housekeeping role in maintaining COX-2 mRNA stability32 via the recognition of the AUUUA motifs present in the 3' untranslated region of COX-2.41 We therefore studied COX-2 mRNA stability in cells stimulated with HDL3. Simultaneous addition of actinomycin D and SB203580 to the cells resulted in a more rapid decrease in COX-2 mRNA compared with actinomycin D alone. This represents a new mechanism by which HDL can influence gene expression at a posttranscriptional level and is likely to contribute to the increase of COX-2 protein levels in endothelial cells.
COX-2 has been proposed to exert both an antiatherogenic or a proatherogenic role depending on the eicosanoids produced and the arterial wall cells where it is expressed.9 Eicosanoids are involved in a variety of physiological processes in atherosclerosis and thrombosis, including leukocyte–endothelial cell adhesion, vasorelaxation, and platelet aggregation.9 The dominant prostaglandin produced by endothelial cells is PGI2.4 PGI2 is believed to play a protective role in atherothrombosis.4 COX-2 contributes significantly to systemic PGI2 synthesis in humans;42 therefore, it is possible that COX-2 induced in endothelial cells at lesion-protected areas catalyzes the formation of the anti-atherogenic molecule prostacyclin. This may be the case in the presence of HDL3 that increases PGI2 release mediated by AA in endothelial cells. This effect is dependent mainly on COX-2 as indomethacin eptyl ester, a specific COX-2 inhibitor, abolished PGI2 release induced by HDL3. This observation may also be relevant to the recent observation that COX-2 inhibitors may increase CHD risk.10 In vitro 30 μg/mL of HDL3 induces maximally COX-2 expression, and no further increase is observed up to 600 μg/mL (a physiological concentration that constantly bathes arteries in vivo), thus suggesting that low concentrations of HDL are enough to support COX-2 expression, and higher levels may only provide the substrate. Alternatively, the in vitro conditions allow for a better interaction of HDL with cultured endothelial cells as compared with in vivo settings, in which proteoglycans may trap lipoproteins and reduce their availability for interactions with the endothelial cells.
The observation that COX-2 induced by HDL3 does not increase PGE2, a proatherogenic eicosanoid, synthesized mainly via COX-2,12 confirms that COX-2 expression in the arterial wall could play both a proatherogenic or anti-atherogenic role, but it is the final eicosanoid produced that is responsible for its proatherogenic or anti-atherogenic properties.
Moreover, HDL3-induced COX-2 protein co-localizes with PGI-S in endothelial cells, thus suggesting that in this model, once induced, COX-2 can drive prostacyclin synthesis. PGI-S is associated with caveolae33 and is activated when shuttled from the plasma membrane in the perinuclear area;35 moreover, disruption of caveolae organization downregulates prostacyclin production and impairs angiogenesis.43,44 Here we demonstrate that HDL3 induces caveolin-1 phosphorylation, which shuttles with PGI-S from the plasma membrane to the perinuclear area where it co-localizes with COX-2. Furthermore, the possibility that the abundant increase in COX-2 observed can be related to an increase of prostanoids synthesis other than prostacyclin cannot be excluded.
Also, endothelial nitric oxide synthase, the enzyme responsible for nitric oxide synthesis in the endothelium, localizes in the caveolae.33,35 Nitric oxide is responsible for several beneficial effects of HDL on endothelial cells,45 such as helping to maintain endothelial integrity, facilitating vascular relaxation, inhibiting cell adhesion to vascular endothelium, decreasing radical oxygen production, and inhibiting apoptosis.45 Even if we have not addressed the role of HDL in modulating endothelial nitric oxide synthase shuttling through caveolin-1 phosphorylation, it is conceivable that some of the effects of HDL are mediated via this pathway46
In summary, our data suggest that in human endothelial cells, HDL can modulate COX-2 expression via p38 MAPK-dependent COX-2 mRNA transcription and stabilization. Moreover, the HDL-dependent caveolin-1 phosphorylation favors PGI-S shuttling and COX-2 coupling. These data add further insights into the molecular mechanisms involved in the anti-atherogenic activity of HDL.
