Comparison of static incubation versus physiologic perfusion techniques for quantitation of luminal release of prostacyclin and thromboxane in canine arteries and veins

Intraluminal pressure modulates eicosanoid enzyme expression in vascular endothelium of intact human conduit vessels at physiological levels of shear stress

OBJECTIVE

Biosynthesis of eicosanoid metabolites in blood vessels regulates vascular tone and platelet function. We investigated whether intraluminal pressure modulates gene and protein expression of key eicosanoid enzymes in intact human conduit vessels and/or release of their vasoactive metabolites.

METHODS

Paired segments of human umbilical veins were perfused under laminar flow for 1.5, 3 and 6 h at high versus low intraluminal pressure (40/20 mmHg) with identical shear stress (10 dyn/cm(2)). Endothelial cell mRNAs encoding cyclooxygenase-1 (COX-1), cyclooxygenase-2 (COX-2), prostaglandin synthase (PGS), and thromboxane synthase (TXS) were measured by quantitative real-time RT-PCR. Secretion of PGI2 and TXA2 to the perfusion medium was measured by enzyme immunoassay of their metabolites 6-keto-prostaglandin F(1alpha) and TXB2.

RESULTS

Intraluminal pressures were 39.9 +/- 0.02 and 20.0 +/- 0.03 mmHg (P < 0.0001) in high and low pressure circuits, and shear stress levels were 10.6 +/- 0.60 and 9.7 +/- 0.36 dyn/cm(2) (NS, not significant). COX-1 mRNA was significantly up-regulated after 1.5 h of high pressure stimulation and continued up to 3 h, but fell thereafter significantly below baseline after 6 h. COX-2 mRNA was initially significantly down-regulated, followed by a significant up-regulation after 6 h. Gene expressions of PGS and TXS were significantly induced after 6 h of high pressure perfusion. High pressure depressed the production of PGI(2) (P < 0.05) but did not alter TXA(2) formation.

CONCLUSIONS

Intraluminal pressure has differential effects on gene and protein expression of key eicosanoid enzymes and biosynthesis of prostanoid metabolites in intact human conduit vessels. The new, computerized biomechanical perfusion system may be a useful tool to elucidate specific effects of various biomechanical forces on intact mammalian conduit vessels.

Effects of shear stress on eicosanoid gene expression and metabolite production in vascular endothelium as studied in a novel biomechanical perfusion model

This study investigated the effects of shear stress on gene expression of prostacyclin synthesis-related enzymes cyclooxygenases (COX-1 and COX-2), prostacyclin synthase (PGS), and thromboxane synthase (TXS) and their metabolites prostaglandin (PGI(2)) and thromboxane A(2) (TXA(2)) in endothelium of intact conduit vessels. Paired human umbilical veins were perfused at high/low shear stress (25/<4 dyn/cm(2)) at identical intraluminal pressure (20 mmHg) for 1.5, 3, or 6 hours in a computerized vascular model. High shear perfusion induced a significant, monophasic upregulation of PGS and TXS gene expressions after 6 hours. COX-1 and COX-2 mRNA showed a biphasic response with peaks at 1.5 and 6 hours, with a nadir level at 3 hours. Shear-induced gene expression was associated with a significantly greater accumulation of 6-keto prostaglandin F(1alpha) and TXA(2) in the perfusion medium. Thus, shear stress independently of perfusion pressure alters the expression of prostacyclin synthesis-related enzymes and the biosynthesis of their vasoactive metabolites.

Prostanoid production in the umbilicoplacental arterial tree relative to impaired glucose tolerance

The purpose of this study was to investigate prostanoid synthesis in different segments of the umbilicoplacental vascular tree and its relationship to impaired maternal glucose tolerance. Segments from the umbilical artery and vein, allantochorionic artery branches, and the cotyledon artery from 21 women with diabetes or impaired glucose tolerance and 10 healthy women were studied. Production of prostacyclin (PGI2) and thromboxane (TxA2) metabolites was determined. The Mann-Whitney U test, Wilcoxon signed-ranks matched-pairs test, Kruskal-Wallis test, analysis of variance, and simple linear regression analysis were used. A two-tailed P value of < 0.05 was considered statistically significant. From the umbilical artery distal to the cotyledon artery, the PGI2 synthesis decreased and the TxA2 synthesis increased gradually towards the periphery in normal pregnancy. The PGI2/TxA2 ratio was more than 200 times higher in the umbilical artery than in the cotyledon artery. The TxA2 production tended in general to be higher in the diabetic group than in the control group, resulting in significantly lower PGI2/TxA2 ratios in some vessels. The prostanoid production was not significantly correlated to maternal HbA1c or cord C-peptide concentrations. The balance between vascular prostacyclin and thromboxane synthesis in the umbilicoplacental arterial tree changed gradually towards the periphery: the more peripheral, the lower the prostacyclin and the higher the thromboxane production. The physiological role of this phenomenon is unknown, but may be of importance for the equilibration of vascular tone between arteries of different calibers. The altered prostanoid balance found in diabetic pregnancy was not directly attributable to the degree of maternal glycemic control, but may reflect increased free radical activity and peroxide production in diabetes.

