Replication of human endothelial cells in culture.

Investigative studies dealing with the properties and functions of endothelial cells have been hampered because there has been little or no success in the isolation, growth, and passage of individual cells in large numbers. We have developed a system whereby pure cultures of endothelial cells derived from umbilical veins can be subcultured for at least five serial passages. Many facets of endothelial function and interaction can be evaluated with the use of this new adaptive system of isolation and culture.

Prostacyclin biosynthesis in cultured vascular endothelium is limited by deactivation of cyclooxygenase.

Primary monolayer cultures of human umbilical vein endothelium produce prostacyclin (PGI2) in response to stimulation by thrombin, ionophore A23187, arachidonic acid, and the prostaglandin endoperoxide, PGH2. None of these treatments had a significant effect on the capacity of the endothelium to produce PGI2 in response to subsequent stimulation by PGH2. By contrast, endothelium initially exposed to thrombin, A23187, or arachidonic acid produced approximately 37, 68, and 84% less PGI2, respectively, upon subsequent stimulation by arachidonic acid. These findings suggest that PGI2 biosynthesis in cultured endothelium results in deactivation of cyclooxygenase-hydroperoxidase but not PGI2 synthetase. To test the hypothesis that PGI2 biosynthesis alone causes deactivation of cyclooxygenase, thrombin, A23187, and arachidonic acid were added to monolayers that had been preincubated with ibuprofen (250 microM), a rapidly reversible, competitive inhibitor of this enzyme. After removal of the ibuprofen and the initial stimulus, PGI2 production in response to subsequent stimulation by arachidonic acid was maximal. These findings suggest that the metabolism of arachidonic acid itself causes a direct deactivation of cyclooxygenase. After an initial exposure to arachidonic acid, PGI2 production in response to a second stimulation by arachidonic acid was restored to approximately 34, 69, and 74% of maximal, after recovery periods of 1, 24, and 48 h, respectively. We conclude that the regulation of PGI2 biosynthesis in normal vascular endothelium may be in part a function of the activity and biosynthesis of cyclooxygenase-hydroperoxidase and the deactivation of this enzyme may be a primary factor limiting the capacity of the endothelium to produce PGI2.

Effect of aspirin on thrombin-induced adherence of platelets to cultured cells from the blood vessel wall.

An in vitro method was used to detect adherence of (51)Cr-labeled platelets to monolayers of cultured human endothelial, fibroblast, and smooth muscle cells. Washed platelets did not adhere to untreated or aspirin-treated endothelial monolayers in the absence of thrombin. In contrast, thrombin-induced platelet aggregates adhered to all of the monolayers but adherence to endothelium was significantly less than to the other cells. Additional evidence for adherence of platelets to the endothelium was provided by scanning and transmission electron microscopy. Thrombin-induced platelet adherence to endothelium was inhibited by hirudin. Platelet adherence induced by thrombin was enhanced significantly by treatment of the endothelial monolayer with 1-2 mM aspirin. This increase in adherence was seen even when aspirin-treated platelets were used; adherence values approached those seen with fibroblasts and smooth muscle cells. An aspirin concentration of 0.1 mM was sufficient to block thrombin-induced malonaldehyde production in platelets but it did not interfere with the inhibitory effect of the endothelium against platelet adherence. The effect of aspirin on the endothelium was temporary and inhibitory activity of the endothelium was restored 1 h after aspirin had been removed from the incubation system. The ability of thrombin to cause adherence of platelets to undamaged endothelium, and the potential for aspirin to enhance this adherence have implications for mechanisms which operate in platelet interaction with the blood vessel wall.

Toxic effects of glucagon-induced acute lipid mobilization in geese.

The toxic effects associated with rapid lipid mobilization and a high plasma free fatty acid (FFA) concentration produced by glucagon were evaluated. Glucagon (0.5 mg/kg of body wt) was injected intravenously into nonfasting geese. The geese developed rapid respirations and high plasma FFA levels within 15 min after the glucagon injection; three of eleven died. Control geese, injected with saline, did not exhibit toxic signs. Peak FFA concentrations developed 15 min after glucagon and high levels persisted for over 90 min. Geese injected with glucagon frequently developed electrocardiographic abnormalities that included supraventricular tachycardia, premature ventricular contractions, and signs of myocardial ischemia. Light and electron microscopy revealed acute myocardial degeneration and fatty infiltration of the liver. The increase in plasma FFA concentrations and toxic effects were not prevented by pretreatment with nicotinic acid or propranolol.

Utilization of arachidonic and linoleic acids by cultured human endothelial cells.