Acknowledgments
We are grateful to Roberto Zecca for software assistance, to Giulio Simonutti for technical assistance with confocal microscopy, to Fabio Pellegatta, who provided the HUVECs, and to Francesco Cipollone, who provided the anti-mPGES antibody.
References
Assmann G, Schulte H, von Eckardstein A, Huang Y. High-density lipoprotein cholesterol as a predictor of coronary heart disease risk. The PROCAM experience and pathophysiological implications for reverse cholesterol transport. Atherosclerosis. 1996; 124: S11–S20.
Nofer JR, Kehrel B, Fobker M, Levkau B, Assmann G, von Eckardstein A. HDL and arteriosclerosis: beyond reverse cholesterol transport. Atherosclerosis. 2002; 161: 1–16.
Fleisher LN, Tall AR, Witte LD, Miller RW, Cannon PJ. Stimulation of arterial endothelial cell prostacyclin synthesis by high density lipoproteins. J Biol Chem. 1982; 257: 6653–6655.
Thiemermann C. Biosynthesis and interaction of endothelium-derived vasoactive mediators. Eicosanoids. 1991; 4: 187–202.
Vane JR, Bakhle Y, Botting R. Cyclooxygenase 1 and 2. Ann Rev Pharmacol Toxicol. 1998; 38: 97–120.
Bartlett SR, Sawdy R, Mann GE. Induction of cyclooxygenase-2 expression in human myometrial smooth muscle cells by interleukin-1 beta: involvement of p38 mitogen-activated protein kinase. J Physiol. 1999; 520: 399–406.
Lee SH, Soyoola E, Chanmugam P, Hart S, Sun W, Zhong H, Liou S, Simmons D, Hwang D. Selective expression of mitogen-inducible cyclooxygenase in macrophages stimulated with lipopolysaccharide. J Biol Chem. 1992; 267: 25934–25938.
Caughey GE, Cleland LG, Penglis PS, Gamble JR, James MJ. Roles of cyclooxygenase (COX)-1 and COX-2 in prostanoid production by human endothelial cells: selective up-regulation of prostacyclin synthesis by COX-2. J Immunol. 2001; 167: 2831–2838.
Linton MF, Fazio S. Cyclooxygenase-2 and atherosclerosis. Curr Opin Lipidol. 2002; 13: 497–504.
Ray WA, Stein CM, Daugherty JR Hall K, Arbogast PG, Griffin MR. COX-2 selective non-steroidal anti-inflammatory drugs and risk of serious coronary heart disease. Lancet. 2002; 360: 1071–1073.
Vinals M, Martinez-Gonzalez J, Badimon L. Regulatory effects of HDL on smooth muscle cell prostacyclin release. Arterioscler Thromb Vasc Biol. 1999; 19: 2405–2411.
Cockerill GW, Saklatvala J, Ridley SH, Yarwood H, Miller NE, Oral B, Nithyanathan S, Taylor G, Haskard DO. High-density lipoproteins differentially modulate cytokine-induced expression of E-selectin and cyclooxygenase-2. Arterioscler Thromb Vasc Biol. 1999; 19: 910–917.
Guan Z, Buckman SY, Miller BW, Springer LD, Morrison AR. Interleukin-1?-induced cyclooxygenase-2 expression requires activation of both c-Jun NH2-terminal kinase and p38 MAPK signal pathway in rat mesangial cells. J Biol Chem. 1998; 273: 28670–28676.
LaPointe MC, Isenovic E. Interleukin-1? regulation of inducible nitric oxide synthase and cyclooxygenase-2 involves the p44/44 and p38 MAPK signalling pathways in cardiac myocytes. Hypertension. 1999; 33: 276–282.
Seger R, Krebs EG. The MAPK signalling cascade. FASEB J. 1995; 9: 726–735.
Hazzalin CA, Mahadevan LC. MAPK-regulated transcription: a continuously variable gene switch? Nat Rev Mol Cell Biol. 2002; 3: 30–40.
