Acyl chain-dependent effect of lysophosphatidylcholine on cyclooxygenase (COX)-2 expression in endothelial cells.

OBJECTIVE: Previously we identified palmitoyl-, oleoyl- linoleoyl-, and arachidonoyl-lysophosph-atidylcholine (LPC 16:0, 18:1, 18:2 and 20:4) as the most prominent LPC species generated by endothelial lipase (EL). In the present study, we examined the capacity of those LPC to modulate expression of cyclooxygenase (COX)-2 in vascular endothelial cells. METHODS &
RESULTS: LPC 16:0 and 20:4 promoted both COX-2 mRNA- and protein synthesis with different potencies and kinetics. While LPC 18:1 induced a weak and transient increase in COX-2 mRNA, but not protein, LPC 18:2 increased COX-2 protein, without impacting mRNA. Chelation of intracellular Ca(2+) and inhibition of p38 MAPK markedly attenuated 16:0 LPC- and 20:4 LPC- elicited induction of COX-2 expression, whereas inhibition of phospholipase C (PLC) attenuated only the effect of 16:0 LPC. LPC 16:0 and 20:4 differed markedly in their potencies to increase cytosolic Ca(2+) concentration and in the kinetics of p38 MAPK activation. While the effects of 16:0 and 20:4 LPC on COX-2 expression were profoundly sensitive to silencing of either c-Jun or p65 (NF-κB), respectively, silencing of cyclic AMP responsive element binding protein (CREB) attenuated markedly the effect of both LPC.
CONCLUSION: Our results indicate that the tested LPC species are capable of inducing COX-2 expression, whereby the efficacy and the relative contribution of underlying signaling mechanisms markedly differ, due to the length and degree of saturation of LPC acyl chains.

Introduction

Saturated lysophosphatidylcholine (LPC), palmitoyl (16:0) LPC is generated by a variety of reactions including: i) the cleavage of plasma membrane- and lipoprotein-phosphatidylcholine (PC) by various phospholipase A2 (PLA2) enzymes [1], ii) lecithin cholesterol acyltransferase (LCAT) activity in high-density lipoprotein (HDL) [2], and iii) oxidation of low-density lipoprotein (LDL) [3]. Additional sources of LPC are endothelial lipase (EL) and hepatic lipase (HL), which by cleaving HDL-PC generate in addition to 16:0 LPC, substantial amounts of unsaturated LPC 18:1, 18:2 and 20:4, respectively [4,5]. These LPC are among the most abundant LPC in human plasma [6].

The physiological concentrations of LPC in plasma is high, around 190 μM [6] with even millimolar levels in hyperlipidemic subjects [7]. LPC in plasma are distributed between albumin and other carrier proteins and lipoproteins [8,9] with the likely transient existence of minute amounts of free LPC. This free LPC might occur during an excessive lipolysis and concomitant saturation of albumin and carrier proteins with fatty acids (FA) and LPC, leading to interaction of the free LPC with cells. In vascular endothelial cells 16:0 LPC was shown to activate numerous signaling pathways thereby promoting expression of various molecules [10,11], including cyclooxygenase-2 (COX-2) [12,13].

COX enzymes are rate-limiting in the conversion of arachidonic acid to prostanoids. Vascular endothelial cells constitutively express both COX isoforms, COX-1 and COX-2 [14-16]. The expression of COX-2 can markedly be augmented by various stimuli, including growth factors and cytokines [12,13]. The COX-2 promoter contains binding sites for various transcription factors including cyclic AMP-response element (CRE)-binding protein, activator protein-1 (AP-1), nuclear factor-IL6/CCAAT enhancer-binding protein (C/EBP), signal transducer and activator of transcription (STAT3), SP1 and nuclear factor (NF)-κB [17].

Studies addressing the impact of LPC on endothelial COX-2 expression used exclusively 16:0 LPC [12,13]. In our previous study in human aortic endothelial cells (HAEC), LPC 16:0, 18:1, 18:2 and 20:4 only slightly increased COX-2 mRNA without affecting COX-2 protein expression [18]. Therefore, we addressed in the present study the capacity and underlying mechanisms of those LPC on COX-2 expression in human endothelial cell line EA.hy 926 [19], found to be responsive to LPC in terms of upregulation of both COX-2 mRNA and protein.