When cultured human umbilical vein endothelial cells are supplemented with linoleic acid, the arachidonic acid content of the cellular phospholipids is reduced approximately 35%. Most of the fatty acid compositional change occurs during the first 24 h. One factor responsible for this effect is the inability of the endothelial cells to convert appreciable amounts of linoleic to arachidonic acid, due to a fatty acid delta 6-desaturase deficiency. By contrast, these endothelial cultures contain delta 5- and delta 9-desaturase activity and are able to elongate long-chain polyunsaturated fatty acids. The other factor that contributes to the decrease in arachidonic acid is that high concentrations of linoleic acid reduce the incorporation of arachidonate into cellular phospholipids. Stearic acid, a long-chain saturate, does not produce any reduction, whereas eicosatrienoic acid is an even more effective inhibitor than linoleic acid. In spite of the fact that high concentrations of these polyunsaturates produced inhibition, the endothelial cells were found to efficiently incorporate exogenous arachidonic acid into cellular phospholipids and triglycerides. This may serve to compensate for the inability of these cells to synthesize arachidonic acid from linoleic acid. These findings suggest that the endothelium obtains arachidonic acid from an extracellular source, that this cannot be provided in the form of linoleic acid and, in fact, that high concentrations of linoleic acid actually may interfere with the ability of the endothelium to maintain an adequate supply of intracellular arachidonic acid.

Utilization of long-chain free fatty acids by human platelets.

There was a rapid net uptake of free fatty acid (FFA) by human platelets when long-chain FFA, bound to human serum albumin, were incubated with platelet suspensions. Results from experiments in which both palmitate and albumin were labeled indicated that the fatty acid dissociated from the protein during uptake. Much of the FFA taken up by the platelet in short-term incubations remained in unesterified form, i.e., it was recovered as platelet FFA. As the incubation continued, increasing amounts of FFA were oxidized to CO(2) and incorporated into platelet lipid esters, particularly lecithin. Essentially all of the fatty acid that was incorporated into the platelet FFA fraction was released rapidly from the cells when they were exposed to a medium containing FFA-free albumin. The magnitude of uptake into the platelet FFA fraction was similiar at 0 degrees and 37 degrees C. Likewise, the rate and magnitude of FFA release from the platelet were similar at 0 degrees and 37 degrees C. Therefore, it is likely that both FFA uptake and FFA release occur by energy-independent mechanisms. The major effect of increasing the FFA concentration of the incubation medium was increased fatty acid uptake into the platelet FFA fraction. Similar results occurred when platelets were incubated in human plasma containing increasing amounts of added palmitate. At a given extracellular FFA concentration, considerably more of the saturated fatty acids, palmitate and stearate, were taken up as platelet FFA than either oleate or linoleate.

Effect of fatty acid modification on prostacyclin production by cultured human endothelial cells.

We have investigated whether changes in cellular fatty acid saturation can influence prostacyclin (PGI2) production by cultured human umbilical vein endothelial cells. As compared to control cells, those enriched with linoleic acid released 60--75% less PGI2 in response to thrombin or the calcium ionophore A23187. A similar but considerably smaller effect was observed when the cells were enriched with oleic or linolenic acid, but no reduction occurred with palmitic or linoelaidic acids. Some reduction in PGI2 release was noted as early as 1 h after exposure to linoleic acid. When the culture medium was supplemented with linoleic acid, the cell phospholipids contained four to five times more linoleate and 25--40% less arachidonate. These changes were most marked in the choline and serine plus inositol phosphoglyceride fractions. When the fatty acid composition of the cells enriched with linoleic acid was allowed to revert, there was a progressive increase in the capacity of the cells to release PGI2 in response to thrombin. The increase correlated with a reduction in linoleate content of the cell lipids, but there was no change in arachidonate content. This suggests that linoleic acid may act as an inhibitor of PGI2 production. The cultured endothelial cells were also able to produce PGI2 directly from added arachidonic acid. As the arachidonic acid concentration of the medium was raised, PGI2 formation by the linoleate-enriched cells increased relative to control cells, suggesting that the inhibition produced by linoleic acid may be competitive.

Inhibition of prostacyclin by treatment of endothelium with aspirin. Correlation with platelet adherence.