Schulze-Osthoff K, Ferrari D, Riehemann K, Wesselborg S. Regulation of NF-kappa B activation by MAP kinase cascades. Immunobiology. 1997; 198: 35–49.
Tan Y, Rouse J, Zhang A, Cariati S, Cohen P, Comb MJ. FGF and stress regulate CREB and ATF-1 via a pathway involving p38 MAP kinase and MAPKAP kinase-2. EMBO J. 1996; 15: 4629–4642.
Xie W, Herschman HR. v-src induces prostaglandin synthase 2 gene expression by activation of the c-Jun N-terminal kinase and the c-Jun transcription factor. J Biol Chem. 1995; 270: 27622–27628.
Inoue H, Yokoyama C, Hara S, Tone Y, Tanabe T. Transcriptional regulation of human prostaglandin-endoperoxide synthase-2 gene by lipopolysaccharide and phorbol ester in vascular endothelial cells. Involvement of both nuclear factor for interleukin-6 expression site and cAMP response element. J Biol Chem. 1995; 270: 24965–24971.
Inoue H, Umesono K, Nishimori T, Hirata Y, Tanabe T. Glucocorticoid-mediated suppression of the promoter activity on the cyclooxygenase-2 gene is modulated by expression of its receptor in vascular endothelial cells. Biochem Biophys Res Comm. 1999; 254: 292–298.
Norata GD, Pellegatta F, Hamsten A, Catapano AL, Eriksson P. Effects of high density lipoprotein subfraction 3 on the expression of matrix-degrading proteases in human endothelial cells. Int J Mol Med. 2003; 12: 73–78.
Norata GD, Pirillo A, Callegari E, Hamsten A, Catapano AL, Eriksson P. Gene expression and intracellular pathways involved in endothelial dysfunction induced by VLDL and oxidised VLDL. Cardiovasc Res. 2003; 59: 169–180.
Norata GD, Bjork H, Hamsten A, Catapano AL, Eriksson P. High density lipoprotein decrease ADAMTs1 expression induced by LPS and TNF in human endothelial cells. Matrix Biology. 2004; 22: 557–560.
Kalgutkar AS, Marnett AB, Crews BC, Remmel RP, Marnett LJ. Ester and amide derivatives of the non-steroidal anti-inflammatory drug, indomethacin, as selective cyclooxygenase-2 inhibitors. J Med Chem. 2000; 43: 2860–2870.
Singer CA, Baker KJ, McCaffrey A, AuCoin DP, Dechert MA, Gerthoffer WT. P38 MAPK and NF-{kappa}B Mediate COX-2 Expression in Human Airway Myocytes. Am J Physiol. 2003; 285: L1087–L1098.
Loudon JA, Elliott CL, Hills F, Bennett PR. Progesterone represses interleukin-8 and cyclo-oxygenase-2 in human lower segment fibroblast cells and amnion epithelial cells. Biol Reprod. 2003; 69: 331–337.
Tamura M, Sebastian S, Yang S, Gurates B, Fang Z, Okamura K, Bulun SE. Induction of cyclooxygenase-2 in human endometrial stromal cells by malignant endometrial epithelial cells: evidence for the involvement of extracellularly regulated kinases and CCAAT/enhancer binding proteins. J Mol Endocrinol. 2003; 31: 95–104.
Baez JM, Barbour SE, Cohen DE. Phosphatidylcholine transfer protein promotes apolipoprotein A-I-mediated lipid efflux in Chinese hamster ovary cells. J Biol Chem. 2002; 277: 6198–6206.
Ioka RX, Kang MJ, Kamiyama S, Kim DH, Magoori K, Kamataki A, Ito Y, Takei YA, Sasaki M, Suzuki T, Sasano H, Takahashi S, Sakai J, Fujino T, Yamamoto TT. Expression cloning and characterization of a novel glycosylphosphatidylinositol-anchored high density lipoprotein-binding protein, GPI-HBP1. J Biol Chem. 2003; 278: 7344–7349.