Cyclic strain is a weak inducer of prostacyclin synthase expression in bovine aortic endothelial cells

Recent studies indicate that hemodynamic forces such as cyclic strain and shear stress can increase prostacyclin (PGI2) secretion by endothelial cells (EC) but the effect of these forces on prostacyclin synthase (PGIS) gene expression remains unclear and is the focus of this study. Bovine aortic EC were seeded onto type I collagen coated flexible membranes and grown to confluence. The membranes and attached EC were subjected to 10% average strain at 60 cpm (0.5 sec deformation alternating with 0.5 sec relaxation) for up to 5 days. PGIS gene expression was determined by Northern blot analysis and protein level by Western blot analysis. The effect of cyclic strain on the PGIS promoter was determined by the transfection of a 1-kb human PGIS gene promoter construct coupled to a luciferase reporter gene into EC, followed by determination of luciferase activity. PGIS gene expression increased 1.7-fold in EC subjected to cyclic strain for 24 hr. Likewise, EC transfected with a pGL3B-PGIS (-1070/-10) construct showed an approximate 1.3-fold elevation in luciferase activity in EC subjected to cyclic strain for 3, 4, 8, and 12 hr. The weak stimulation of PGIS gene expression by cyclic strain was reflected in an inability to detect alterations in PGIS protein levels in EC subjected to cyclic strain for as long as 5 days. These data suggest that strain-induced stimulation of PGIS gene expression plays only a minor role in the ability of cyclic strain to stimulate PGI2 release in EC. These findings coupled with our earlier demonstration of a requisite addition of exogenous arachidonate in order to observe strain-induced PGI2 release, implicates a mechanism that more likely involves strain-induced stimulation of PGIS activity.

Recovery of endothelial cells and prostanoid production in endothelial cell-seeded grafts

OBJECTIVE

To investigate the function and morphology of endothelial cell (EC) seeded grafts.

DESIGN

Experimental, open study.

CHIEF OUTCOME MEASURES

Endoluminal release of prostacyclin (6-Keto-PGF1 alpha) and thromboxane B2 (TxB2), patency, EC coverage and cell identity.

MATERIALS

In 12 sheep, segments of both carotid arteries were excised. On one side a seeded and on the other an unseeded dacron graft were inserted. After 3 months the grafts were excised. In grafts and arteries, the endoluminal release of 6-keto-PGF1 alpha and TxB2 was determined in a perfusion system. Scanning electron microscopy (SEM) and light microscopy were used to determine the EC coverage and cell identity.

RESULTS

Eight animals survived. Three seeded and two unseeded grafts were occluded. Prostacyclin release did not differ significantly between seeded and unseeded grafts and arteries, when the arteries were looked upon as one group. When the graft was compared with its corresponding artery, i.e. the artery it replaced, a significantly lower release was found in the unseeded group. Thromboxane release was undetectable in arteries but significantly higher in both graft groups. SEM revealed a cellular coverage of 75% in the seeded grafts and 50% in the unseeded (not significant). Light microscopy showed a patchy staining for Factor VIII-related antigen in some grafts in both groups.

CONCLUSION

Prostacyclin release in unseeded and seeded dacron grafts did not differ 3 months after implantation in sheep, except when the graft was compared with its corresponding artery. The significance of this remains to be settled. Seeded grafts did not have a higher proportion of endothelial coverage than unseeded grafts.

Heterogeneity in prostacyclin and thromboxane synthesis in ovine pulmonary vascular tree: effect of age and oxygen tension