Materials and methods

LPC

LPCs (16:0, 18:1, 18:2 and 20:4 LPC) were purchased from Avanti Polar Lipids, Alabaster, AL. LPCs were dissolved in chloroform/methanol solution under argon atmosphere and stored at −20 °C. For cell culture experiments, required amounts of LPCs were dried/evaporated under a stream of nitrogen or argon and re-dissolved in PBS (pH 7.4).

Cell culture

Human endothelial cell line EA.hy 926 [19] was cultured in Dulbecco's modified Eagle medium (DMEM) (Gibco) containing 10% fetal bovine serum (FBS) (PAA, Pasching, Austria) and 1% HAT Media Supplement (100 μM hypoxanthine, 0.4 μM aminopterin, 16 μM thymidine). Cell culture medium was supplemented with penicillin G sodium sulfate (100 units/ml), streptomycin sulfate (100 μg/ml), and amphotericin B (2.5 μg/ml). Cells were cultured in humidified atmosphere of 5% CO2/95% air at 37 °C and were sub-cultured using 0.025% trypsin/0.01% EDTA.

LPC treatment of EA.hy 926 cells

Twenty 4 h after plating of cells into 6- or 12-well plates, the culture medium was replaced with the fresh culture medium (10% FBS) supplemented with 200 μM LPCs. After the respective incubation times, cells were washed with PBS twice and lysed in buffers for isolation of RNA or proteins.

Pharmacological inhibitors

Cells were pre-treated with respective pharmacological inhibitors or vehicle (DMSO) for 30 min before the addition of fresh culture medium containing LPC supplemented with vehicle or inhibitors, respectively. Following inhibitors were applied: Ca2+ chelator BAPTA/AM (10 μM), p38 MAPK inhibitor SB203580 (5 μM), Actinomycin D (800 nM), phospholipase C (PLC) inhibitor U73122 (2 μM) and cycloheximide (100 μM). All inhibitors except BAPTA/AM (Calbiochem) were from Sigma, St Louis, MO. The applied LPC, pharmacological inhibitors or their combination, respectively, were not toxic to cells as determined by monitoring the release of lactate dehydrogenase (LDH) using the cytotoxicity detection kit (Roche, Mannheim, Germany).

Quantitative real-time PCR

Total cellular RNA was isolated using the peqGOLD Total RNA Kit (Peqlab-biotechnology, Erlangen, Germany) according to the manufacturer's protocol. Then, 0.8 μg of total cellular RNA were reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) and 0.8 U of an RNase Inhibitor (Qiagen, Hilden, Germany). Real-time PCR analysis was performed as described previously [18].

Western blotting

EA.hy 926 cells were stimulated with LPCs (200 μM, 10% FBS) or PBS for indicated times, rinsed twice with ice-cold PBS and lysed in RIPA buffer [25 mM Tris–HCl (pH 7.6), 1% NP-40, 150 mM NaCl, 1% sodium deoxycholate, 0.1% SDS] (Pierce) containing protease inhibitor cocktail (Sigma–Aldrich). Protein concentration was determined with the BCA protein assay kit (Novagen, Darmstadt, Germany), according to the manufacturers instructions. Equal amounts of cell lysate protein samples were subjected to gel electrophoresis using 10% SDS-polyacrylamide gels followed by the transfer to nitrocellulose membrane. Proteins were detected by antibodies specific for COX-2 (1:200), β-actin (1:1000) (both Santa Cruz Biotechnology, Heidelberg, Germany), RelA, CREB, c-Jun, p38 or phospho-p38 (Thr180/Tyr182) (all 1:1000; Cell Signaling, Danvers, MA), followed by appropriate HRP-conjugated secondary antibodies (Dako). Antibody binding was visualized using a SuperSignal West Pico (Thermo Scientific) detection system. Densitometry analyses were carried out using ImageJ.

siRNA transfection

Transfections were carried out using Nanofectin siRNA (PAA, Pasching, Austria) according to the manufacturer's instructions. EA.hy 926 cells were plated (50 000/well) in 12-well plates 24 h before transfection. One hour before transfection, 50–60% confluent cells were washed with PBS and incubated in DMEM lacking the serum and antibiotics. The transfection mixture was prepared by mixing 40 pmol (0.5 μg siRNA) of the respective siRNA with Nanofectin siRNA reagent in the ratio of 1:3 (μg of siRNA: μl of Nanofectin siRNA reagent) in medium without antibiotics and serum. After 20 min, the transfection complex was added drop-wise to the cells. After 3 h the medium was exchanged with the fresh one containing antibiotics and serum. Cells were collected after 48 h for RNA or protein isolation. The siRNA used were p65 (RelA) (SI003001672), c-Jun (SI00300580), CREB1 (SI00299894) and control (nonsilencing) siRNA (SI001027281) (Qiagen).