Aspirin treatment of cultured endothelial cells from the umbilical vein increased the adherence of 51Cr-platelets when thrombin was present. If the cyclooxygenase activity of endothelium was inhibited by aspirin, as it is in the platelet, reduction of endogenous prostacyclin (PGI2) production could have been responsible. By correlating thrombin-induced adherence of platelets to endothelial monolayers with PGI2 release (as measured by radioimmunoassay for 6-keto-prostaglandin FI1 alpha [6-keto-PGF1 alpha]), we have demonstrated an inverse relationship between platelet adherence and PGI2 levels. Untreated endothelial monolayers exposed to thrombin and platelets resulted in 4% platelet adherence and 107 nM 6-keto-PGF1 alpha. With 0.1 mM aspirin treatment, which is known to block platelet cyclooxygenase, adherence was 5% and 6-keto-PGF1 alpha decreased to 45 nM. Increasing the aspirin concentration to 1 mM resulted in 44% adherence and less than 3 nM 6-keto-PGF1 alpha. When 25 nM exogenous PGI2 was added to 1 mM aspirin-treated endothelium, adherence returned to 5%. The increase in thrombin-induced platelet adherence to 1 mM aspirin-treated monolayers was reversed 2 h after removal of the aspirin solution. 6-Keto-PGF1 alpha returned to 37% of the untreated monolayer value. Recovery from the aspirin effect did not occur when cycloheximide, an inhibitor of protein synthesis, was present during the 2-h period.

Effects of triiodothyronine-induced hypermetabolism on factor 8 and fibrinogen in man.

Triiodothyronine (liothyronine sodium) (400-500 mug/day for 14 days) was given to six normal subjects. Factor VIII (antihemophilic globulin) activity increased from 109 to 167% (P < 0.05); fibrinogen increased from 344 to 581 mg/100 ml (P < 0.01). To test whether the increases in factor VIII activity and fibrinogen were mediated by beta adrenergic receptors, propranolol (20 mg every 6 hr) was given orally to four other normal subjects in addition to triiodothyronine for 14 days. Factor VIII increased from 100 to 161%; fibrinogen increased from 374 to 564% (P < 0.01). Factor VIII activity did not change in a severe classical hemophiliac made hypermetabolic with triiodothyronine, but it increased from 39 to 82% in a patient with von Willebrand's disease. Triiodothyronine-induced hypermetabolism increased the incorporation of selenomethionine-(75)Se into plasma fibrinogen. These results suggest that the increases in clotting factor activity during triiodothyronine-induced hypermetabolism reflect an effect of increased protein synthesis rather than enhanced stimulation of beta adrenergic receptors.

Heterogeneity of antibody response to human platelet transfusion.

To study the antibody response to human platelet transfusions, nine thrombocytopenia patients with bone marrow failure were given 6 U (3X10(11)) of random platelet concentrates twice a week. Before transfusion, none of the patients had preexisting antibodies detectable with lymphocytotoxicity, platelet aggregation, or capillary leukoagglutination techniques. After receiving 18-78 U of platelets, they became refractory to further transfusions of random platelets and alloantibodies were detectable. Two patterns of antibody response could be identified. In three patients, the sera were not lymphocytotoxic with a panel of standard cells in which all the known HLA antigens in the first and second series were represented at least once. Yet, they caused platelet aggregation with 30, 24, and 60%, respectively, of a donor population studied. The aggregating activities were inhibited by antihuman IgG but not by antihuman IgA or antihuman IgM antiserum. The aggregating antibodies could be absorbed out with donor platelets but not lymphocytes or granulocytes. Antibodies from two of these patients aggregated platelets of their respective siblings matched for both HLA haplotypes. Transfusion of platelets from these two siblings did not increase the platelet count while platelets obtained from aggregation-negative donors did. The sera from the remaining six patients were lymphocytotoxic with 15-100% of the panel of standard cells. They also had aggregating antibodies, which could be absorbed out by both platelets and lymphocytes, suggesting that they were HLA antibodies. These data suggest that the development of platelet-specific antibodies may play an important role in the immunological rejection of isologous platelets, and should be considered in the selection of donors for patients who are refractory to platelets from random donors.

Differential binding of insulin to human arterial and venous endothelial cells in primary culture.

The properties of 125I-insulin binding were assessed in endothelial cells prepared from the veins and the arteries of human umbilical cords. The endothelial nature of both the natural and venous cultures were documented by the presence of characteristic endothelial features, including Weibel-Palade bodies, factor VIII antigen, and morphology. Both arterial and venous cells possessed typical receptors for insulin on the basis of specificity of binding, curvilinear Scatchard plots, affinity profiles, pH dependency, and dissociation kinetics. Arterial cells bound at least 2.5 times more insulin than did venous cells, whether studied at 4 h, 24 h, or 72 h after in vitro plating. We conclude that (1) specific receptors for insulin are present on human arterial as well as human venous endothelial cells and (2) the concentration of insulin receptors varies among endothelial cells derived from different vascular sources.