Intrapulmonary arteries and veins of 8 near-term fetal lambs (141-145 days gestation) and 8 ewes were isolated into segments of > 3mm, 1-3 mm, and < mm in diameter. Vessels were incubated in Krebs' buffer at 37 degrees C at PO2 approximately 100 torr (normoxia) and PO2 < 50 torr (hypoxia) to study local vascular production of prostanoids. Prostacyclin and thromboxane (Tx) A2 produced were measured by radioimmunoassay and expressed in ng/mg dry wt, means +/- SEM. During normoxia, fetal arteries > 3 mm synthesized more prostacyclin than adult arteries of the same size (1.71 +/- 0.3 vs 0.45 +/- 0.04). However, fetal arteries < 1 mm synthesized less prostacyclin than adult arteries < 1 mm (0.47 +/- 0.1 vs 1.75 +/- 0.16). Prostacyclin production by veins > 3 mm was similar in the fetus and adult (0.49 +/- 0.06 vs 0.67 +/- 0.08), but in veins < 1 mm was greater in adult than in fetal vessels (1.73 +/- 0.17 0.54 +/- 0.06). Hypoxia-attenuated prostacyclin production by fetal arteries and veins of all sizes, but only in 1 to 3-mm-size adult arteries. In general, production of TxA2 by segments of fetal and adult vessels was less than 50% of that of prostacyclin. Protein and DNA concentrations in similar sized fetal and adult vessels were similar. The data show that there is heterogeneity in the production of prostacyclin and TxA2 along the ovine pulmonary vascular tree. Prostanoid synthesis of fetal vessels is markedly influenced by hypoxia, with a greater suppression of prostacyclin synthesis. Similar protein and DNA concentrations in fetal and adult vessels suggest that differences in prostanoid production by vessel segments may be due to differences in enzyme activity rather than cell number or tissue mass.

Prostacyclin release from the human saphenous vein in diabetics in lower than in nondiabetics

The balance between prostacyclin and thromboxane has been suggested to be of great importance for the maintenance of patency in veins. In order to investigate prostacyclin and thromboxane release, segments from the human saphenous veins were investigated in 53 patients. Twenty-seven patients (10 males, 17 females) underwent surgery for varicose veins. Twenty-six patients (14 nondiabetics, 12 diabetics) underwent surgery for lower limb ischemia (rest pain or gangrene) with use of the saphenous vein as arterial conduit. Vein segments were gently excised and perfused ex vivo for five 15 minute periods, with a balanced salt solution and determination of the stable degradation products 6-keto-PGF1 alpha and TxB2. Saphenous veins from patients with varicose veins had an initial prostacyclin release of 61 +/- 13 pg/mm2/15 min declining to 4 +/- 1 pg/mm2/15 min after 60 min (p < 0.001) and increasing after addition to arachidonic acid to 37 +/- 7 pg/mm2/15 min (p < 0.001). Segments from nondiabetic patients with lower limb ischemia did not differ from those of varicectomy patients, but diabetic segments had a significantly lower prostacyclin release than both these groups, 34 +/- 11 pg/mm2/15 min, 1 +/- 1 pg/mm2/15 min, and 7 +/- 5 pg/mm2/15 min, respectively (p < 0.05). The addition of arachidonic acid failed to increase the prostacyclin release in diabetics. Three patients from each group were studied regarding thromboxane release and there was almost no detectable thromboxane in any group. These findings suggest that diabetics have a lowered prostacyclin release from the saphenous vein and that the deficiency is at the cyclo-oxygenase level.(ABSTRACT TRUNCATED AT 250 WORDS)

Pharmacological modulation of prostacyclin and thromboxane production of rat and cat venous tissue slices

To reveal a potential modulating effect of vasoactive pharmacological agents on the prostanoid production of the venous wall, prostacyclin and thromboxane release from venous tissue slices was studied. Aortic and caval vein samples from 20 rats as well as from 21 cats were studied. Prostacyclin and thromboxane productions were determined by radioimmunoassay as 6-keto-PGF1 alpha and TxB2 released into the incubation medium. Venous tissue produced significantly less prostacyclin per unit weight than arterial tissue in rats (30.7 +/- 4.6 vs. 52.1 +/- 8.2 pg/mg/min), while in cats an opposite situation was found (16.6 +/- 3.2 vs. 7.06 +/- 1.9 pg/mg/min). Thromboxane production of venous tissue was consequently higher than corresponding values for aortic tissue (3.72 +/- 0.46 vs. 1.54 +/- 0.14 in rats and 3.4 +/- 0.6 vs. 1.33 +/- 0.19 in cats, all values in pg/mg/min). Norepinephrine and dopamine significantly increased both the prostacyclin and the thromboxane release from venous tissue, while isoproterenol had no effect. Vasopressin significantly increased thromboxane release and decreased the ratio of prostacyclin vs. thromboxane production (from 10.4 +/- 1.6 to 7.5 +/- 1.6, in acetylsalicylic acid pretreated cats). Angiotensin and thrombin had no significant effects. Bradykinin (0.5 microgram/ml) significantly augmented prostacyclin release from venous tissue (14.4 +/- 2.6 from 10.9 +/- 2.4 pg/mg/min) and decreased thromboxane release (0.65 +/- 0.18 from 1.35 +/- 0.22 pg/mg/min). Methionine-enkephalin (5 micrograms/ml) significantly reduced the thromboxane release from venous tissue slices. The presented material demonstrates that several vasoactive agents modulate the vasoactive prostanoid release of the venous wall. In some cases, the prostacyclin and the thromboxane productions are influenced separately, which in turn will have its impact on smooth muscle activity and thrombocyte aggregation.