Measurements of intracellular Ca2+

Intracellular Ca2+ concentration [Ca2+]i was determined using the fluorescent Ca2+ indicator fura-2-acetoxymethyl ester (Fura-2/AM) as described [20]. EA.hy 926 cells were grown in T75 flasks until reaching confluency followed by serum starvation for 2 h and loading with 2 μM Fura-2/AM for 45 min. Then the cells were washed, trypsinized and for the measurements in the presence of Ca2+ resuspended either in Ca2+-containing HEPES-buffer (138 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM glucose and 10 mm N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES)/NaOH, pH = 7.4), or a Ca2+ free buffer, which contained 0.1 mM EGTA instead of CaCl2. The Fura-2 fluorescence intensity was monitored spectrophotometrically in a stirring cuvette at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm before and after the injection of 3 μM LPCs. Data were expressed as ratio values of Fura-2 fluorescence (340/380 nm).

Statistical analysis

Experiments were performed at least three times and the data are represented as the mean ± standard deviation (S.D.). Mean values were compared using one-way Anova followed by Bonferroni's multiple comparison test. Statistically significant differences between groups are indicated by P-values of <0.05 (*), <0.01 (**), or <0.001 (***).

Experimental results

LPC induce COX-2 mRNA and protein in endothelial cells

To examine the impact of various LPC on COX-2 expression cells were exposed to respective LPCs at concentration of 200 μM in the presence of 10% FBS. The relationship between LPC and FBS in the present study was similar as in previous studies [10,12], and required to overcome scavenging of LPC by serum proteins thus yielding the free LPC capable of evoking cellular responses. As shown in Fig. 1A, 16:0 LPC elicited a 3.7-fold increase in COX-2 mRNA as early as after 1 h of incubation, followed by a sustained increase and a maximal 61- and 41-fold upregulation after 16 and 24 h, respectively. In contrast to 16:0 LPC, 18:1 LPC elicited a weak and transient 1.5-fold increase in COX-2 mRNA obvious only after 1 h of incubation, whereas 18:2 LPC had no effect on COX-2 mRNA expression. 20:4-LPC elicited a significant 2.4-fold increase after 3 h of incubation with a further sustained increase after 5 h (3.6-fold) and 8 h (4.7-fold), respectively, followed by a decrease after 16 h (2.4-fold) and a maximal 13-fold upregulation after 24 h. The observed, pronounced 16:0 LPC- and 20:4 LPC-elicited upregulation of COX-2 mRNA could be completely suppressed by actinomycin D (Fig. 1B). In line with the pronounced upregulation of COX-2 mRNA, both 16:0 and 20:4 LPC yielded elevation of cellular COX-2 protein content, however, with different potency and kinetics (Fig. 1C). While 18:1 LPC failed to promote COX-2 protein expression (not shown), COX-2 protein was increased after 5 h of incubation with 18:2 LPC and this increase could be completely abolished with cycloheximide (Fig. 1D).

Fig. 1

LPC induce COX-2 mRNA and protein expression. EA.hy 926 cells were incubated with 200 μM LPC or PBS in medium containing 10% FBS for the indicated time points, followed by the determination of (A) COX-2 mRNA abundance by qRT-PCR and (C) COX-2 protein by Western blotting. In (B) cells were pre-incubated with actinomycin D (ActD) (800 nM) for 30 min before addition of LPC followed by incubation without or with LPC 16:0 for 1 h and LPC 20:4 for 4 h in the absence or presence of ActD. Subsequently, relative COX-2 mRNA expression was determined by qRT-PCR as detailed in Section 2. In (D) cells were pre-incubated with cycloheximide (+CHX) (100 μM) or vehicle (-CHX) followed by incubation without or with LPC 18:2 in the presence or absence of CHX for 5 h. Subsequently, expression of COX-2 protein was analyzed by Western blotting. Results shown in (A) and (B) are mean ± SD of three independent experiments done in triplicates. In (A) the values of LPC samples were normalized to those of PBS, which were set to 1. Results in (C) and (D) are representative Western blots. *, **, *** indicate significant differences between LPC and PBS treated cells. #, ##, ### indicate significant differences between LPC treated cells in the presence and absence of applied inhibitors.