Insulin receptors in human endothelial cells: identification and characterization.

Specific binding of 125I-insulin was found in cultured human endothelial cells obtained from human umbilical veins. The binding reaction was rapid and reversible, demonstrated receptor site-site interactions of the negatively cooperative type, and was dependent on the temperature, pH and duration of incubation. At 21 degrees C, steady-state conditions of binding occurred in 90 minutes, the pH optimum was 7.8 and less than 10% of the labeled hormone was degraded. Binding of tracer amounts of 125I-insulin was inhibited by concentrations of unlabeled insulin as low as 0.2 ng/ml and 50% inhibition was obtained at 2-5 ng/ml of unlabeled insulin. Unlabeled porcine insulin, porcine proinsulin and desoctapeptide insulin inhibited the binding of 125I-porcine insulin in direct proportion to their biological potencies, whereas to the cells was inhibited by antibodies against insulin receptors. We conclude that human endothelial cells possess specific receptors for insulin whose physio-chemical properties are similar to those of insulin receptors in other tissues.

Platelet-endothelial cell interactions.

We are in an era in which there has been a rapid accumulation of new information about the endothelium and the interactions of platelets with the vessel wall. The decade of the 70s could aptly be designated as the "prostaglandin and endothelium period." One can only speculate that the future will be equally interesting and challenging. It is quite possible that discoveries of congenital and acquired defects of the endothelium will parallel those already described for the platelet and coagulation factors. What is currently looked upon today as a baffling and frustrating case of resistant thrombosis by the clinician may be unraveled tomorrow by the talented and creative investigator who can develop better approaches to the study of the endothelium and platelet interactions with the vessel wall.

Stearic acid, clotting, and thrombosis.

Stearic acid causes hypercoagulability of the blood by activation of factor XII and by aggregation of blood platelets. Injection of unbound stearic acid (sodium salt) into the systemic circulation of dogs was followed by massive generalized thrombosis and sudden death. Similar infusions into birds, which are deficient in factor XII, did not cause hypercoagulability or thrombosis. The effects of the long-chain saturated fatty acids could be prevented by using albumin to bind the stearic acid at a molar ratio of free fatty acid (FFA) to albumin of < 2. The major issue is whether eating foods rich in stearic acid can cause a thrombogenic effect. We have no experimental evidence to support this concept. If a thrombogenic effect of long-chain saturated fatty acids exists in humans, it is most likely to occur as an aberration of fatty acid transport in which the FFA-albumin molar ratio exceeds 2 either as a result of very high plasma FFA concentrations from lipid mobilization or a low concentration of albumin in the blood as found in disease states such as the nephrotic syndrome.

Fatty acids in animals: thrombosis and hemostasis.

Long-chain fatty acids, including stearic acid, can cause thrombosis when they are injected into the systemic circulation of animals. The toxicity is decreased if the fatty acids are bound to albumin. To date, the effect of individual fatty acid feeding is not known. In general, feeding animals diets high in saturated fat followed by the injection of a thrombogenic stimulus is associated with a greater incidence of thrombosis than when a normal diet or a diet high in unsaturated or polyunsaturated fat is fed. A high dietary intake of long-chain polyunsaturated fatty acids of the n-3 family may prevent thrombosis in animals. In assessing the effects of dietary fatty acids, the ability of the organism, be it animal or human, to convert fatty acids to less thrombogenic fatty acids (stearic to oleic) or create an antithrombotic fatty acid (linolenic to eicosapentaenoic) is a major attribute. Further studies should consider the storage of fatty acids in tissues and their release into the blood and the potential impairment of albumin binding and fatty acid transport mechanisms.

Arterial platelet accumulation in experimental hypercholesterolemia.

Accumulation of arterial platelets was calculated from 51Cr radioactivity in the intima-inner media of flushed, perfuse-fixed aortas and branch vessels of cynomolgus monkeys 48 hours after labeling of the blood platelets. In normo-cholesterolemic controls the radioactivity per square centimeter of tissue was consistently higher in aortic branching sites (circumostial aorta) than in the remainder of the aorta. In monkeys given 10 and 100 days of hypercholesterolemic diet radioactivity rose in circumostial aorta in proportion to increases in streaks and in areas of Evans blue uptake. In branch artery inner wall, where intimal changes were minimal and usually absent, counts were significantly greater in animals given 100 days of hypercholesterolemic diet than in controls. After 10 days of hypercholesterolemic diet the mean radioactivity in aortic branch artery inner wall uniformly exceeded control values, but usually nonsignificantly, permitting the speculation that changes in platelet-intima interactions may occur in widespread fashion throughout the arterial tree very early in hypercholesterolemia when lesion formation is incipient or absent.