Prostacyclin is produced from endothelial cell-seeded grafts: an experimental study in sheep

Endothelial cell seeding might be of value in reducing the thrombogenicity of small-diameter vascular grafts. We investigated the capacity of endothelial cell-seeded grafts to produce prostacyclin and compared this with that of the unseeded graft as well as the native artery. Twelve sheep were operated on with carotid interposition of externally supported knitted dacron grafts. On one side of the neck the graft was seeded with endothelial cells, enzymatically harvested from the left jugular vein. After 3 weeks, three out of 12 seeded grafts, and one out of 12 unseeded grafts were occluded (N.S.). After excision, the grafts were mounted and perfused ex vivo for five 15-min periods. During the last period, arachidonic acid (4 micrograms/ml) was added to the perfusate. The resected carotid artery was used as a control. Prostacyclin was determined as the stable degradation product 6-keto-PGF1 alpha using radio-immunoassay, and expressed as pg mm-2 luminal surface. The native artery had a significantly higher release of prostacyclin than the seeded graft, which in turn had a significantly higher release than the unseeded graft. Histological examination showed weakly positive staining for factor VIII-related antigen on the luminal surface of seeded grafts. Scanning electron microscopy showed endothelial cells with typical endothelial tufts and was evaluated blindly from 10 areas of each graft. The extent of endothelial cell coverage was evaluated and scored from 0 to 2.5. The median score for the unseeded grafts was 0.3 and for the seeded grafts 1.5 (p = 0.008). Prostacyclin production was higher in seeded than unseeded grafts, but did not influence patency in this model.

Storage in sodium chloride does not impair arterial prostacyclin release

Heparinised sodium chloride solution is often used for the storage of veins and to prevent clotting in the arterial tree during vascular surgery. Sodium chloride deranges the morphology of endothelial cells and has therefore been interpreted as "toxic" to the endothelial cell. Perfused human saphenous veins and rabbit aortas show the same pattern of prostanoid release even though veins have a lower release than arteries. Excised rabbit aortas were stored in either (a) 0.9% sodium chloride or (b) 0.9% sodium chloride with heparin 5 I.U. ml-1 prior to being mounted in a perfusion model. The vessels were perfused ex vivo for 5 x 15 min with either Hanks' balanced salt solution (HBSS), calcium- and magnesium-free HBSS or 0.9% sodium chloride. For the last period, arachidonic acid (AA) 4 micrograms ml-1 was added. The release of prostacyclin, measured as the stable degradation product 6-keto-PGF1 alpha by radioimmunoassay, was not altered by storage in sodium chloride, or 0.9% sodium chloride plus heparin when compared with control segments. Perfusion with 0.9% sodium chloride did, however, significantly (p greater than 0.05) decrease the prostacyclin production when AA was added. This is most likely to be due to the low pH of the sodium chloride solution. It is concluded that short-term storage of rabbit aorta in sodium chloride plus heparin or 0.9% sodium chloride does not impair the prostacyclin cascade from the vessel wall, which might be of importance when choosing storage medium for reversed veins and veins used for coronary bypass surgery.

Thromboxane and prostacyclin production in the perfused rabbit lung

We investigated whether prostacyclin formation by the isolated rabbit lung can serve as a measure of pulmonary distress. The basal TXA2 and PGI2 formation was very low, and depended on the preperfusion history of the lung (low or high flow, use of dextran or artificial perfusate). The basal prostanoid production remained unchanged over a time period of 2 h. Neither was it influenced by the serotonin uptake inhibitor chlorimipramine and by small changes in temperature (33 degrees C vs 39 degrees C). The PGI2 formation was almost independent of hemodynamic alterations such as embolism or vasoconstriction. An enhanced production was only seen after a dramatic increase in flow (from 1.7-5 ml/sec), and a transient 3-fold increase was observed after administration of 1 mM H2O2. A substantial (up to 40-fold) but transient increase in TXA2 production was measured after 1 mM of H2O2, and the TXA2 production was positively correlated to the increase in pulmonary arterial pressure. However, thromboxane production was also dramatically augmented by hemodynamic alterations such as embolism, increased flow and--to a lesser extent--vasoconstriction. We conclude that the determination of the prostanoid production (and particularly the TXA2 formation) by the rabbit lung cannot be used as a direct measure of endothelial distress. To this end it is excessively biased by hemodynamic alterations such as recruitment and shear stress.