Parfenova H, Parfenov VN, Shlopov BV, Levine V, Falkos S, Pourcyrous M, Leffler CW. Dynamics of nuclear localization sites for COX-2 in vascular endothelial cells. Am J Physiol. 2001; 281: C166–C178.
Dean JLE, Brook M, Clark AR, Saklatvala J. P38 MAPK regulates cyclooxygenase-2 mRNA stability and transcription in lipopolysaccharide-treated human monocytes. J Biol Chem. 1999; 274: 264–269.
Frank PG, Woodman SE, Park DS, Lisanti MP. Caveolin, caveolae, and endothelial cell function. Arterioscler Thromb Vasc Biol. 2003; 23: 1161–1168.
Nomura R, Fujimoto T. Tyrosine-phosphorylated caveolin-1: immunolocalization and molecular characterization. Mol Biol Cell. 1999; 10: 975–986.
Spisni E, Bianco MC, Griffoni C, Toni M, D’Angelo R, Santi S, Riccio M, Tomasi V. Mechanosensing role of caveolae and caveolar constituents in human endothelial cells. J Cell Physiol. 2003; 197: 198–204.
Lim H, Dey SK. A Novel Pathway of Prostacyclin Signalling–Hanging Out with Nuclear Receptors. Endocrinology. 2002; 143: 3207–3210.
Kimura T, Sato K, Malchinkhuu E, Tomura H, Tamama K, Kuwabara A, Murakami M, Okajima F. High-density lipoprotein stimulates endothelial cell migration and survival through sphingosine 1-phosphate and its receptors. Arterioscler Thromb Vasc Biol. 2003; 23: 1283–1288.
Chen X, Resh MD. Activation of mitogen-activated protein kinase by membrane-targeted Raf chimeras is independent of raft localization. J Biol Chem. 2001; 34617–34623.
Smith LH, Boutaud O, Breyer M Morrow JD, Oates JA, Vaughan DE. Cyclooxygenase-2-Dependent Prostacyclin Formation Is Regulated by Low Density Lipoprotein Cholesterol In Vitro. Arterioscler Thromb Vasc Biol. 2002; 22: 983–988.
Schroer K, Zhu Y, Saunders MA, Deng WG, Xu XM, Meyer-Kirchrath J, Wu KK. Obligatory role of cyclic adenosine monophosphate response element in cyclooxygenase-2 promoter induction and feedback regulation by inflammatory mediators. Circulation. 2002; 105: 2760–2765.
Inoue H, Taba Y, Miwa Y, Yokota C, Miyagi M, Sasaguri T. Transcriptional and posttranscriptional regulation of cyclooxygenase-2 expression by fluid shear stress in vascular endothelial cells. Arterioscler Thromb Vasc Biol. 2002; 22: 1415–1420.
McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX)-2: The human pharmacology of selective inhibitors of COX-2. Proc Natl Acad Sci U S A. 1999; 96: 272–277.
Griffoni C, Spisni E, Santi S, Riccio M, Guarnieri T, Tomasi V. Knockdown of caveolin-1 by antisense oligonucleotides impairs angiogenesis in vitro and in vivo. Biochem Biophys Res Commun. 2000; 276: 756–761.
Spisni E, Griffoni C, Santi S, Riccio M, Marulli R, Bartolini G, Toni M, Ullrich V, Tomasi V. Colocalization of PGI-S and caveolin-1 in endothelial cells and new role for prostacyclin in angiogenesis. Exp Cell Res. 2001; 266: 31–43.
Calabresi L, Gomaraschi M, Franceschini G. Endothelial protection by HDL. From bench to beside. Arterioscler Thromb Vasc Biol. 2003; 23: 1724–1731.
Uittenbogaard A, Shaul PW, Yuhanna IS, Blair A, Smart EJ. HDL prevents Ox-LDL-induced inhibition of eNOS localization and activation in caveolae. J Biol Chem. 2000; 275: 11278–11283.(G.D. Norata; E. Callegari)