Cyclooxygenase COX-2a, a novel COX-2 mRNA variant, in platelets from patients after coronary artery bypass grafting

There are two principal cyclooxygenase isoforms referred to as COX-1 and COX-2. Recently, COX-3 has been identified. We have demonstrated the expression of COX-2 in platelets from patients after coronary artery bypass grafting (CABG). Careful biochemical analysis revealed that, when compared to recombinant COX-2, platelet COX-2 had a slightly higher electrophoretic mobility. Two COX-2 sequences (∼1.8 kb, ∼1.7 kb) were cloned from platelet mRNA. The ∼1.7 kb sequence, designated COX-2a, differed from the human COX-2 sequence only in a deletion from position +458 to +567. Similar to the human COX-3, there is a frame shift in the COX-2a sequence resulting in a TAA stop codon at position +490. Thus, the expression of a COX-2a protein corresponding to the 67 kDa COX-2 protein is not clear. However, the marked shifting from COX-2 to COX-2a in platelets from some patients after CABG is a striking finding.

Alternative splicing of platelet cyclooxygenase-2 mRNA in patients after coronary artery bypass grafting

Recently, we cloned from platelet mRNA a novel cyclooxygenase (COX)-2 splice variant, designated COX-2a, which is characterized by a partial deletion of exon 5. Preliminary studies of mRNA distribution of COX-2 isoforms in platelets from coronary artery bypass grafting (CABG) patients showed a variable increase in COX-2a mRNA expression after cardiac surgery. Thus, we assessed whether this variant may play a functional role in these patients. We report a marked (about 20346-fold) increase in the expression of COX-2a mRNA after CABG. Evidence is presented that ribosomal frame-shifting may correct the coding sequence resulting in the expression of a full-length COX-2a pro-tein. In addition, a reading frame-corrected COX-2a mutant (COX-2aΔG) was generated by site-directed mutagenesis and expressed in COS-7 cells using an adenoviral expression system. However, COX-2a protein was not active in terms of prostaglandin formation. Thus, alternative mRNA splicing might represent an intriguing posttranscriptional mechanism to oppose a transcriptional activation of the COX-2 gene. Evolutionary, this mechanism may prevent COX-2-dependent thromboxane synthesis in the platelet, which would potentiate the likelihood of thrombosis; pharmacologically, this mechanism would prevent an aspirin-insensitive pathway of thromboxane formation.

Alternatively spliced human tissue factor (asHTF) is not pro-coagulant

It has been proposed that alternatively-spliced human tissue factor (asHTF) is pro-coagulant. We have evaluated the function of asHTF in a mammalian expression system. Full-length human tissue factor (HTF) and asHTF were cloned from smooth muscle cells and over-expressed in HEK293 cells. As expected, a marked pro-coagulant activity (FX activation, thrombin generation) was observed on the surface, in lysates, and on microparticles from HTF transfected cells. In contrast, no pro-coagulant activity of as HTF was observed.

Expression and translation of the COX-1b gene in human cells--no evidence of generation of COX-1b protein

Cyclooxygenase 1b (COX-1b) is a splice variant of COX-1, containing a retained intron 1 within the signal peptide sequence. COX-1b mRNA is found in many species, but the existence of a functionally active protein, which is possibly related to different species-dependent lengths of intron 1, is controversially discussed. The human intron 1 comprises 94 bp, and the resulting frameshift at the intron 1-exon 2 junction creates a premature stop codon. Nevertheless, full-length human COX-1b protein expression, including translated intron 1 and the signal peptide, has been reported and was explained by a frameshift repair. In this study, the fate of COX-1b mRNA in a human overexpression system is analyzed. Independent of the hypothetical frameshift repair mechanism, the splicing of the COX-1b intron 1, resulting in COX-1 mRNA and removal of the signal peptide during protein maturation, with subsequent generation of a COX-1 protein is demonstrated.