The role of intracellular Ca2+ in LPC-elicited COX-2 mRNA upregulation

Since alterations in intracellular Ca2+ concentration [Ca2+]i are implicated in activation of various signaling pathways [21] and in turn in regulation of gene expression [22], we examined the impact of intracellular Ca2+ buffering (by loading the cells with the Ca2+ chelator BAPTA/AM) as well as of the PLC inhibitor U73122 on the COX-2 mRNA upregulation elicited by LPC 16:0 and 20:4. As shown in Fig. 2A intracellular Ca2+ buffering and inhibition of PLC suppressed 16:0 LPC- elicited induction of COX-2 mRNA. In contrast, 20:4 LPC-elicited induction of COX-2 mRNA was decreased by BAPTA/AM but the effect of U73122 was weak, without reaching statistical significance (Fig. 2B). Interestingly, BAPTA/AM increased COX-2 mRNA in control PBS-treated cells (Fig. 2B). In the presence of extracellular Ca2+ 16:0 LPC-induced increase in [Ca2+]i was markedly higher compared to that induced by 20:4 LPC (Fig. 2C). In the absence of extracellular Ca2+ (EGTA) the increase in [Ca2+]i evoked by both LPC was profoundly lower, however, still higher upon exposure to 16:0 than 20:4 LPC (Fig. 2D). Under EGTA-conditions both U73122 and 2-APB, an inositol-3 phosphate (IP3) receptor antagonist, completely blocked the increase in [Ca2+]i (supplementary Fig. 1), indicating that LPC-elicited increase in intracellular Ca2+ via PLC/IP3-mediated mechanism triggered COX-2 upregulation.

Fig. 2

Intracellular Ca2+ is involved in LPC-elicited COX-2 upregulation. EA.hy 926 cells were pre-incubated with the intracellular Ca2+ chelator (BAPTA/AM) (10 μM) and PLC inhibitor U73122 (2 μM) for 30 min before they were exposed to 200 μM LPC without or with inhibitors in medium containing 10% FBS for 1 h (LPC 16:0) (A) and 4 h (LPC 20:4) (B). Subsequently, relative COX-2 mRNA expression was determined by qRT-PCR. Results shown in (A) and (B) are mean ± SD of three independent experiments done in triplicates (*, **, *** indicate significant differences between LPC and PBS treated cells; #, ##, ### indicate significant differences between LPC treated cells in the presence and absence of respective inhibitors). (C) and (D): Fura-2/AM-loaded EA.hy 926 cells were trypsinized and resuspended either in Ca2+-containing buffer (C) or Ca2+-free buffer containing EGTA (D). The ratio of fura-2 fluorescence intensity at the two excitation wavelengths (340/380 ratio) was monitored spectrophotometrically in a stirring cuvette during exposure to 3 μM LPC. Results are representative single traces out of four experiments performed in duplicates.

p38 MAPK mediates LPC-elicited COX-2 mRNA upregulation

Since p38 MAPK has been shown to mediate the induction of COX-2 mRNA triggered by various stimuli, we tested its role in LPC-elicited COX-2 mRNA upregulation. As shown in Fig. 3A and B upregulation of COX-2 mRNA elicited by both LPC 16:0 and 20:4 was efficiently attenuated by SB203580, a specific p38 MAPK inhibitor. Importantly, the kinetics of p38 MAPK activation (phosphorylated form), differed markedly upon exposure of cells to LPC 16:0 and 20:4. While 16:0 LPC induced very rapid and transient phosphorylation with a maximum after 30 min of incubation (Fig. 3C), 20:4 LPC promoted a sustained, long lasting phosphorylation of p38 MAPK with a maximum achieved after 3 h (Fig. 3D).

Fig. 3

p38 MAPK is involved in LPC-elicited COX-2 mRNA upregulation. EA.hy 926 cells were pre-incubated with SB203580 a p38 MAPK inhibitor (5 μM), for 30 min before they were exposed to 200 μM LPC in the presence of vehicle or inhibitor in medium containing 10% FBS for 1 h (LPC 16:0) (A) and 4 h (LPC 20:4) (B). Subsequently, relative COX-2 mRNA expression was determined by qRT-PCR. Results shown in (A) and (B) are mean ± SD of three independent experiments done in triplicates. (*, **, *** indicate significant differences between LPC and PBS treated cells; #, ##, ### indicate significant differences between LPC treated cells in the presence of vehicle or inhibitor). (C) and (D): EA.hy 926 cells were incubated with 200 μM LPC in medium containing 10% FBS for indicated time periods. Subsequently, phosphorylated (p) and total (t) p38 MAPK were analyzed by Western blot as described in Section 2.