Prostacyclin production in the area of arterial endothelial denudation

The goal of this study was to quantify arterial prostacyclin (PGI2) synthesis and platelet aggregation in the immediate area of vessel injury. Twelve mongrel dogs whose platelets aggregated maximally to added arachidonic acid (AA) were equally divided into three groups. Controls received no drug while other dogs were treated with the platelet inhibitors aspirin (ASA) 3 mg/kg/day or BM 13.505 (BM), a thromboxane receptor antagonist, 25 mg/kg/day. After 3 days of treatment, the dogs underwent balloon catheter endothelial denudation of both carotid and femoral arteries. Blood was sampled from the first carotid artery just distal to the injury at 0, 1, 5, and 10 min after restoration of flow. Venous blood samples were also obtained at 1 and 2 weeks postoperatively. Dogs were sacrificed at 2 weeks and arterial rings from a proximal normal and a denuded region were excised and tested for PGI2 production in response to added AA. While control dogs showed no statistical change in AA-induced platelet aggregation postoperatively, there appeared to be a trend for enhanced responsiveness at 1 week. Aspirin and BM inhibited AA-induced aggregation completely in all samples at all timepoints. Levels of the stable metabolite of PGI2, 6-keto-PGF1 alpha, were determined by radioimmunoassay. In control dogs, baseline PGI2 levels were 72 +/- 17 pg/ml which increased to 479 +/- 20 pg/ml immediately after restoration of flow (P less than 0.001, Student t test) and returned to basal values in 10 min (85 +/- 5 pg/ml).(ABSTRACT TRUNCATED AT 250 WORDS)

Influence of transluminal angioplasty on the prostanoid release from the arterial wall

Vasospasm and thrombosis complicate percutaneous transluminal angioplasty (PTA). To study if the release of the prostanoids PGI2 and TxA2 are affected by PTA, the following experiment was undertaken: In ten rabbits, the upper or lower half of the aorta was randomised either to transluminal angioplasty or control segment. After excision the segments were simultaneously but separately perfused ex vivo with Hank's balanced salt solution for five consecutive 15 min periods. Arachidonic acid was added to the perfusate for the last 15 min period. PGI2 and TxA2 were measured by radioimmunoassay in the perfusate as the stable degradation products 6-keto-PGF1 alpha and TxB2. After perfusion, the two aortic segments were prepared for scanning electron microscopy (SEM). Angioplasty decreased the basic release of PGI2 as well as the response to arachidonic acid. This is likely to be due to endothelial denudation as seen by SEM. The release of TxA2 from the vessel wall was very low and was not increased by dilatation. The influence of angioplasty on the prostanoid system may be of importance in the complications of vasospasm and thrombosis.

Arterial 6-keto-PGF1 alpha and TxB2 release in ex vivo perfused canine vessels: effects of pulserate, pulsatility, altered pressure and flow rate

Certain experimental conditions are known to influence the release of prostacyclin and thromboxane from the vessel wall. The specific effects of altered pulsatility, pressure, and flow rate on intraluminal release of 6-keto-PGF1 alpha and thromboxane B2 were assessed in canine arteries perfused ex vivo for five 15 min periods with arachidonic acid (AA) added during the last period. Control arteries were perfused at 100 mmHg with pulsatile flow of 90 ml/min. Experimental arteries were perfused at 7, 50 and 200 mmHg with pulsatile flow of 90 ml/min, and at 100 mmHg pressure with pulsatile flow of 20, 60, 130 and 180 ml/min, as well as at 100 mmHg with 90 ml/min nonpulsatile flow. Perfusion pump rates of 44 and 96 beats/min were also assessed. The lowest perfusion pressure, 7 mmHg, resulted in a lesser initial release of prostacyclin compared to higher pressures, and there was a tendency to a higher release of prostacyclin with increasing pressures. There was also a tendency for a lesser response to AA in arteries perfused at 200 mmHg, perhaps due to endothelial cell damage. Nonpulsatile flow was associated with a decreased initial release of prostacyclin, and diminished release following addition of AA when compared to pulsatile flow. Altered flow rate elicited no difference in prostacyclin release, although there was a tendency towards a lesser release when perfused at 20 ml/min compared to 130 ml/min or 180 ml/min. Thromboxane release was decreased by nonpulsatile flow but was otherwise unaffected by the experimental conditions tested. It is concluded that pulsatility enhances release of prostacyclin from arteries.(ABSTRACT TRUNCATED AT 250 WORDS)