Complete Downmodulation of P-Selectin Glycoprotein Ligand in Monocytes Undergoing Apoptosis

Objectives— Apoptotic monocytes release membrane microparticles which may play a major role in thrombogenicity through a P-selectin glycoprotein ligand (PGSL-1)–mediated mechanism. We have studied systematically the regulation of PSGL-1 expression and function in apoptotic monocytic cells.

Methods and Results— PSGL-1 expression (flow cytometry, immunofluorescence microscopy, immunoblot) was virtually abolished in apoptotic monocytes by proteolytic shedding. This was accompanied by a complete loss of PSGL-1–mediated platelet–leukocyte (flow cytometry) and leukocyte–endothelial cell (parallel plate flow chamber) interactions. Systematic screening of protease inhibitors combined with knock-out and siRNA experiments characterized the PSGL-1-cleaving enzyme as an N-ethylmaleimide-inhibitable metalloproteinase of the ADAM family.

Conclusions— Downmodulation of PGSL-1 in apoptotic monocytes may prevent ectopic cell clearance in the peripheral vasculature to reduce local inflammatory and proliferative responses. Depletion of PSGL-1 expression on apoptotic microparticles may also act as a molecular switch to modulate their thrombogenic activity.

Alternative splicing of platelet cyclooxygenase-2 mRNA in patients after coronary artery bypass grafting

Recently, we cloned from platelet mRNA a novel cyclooxygenase (COX)-2 splice variant, designated COX-2a, which is characterized by a partial deletion of exon 5. Preliminary studies of mRNA distribution of COX-2 isoforms in platelets from coronary artery bypass grafting (CABG) patients showed a variable increase in COX-2a mRNA expression after cardiac surgery. Thus, we assessed whether this variant may play a functional role in these patients. We report a marked (about 200-fold) increase in the expression of COX-2a mRNA after CABG. Evidence is presented that ribosomal frame-shifting may correct the coding sequence resulting in the expression of a full-length COX-2a protein. In addition, a reading frame-corrected COX-2a mutant (COX-2a delta G) was generated by site-directed mutagenesis and expressed in COS-7 cells using an adenoviral expression system. However, COX-2a protein was not active in terms of prostaglandin formation. Thus, alternative mRNA splicing might represent an intriguing posttranscriptional mechanism to oppose a transcriptional activation of the COX-2 gene. Evolutionary, this mechanism may prevent COX-2-dependent thromboxane synthesis in the platelet, which would potentiate the likelihood of thrombosis; pharmacologically, this mechanism would prevent an aspirin-insensitive pathway of thromboxane formation.

Cholesterol Enhances Thrombin-Induced Release of Fibroblast Growth Factor-2 in Human Vascular Smooth Muscle Cells

OBJECTIVE

The mitogenic response to the G protein-coupled receptor agonist thrombin in human vascular smooth muscle cells (SMCs) depends on release of fibroblast growth factor-2 (FGF-2). Yet, intracellular mechanisms triggering FGF-2 release are unknown. The present study investigates possible effects of cholesterol enrichment and depletion, which have been shown to influence FGF-2-dependent signaling and SMC mitogenesis, on thrombin-induced FGF-2 release.

METHODS AND RESULTS

Cultured human aortic and saphenous vein SMCs were enriched with cholesterol by using a cyclodextrin-cholesterol complex. Cholesterol accumulation was determined by a fluorometric assay. ELISA, Western blotting, and RT-PCR were used for quantification of FGF-2 levels. DNA synthesis was determined by [3H]-thymidine incorporation, proliferation by cell counting. Stimulation of SMCs with thrombin (30 nmol/L) resulted in release of FGF-2 into the pericellular space within 10 minutes. Preincubation with cyclodextrin-cholesterol caused accumulation of cellular cholesterol, increased thrombin-induced FGF-2 release, and stimulated FGF-2 de novo synthesis. Thrombin-induced DNA synthesis and proliferation were enhanced in cholesterol-rich SMCs. This effect was inhibited by FGF-2-neutralizing antibodies.