NF-κB, AP-1 and CREB mediate LPC-elicited COX-2 protein upregulation

To examine the role of selected transcription factors in 16:0 and 20:4 LPC-elicited COX-2 protein upregulation, experiments were performed in cells with silenced p65 subunit of NF-κB, CREB and c-Jun, a component of AP-1. As revealed by Western blotting, the silencing procedure completely depleted p65 and CREB and markedly reduced c-Jun. The impact of p65 silencing on COX-2 protein in 16:0 LPC treated cells was less pronounced compared with that of c-Jun and CREB (Fig. 4B). In contrast, the impact of p65 silencing had a more profound attenuating effect on 20:4 LPC-, compared with 16:0 LPC-elicited COX-2 protein induction and was similar to the impact of CREB silencing (Fig. 4B). The impact of CREB silencing was similar for both LPC species (Fig. 4B). While c-Jun efficiently attenuated COX-2 protein expression elicited by 16:0 LPC (Fig. 4B), the data obtained for 20:4 LPC were controversial and not conclusive as in a set of experiments c-Jun silencing decreased COX-2 protein but in another led to increased COX-2 protein (Supplementary Fig. 2).

Fig. 4

NF-κB, c-Jun and CREB are involved in LPC-elicited COX-2 protein upregulation. 48 h after transfection with p65, c-Jun, CREB or scrambled control siRNA (control), EA.hy 926 cells were treated with 16:0- or 20:4-LPC for 16 and 8 h, respectively. Subsequently, protein levels of (A) p65, c-Jun and CREB as well as of (B) COX-2 ere analyzed by Western blot as described in Section 2. Representative images are presented. COX-2 densitometric values are normalized to β-actin and expressed as fold-induction, setting the value of cells transfected with control siRNA and treated with LPCs as 1. Bars represent the mean ± SD (n = 6). (*, **, *** indicate significant difference between LPC treated cells transfected with specific or control siRNA).

Discussion

Identification of LPC 16:0, 18:1, 18:2 and 20:4 as major LPC species generated by the action of EL on HDL [4], capable of promoting endothelial prostacyclin and IL-8 production in acyl-chain-dependent manner [10,18], prompted us to examine the capacity of those LPC to modulate COX-2 expression in human endothelial cell line EA.hy 926.

The LPC concentration of 200 μM in the presence of 10% FBS was not toxic to cells and was within the (patho)physiological range of 190 μM-1.7 mM [6,7,23]. Due to LPC scavenging by albumin and other serum proteins [8], an appropriate ratio between the amounts of LPC and FBS is required to yield the free LPC capable of evoking cellular responses [10,18]. The relationship between LPC and FBS in the present study was similar as in previous studies, where cells were exposed to 100 μM LPC in the presence of 5% FBS [10,12].

In contrast to our previous study in human aortic endothelial cells (HAEC) [18], where LPC elicited only upregulation of COX-2 at mRNA but not protein level, in the present study performed in the human umbilical vein endothelial cell (HUVEC)-derived cell line EA.hy 926, LPC yielded increases in both COX-2 mRNA and protein content. This discrepancy might be explained by the different experimental conditions: 200 μM LPC in the presence of 10% FBS in the present study versus 10 μM LPC in the absence of serum in the previous study. Additionally, the fact that EA.hy 926 cells originate from HUVEC where both COX-2 mRNA and protein were upregulated upon exposure to LPC 16:0 [12], might explain the observed better responsiveness of EA.hy 926 cells than HAEC in terms of COX-2 upregulation.

In the present study we found a remarkable acyl chain-dependent differences in the capacity of LPC to induce COX-2 expression. The most potent induction of COX-2 mRNA by 16:0 LPC might be explained by its more pronounced capacity to increase [Ca2+]i. Indeed, both chelation of intracellular Ca2+ and inhibition of PLC markedly attenuated the effect of 16:0 LPC on COX-2 mRNA. Along these lines, a negligible effect of LPC 18:1 and a complete failure of LPC 18:2 to induce COX-2 mRNA might be due to their weaker capacity to increase [Ca2+]i (Supplementary Fig. 3) [18]. However, the 20:4 LPC-elicited induction of COX-2 mRNA, although lower than that induced by 16:0 LPC, was pronounced and sustained, despite a weak capacity of 20:4 LPC to increase [Ca2+]i (Supplementary Fig. 3), suggesting additional Ca2+-independent mechanism(s). Because BAPTA/AM increased basal COX-2 mRNA and PLC inhibitor U73122, known to be unstable under prolonged cell culture conditions [24], failed to decrease 20:4 LPC-induced COX-2 mRNA, the impact of intracellular Ca2+ on 20:4 LPC-mediated upregulation of COX-2 mRNA could not be conclusively evaluated.