Spontaneous platelet aggregation in arterial insufficiency: mechanisms and implications.

To investigate the clinical implications and mechanisms of spontaneous platelet aggregation (SPA) in man, 150 normal subjects, 22 patient controls and 130 patients with vascular insufficiency were studied. SPA was negative in normal subjects and patient controls whereas it was positive in 36 of 66 (54%) patients with transient ischemic attacks, 6 of 32 (19%) patients with stable angina, 7 of 10 (70%) patients with acute myocardial infarction and 11 of 14 (80%) patients with acute peripheral arterial insufficiency. The SPA was inhibited with aspirin in vivo, and inhibited competitively in vitro by low concentrations of aspirin, 2-chloroadenosine, prostaglandin E1 or apyrase but only by high concentrations of heparin or hirudin. Addition of platelet-poor plasma from patients with positive SPA did not cause normal platelets to aggregate. Treatment of patients who had acute peripheral arterial insufficiency with aspirin and dipyridamole prevented SPA with notable clinical improvement of the ischemic changes.

The inhibitory effect of aspirin on human endothelial cells.

Human endothelial cell monolayers prepared from umbilical veins have been incubated with aspirin (1--2 mM) dissolved in Hepes modified solution and in platelet-rich plasma. They have also been incubated with plasma prepared from subjects before and after intake of aspirin giving a mean plasma concentration of 0.5 mM. The effects of the endothelial cells on ADP and collagen-induced platelet aggregation and malondialdehyde production in platelet-rich plasma have been tested. The endothelial cells had a spontaneous inhibitory effect on all three parameters. This effect was abolished when the cells were incubated with aspirin dissolved in MHS for 20 min and the increase in effect observed when platelet-rich plasma was incubated with endothelial cells for a period of 30 min was similarly inhibited when aspirin was dissolved in plasma or when plasma prepared from subjects who had taken aspirin were used. Aspirin had no inhibitory effect on prostacyclin (PGI2) with regard to the effect of PGI2 on platelets. On the contrary, the two compounds had an additive inhibitory effect on platelet aggregation induced by ADP and collagen. These findings should be considered with regard to the use of aspirin as an antithrombotic agent.

Ocular neovascularization. Tissue culture studies.

The proliferative activity of a number of intraocular fluids, bovine retinal extract, and normal serum (from humans and cynomolgus monkeys) was investigated by in vitro tissue culture studies, with the use of tritiated thymidine incorporation by the cultured endothelial cells of human umbilical veins. There was increased tritiated thymidine incorporation by (1) the aqueous, vitreous, and intraocular fluid (IOF) (which filled the eye after lensectomy and vitrectomy) removed from cynomolgus monkey eyes with iris neovascularization or with neovascular glaucoma (NVG) that developed after experimental retinal vein occlusion, (2) by aqueous and vitreous removed from human eyes with NVG or proliferative diabetic retinopathy; (3) by the serum, and (4) by the bovine retinal extract. However, tritiated thymidine incorporation was not increased by the normal aqueous, vitreous, or IOF.

Detection of deep vein thrombosis with an automatic electrically calibrated strain gauge plethysmograph.

A new strain gauge plethysmography was developed to permit accurate electrical calibration and automatic calculation of limb blood flow from a panel meter. This instrument was used to quantitate maximum venous outflow (MVO) in both limbs of 20 normal volunteers and 387 limbs of 195 patients with clinically suspected leg vein thrombosis. The MVO of normal subjects averaged 45 +/- 18 cc/minute/100 cc of calf tissue (mean +/- 1 SD). In 69 limbs with deep vein thrombosis documented by Doppler ultrasound and/or phlebography, the MVO averaged 17 +/- 13 cc/minute/100 cc, which was significantly less than that of normal limbs (p less than 0.001). Only three of 31 limbs with venous thrombosis above the knee had MVO above 20 cc/minute/100 cc, the lower limit of normal, for a diagnostic sensitivity of 90%. The specificity of a normal MVO in excluding deep vein thrombosis was 81%, with only a 1% rate of false-negative diagnoses above the knee. This automatic strain gauge plethysmograph has the attribute of providing rapid quantitation of limb venous hemodynamics which facilitates noninvasive diagnosis and follow-up assessment of venous disease.