CONCLUSIONS

Enhanced cellular cholesterol stimulates thrombin-induced release of FGF-2 and increases the mitogenic response toward thrombin in human SMCs. This mechanism might also be relevant for thrombin-induced mitogenesis in hypercholesterolemia in vivo.

Alternatively spliced human tissue factor (asHTF) is not pro-coagulant

It has been proposed that alternatively-spliced human tissue factor (asHTF) is pro-coagulant. We have evaluated the function of asHTF in a mammalian expression system. Full-length human tissue factor (HTF) and asHTF were cloned from smooth muscle cells and over-expressed in HEK293 cells. As expected, a marked pro-coagulant activity (FX activation, thrombin generation) was observed on the surface, in lysates, and on microparticles from HTF transfected cells. In contrast, no pro-coagulant activity of asHTF was observed.

Human cyclooxygenase-1b is not the elusive target of acetaminophen

A cyclooxygenase-1 splice variant (cyclooxygenase-1b), cloned from canine brain, was proposed to be an acetaminophen-sensitive enzyme. Unlike in canines, the retention of intron 1 in the human sequence results in a frame shift and predicts a truncation of the protein. We have sought to answer the question whether human cyclooxygenase-1b, if expressed, is a target of acetaminophen. Thus, we studied the pharmacology of complete human cyclooxygenase-1b in which the frame shift was corrected by site-directed mutagenesis (human cyclooxygenase-1bDeltaG). Human cyclooxygenase-1bDeltaG was active but was not inhibited by acetaminophen. In conclusion, full length human cyclooxygenase-1b is clearly not the target of acetaminophen.

Rapid Release of Active Tissue Factor From Human Arterial Smooth Muscle Cells Under Flow Conditions

Circulating tissue factor (TF) is an important determinant of coronary thrombosis. Among other cell types, such as monocytes, vascular smooth muscle cells (SMCs) are capable of releasing TF. When studied under static conditions, SMCs do release TF, but this process is slow and, thus, cannot explain the elevated levels of circulating TF, as observed in patients with acute coronary syndromes. The present study demonstrates that cultured human mammary artery SMCs very rapidly (minutes) release active, microparticle-bound TF when exposed to flow conditions. There was a clear log-linear correlation between the shear rate (range 10 s(-1) to 1500 s(-1)) and the procoagulant activity of SMC perfusates. Flow-dependent release of TF was transient (10 minutes) and did not measurably reduce cell surface TF content. Interestingly, a time-dependent (t(1/2) 30 minutes) re-exposure of releasable TF was detected after a no-flow period. These data demonstrate that SMCs may become a pathophysiologically relevant source of TF that can be rapidly released into the circulation in situations in which endothelial damage occurs and SMCs come into a close contact with the flowing blood.

Rapid Release of Active Tissue Factor From Human Arterial Smooth Muscle Cells Under Flow Conditions

Circulating tissue factor (TF) is an important determinant of coronary thrombosis. Among other cell types, such as monocytes, vascular smooth muscle cells (SMCs) are capable of releasing TF. When studied under static conditions, SMCs do release TF, but this process is slow and, thus, cannot explain the elevated levels of circulating TF, as observed in patients with acute coronary syndromes. The present study demonstrates that cultured human mammary artery SMCs very rapidly (minutes) release active, microparticle-bound TF when exposed to flow conditions. There was a clear log-linear correlation between the shear rate (range 10 s −1 to 1500 s −1 ) and the procoagulant activity of SMC perfusates. Flow-dependent release of TF was transient (10 minutes) and did not measurably reduce cell surface TF content. Interestingly, a time-dependent (t 1/2 30 minutes) re-exposure of releasable TF was detected after a no-flow period. These data demonstrate that SMCs may become a pathophysiologically relevant source of TF that can be rapidly released into the circulation in situations in which endothelial damage occurs and SMCs come into a close contact with the flowing blood. ( Arterioscler Thromb Vasc Biol . 2006;26:e34–e37.)