In contrast to the role of intracellular Ca2+, the role of p38 MAPK in 20:4 LPC induction of COX-2 mRNA was more pronounced (Fig. 3B). Interestingly, in contrast to a rapid and transient activation of p38 MAPK induced by 16:0 LPC (Fig. 3C), which was similar as shown previously [25], 20:4 LPC elicited a sustained, long-lasting p38 MAPK activation, which might be responsible for a substantial COX-2 mRNA increase independent of, or in addition to rise in [Ca2+]i. The observed decrease in COX-2 mRNA after 16 h of incubation with 20:4 LPC followed by a pronounced increase after 24 h might reflect the action of putative, bioactive 20:4 LPC conversion products, such as oxidized or enzymatically modified LPC, whose bioavailability increases during prolonged incubation.

Although COX-2 mRNA upregulation induced by LPC 16:0 and 20:4 was completely suppressed by actinomycin D, indicating that the increase in COX-2 mRNA levels occurs mainly at the transcriptional level, we can not exclude that the observed increase was at least in part due to stabilization of COX-2 mRNA. This is likely, because both LPC 16:0 and 20:4 activate p38 MAPK, a potent promoter of COX-2 mRNA stability [25].

Independently of whether induction of transcription or stabilization of mRNA or both are responsible for COX-2 mRNA upregulation, the relative contribution of underlying signaling pathways and induction of COX-2 protein are markedly different for LPC 16:0 and 20:4. It is well established that the amplitude and duration of Ca2+ signals control differential activation of signaling kinases and transcription factors [22]. Therefore, it is conceivable that LPC-induced Ca2+ signals, which were in particular different for 16:0 LPC and 20:4 LPC, were at least in part responsible for differential relative contributions of the tested transcription factors, which in turn might have impacted the kinetics and magnitude of COX-2 induction. The relative contribution of activated p38 MAPK, capable of activating CREB and AP-1 [26], was most probably independent of Ca2+ since BAPTA/AM failed to prevent 16:0 and 20:4 LPC -induced p38 MAPK activation (not shown).

The reason for observed both COX-2 up- and down-regulation upon c-Jun silencing in 20:4 LPC- treated cells is not clear. However, one can speculate that incomplete c-Jun depletion (Fig. 4A), together with slight inter-experimental differences in the efficiency of c-Jun silencing might impact the relative abundance of various AP-1 complexes in terms of homo- or heterodimerization of c-Jun with c-fos or ATF1/2 [27]. Thus in case of a prevalence of c-Jun/c-fos heterodimers, known to possess strong AP-1 binding activity [28], the c-Jun silencing would decrease 20:4 LPC-induced COX-2 upregulation. Conversely, due to a strong CRE binding activity of c-Jun/ATF [29], and concomitantly diminished CREB-mediate COX-2 upregulation (Fig. 4B), the c-jun silencing might lead to COX-2 upregulation.

Another striking finding of the present study was upregulation of COX-2 protein but not mRNA by 18:2 LPC. Future experiments should reveal mechanisms underlying 18:2 LPC-elicited COX-2 protein upregulation unrelated to COX-2 mRNA.

Together, here we show that the tested LPC exhibited remarkably different, acyl chain-related potencies and kinetics of COX-2 induction, dependent on intracellular Ca2+, p38 MAPK, NF-κB, c-Jun and CREB.

Considering their high plasma levels and their simultaneous action on vascular endothelium in vivo, the tested LPC might be important stimuli implicated in the maintenance of the basal endothelial COX-2 expression. This is crucial for vascular health, taking into account an increased incidence of cardiovascular events in patients upon inhibition of COX-2 by “coxibs” [30]. On the other hand, strikingly increased plasma levels of those LPC, as found in hyperlipidemia [7], might trigger overexpression of COX-2, leading to overproduction of vasoconstricting and proinflammatory prostanoids and in turn impaired endothelial and vascular function.

Conflict of interest

Authors declare no conflict of interest.