Thermodynamics of apocalmodulin and nitric oxide synthase II peptide interaction

The Ca2+-free form of calmodulin (CaM), apocalmodulin (ApoCaM), regulates a variety of target proteins including nitric oxide synthase II (NOS-II). The CaM-binding site of NOS-II can bind ApoCaM with high affinity. Substitution of hydrophobic amino acids by charged amino acids at crucial positions 3, 9 and 13 within the CaM-binding motif did not abolish the ApoCaM interaction that occurred with significant affinity, though the affinity of the interaction was decreased remarkably. Isothermal titration calorimetry revealed that interaction of ApoCaM and synthetic NOS-II peptides was driven entropically.

Regulation of thrombomodulin expression in human vascular smooth muscle cells by COX-2-derived prostaglandins

There is concern that cyclooxygenase (COX)-2 inhibitors may promote atherothrombosis by inhibiting vascular formation of prostacyclin (PGI2) and an increased thrombotic risk of COX-2 inhibitors has been reported. It is widely accepted that the prothrombotic effects of COX-2 inhibitors can be explained by the removal of platelet-inhibitory PGI2. Using microarray chip technology, we have previously demonstrated that thrombomodulin (TM) mRNA is upregulated in cultured human coronary artery smooth muscle cells by the stable prostacyclin mimetic iloprost. This study is the first to demonstrate a stimulation of the expression of functionally active thrombomodulin in human smooth muscle cells by prostaglandins, endogenously formed via the COX-2 pathway. Because TM is an important inhibitor of blood coagulation, these findings provide a novel platelet-independent mechanism to explain the prothrombotic effects of COX-2 inhibitors. The full text of this article is available online at http://circres.ahajournals.org.

Reaction rate constants of superoxide scavenging by plant antioxidants

Plant phenols may exert protective effects by scavenging superoxide, which is implicated in tissue damage and accelerated inactivation of vasorelaxing nitric oxide. Preventing the interaction of superoxide with tissue biomolecules depends not only on the extent of superoxide scavenging but also on scavenging velocity. However, information on superoxide scavenging kinetics of plant phenols is scarce. We describe an improved lucigenin-based chemiluminescence assay for kinetic analysis. The use of potassium superoxide (KO2) as a nonenzymatic superoxide source allowed simple and reliable determination of the second-order reaction rate constants between superoxide and plant antioxidants at physiologically relevant conditions, avoiding unspecific effects of other reactive oxygen species or superoxide-generating enzymes. We calculated the rate constants for phenols of different structures, ranging from 2.9 x 10(3) mol(-1) l s(-1) for morin to 2.9 x 10(7) mol(-1) l s(-1) for proanthocyanidins. Compounds with pyrogallol or catechol moieties were revealed as the most rapid superoxide scavengers, and the gallate moiety was found to be the minimal essential structure for maximal reaction rate constants with superoxide.

Target recognition of apocalmodulin by nitric oxide synthase I peptides

An increasing number of proteins are found that are regulated by the Ca(2+)-free state of calmodulin, apocalmodulin. Many of these targets harbor a so-called IQ motif within their primary sequence, but several target proteins of apocalmodulin lack this motif. We investigated whether the Ca(2+)-dependent calmodulin-binding site of nitric oxide synthase I could be transformed into a target site of apocalmodulin. Synthetic peptides representing the wild-type amino acid sequence and several peptides carrying mutations were studied by isothermal titration calorimetry and fluorescence spectroscopy. A single amino acid substitution of a negative charge to a positive charge can convert a classical Ca(2+)-dependent binding site of calmodulin into a target site for apocalmodulin. In addition, the introduction of hydrophobic amino acids increases the apparent binding affinity from the micromolar to the nanomolar range. Binding of wild-type and mutant peptides to Ca(2+)-calmodulin was enthalpically driven, and binding to apocalmodulin was entropically driven. Our data indicate that only a few selected amino acid positions in a calmodulin-binding site determine its Ca(2+) dependency.