Distribution of lecithin-retinol acyltransferase activity in different types of rat liver cells and subcellular fractions

It is now well documented that lecithin-retinol acyltransferase (LRAT) is the physiologically important enzyme activity involved in the esterification of retinol in the liver. However, no information regarding the cellular distribution of this enzyme in the liver is presently available. This study characterizes the distribution of LRAT activity in the different types of rat liver cells. Purified preparations of isolated parenchymal, fat-storing, and Kupffer + endothelial cells were isolated from rat livers and the LRAT activity present in microsomes prepared from each of these cell fractions was determined. The fat-storing cells were found to contain the highest level of LRAT specific activity (383 +/- 54 pmol retinyl ester formed min-1.mg-1 versus 163 +/- 22 pmol retinyl ester formed min-1.mg-1 for whole liver microsomes). The level of LRAT specific activity in parenchymal cell microsomes (158 +/- 53 pmol retinyl ester formed min-1.mg-1) was very similar to LRAT levels in whole liver microsomes. The Kuppfer + endothelial cell microsome fractions were found to contain LRAT, at low levels of activity. These results indicate that the fat-storing cells are very enriched in LRAT but the parenchymal cells also posses significant levels of LRAT activity.

Characterization of liver lysosomal enzyme activity in hepatocytes, Kupffer and endothelial cells during aging: effect of dietary restriction

The specific activity of 4 lysosomal enzymes was studied in homogenate, hepatocytes, Kupffer and endothelial cells isolated from the livers of female Sprague-Dawley rats aged 3.5, 12 and 24 months. Cells were obtained by enzymatic digestion and centrifugal elutriation. Cell viability was not affected by age or diet. In hepatocytes, the activities of all enzymes (acid phosphatase, beta-galactosidase, arylsulfatase B and cathepsin D) increased with age in rats fed ad libitum (A) but were not altered significantly by dietary restriction. The activities of all enzymes except acid phosphatase were systematically higher at 3.5 months of age in Kupffer and endothelial cells than in hepatocytes. Acid phosphatase, arylsulfatase B and cathepsin D activities increased with age in both Kupffer and endothelial cells. Beta galactosidase was decreased significantly with age in Kupffer cells but was elevated in endothelial cells. Rats exposed to dietary restriction (R) showed higher activities of beta-galactosidase, arylsulfatase B and cathepsin D when compared to corresponding A animals with the exception of the younger age group. No clear cut pattern was observed in acid phosphatase activity. Thus, the activities of liver lysosomal enzymes increase with age but the pattern of change differs with respect to enzyme and cell populations. The heightened enzyme activity in Kupffer and endothelial cells from R rats may reflect a more efficient phagocytic capacity in these animals.

The embryonic environment strongly attenuates v-src oncogenesis in mesenchymal and epithelial tissues, but not in endothelia

We demonstrate that the behavior of cells expressing v-src, a tyrosine kinase oncogene, differs profoundly between the embryonic and culture environments. V-src was introduced into avian embryo cells both in culture and in stage-24 embryo limbs, using replication-defective retroviral vectors. These vectors were used as single-hit, cellular markers to determine the environmental influences imposed by normal cells and tissues on clonal cell growth. The marker gene lacZ was coexpressed with v-src in order to locate the descendent cells. In culture, v-src induced rapid morphological transformation and anchorage-independent growth of embryo fibroblasts; the vectors were also tumorigenic in hatchling chickens. In contrast, most of the cell clones expressing v-src in the embryo grew normally without neoplasia. Expression of v-src vectors could be found in a wide range of cell types, demonstrating not only that neoplastic transformation is attenuated in ovo, but also that differentiation commitment in many lineages can be maintained concurrently with oncogene expression. Significantly, the embryonic control of cell growth could be perturbed by v-src under certain conditions. Rare, marked clones showed hyperplasia or dysplasia, and the primitive endothelium could succumb to rapid neoplasia; thus, these embryonic tissues are not inherently deficient in transformation factors. We propose that the environmental conditions imposed on cells in ovo are critical for the attenuation of neoplasia, while cultured cells lose this requisite environment.

Mechanism of increased angiotensin-converting enzyme activity stimulated by platelet-activating factor

We studied the effects of platelet activating factor (PAF) on angiotensin-converting enzyme (ACE). PAF (1 x 10(-10) to 1 x 10(-6) M) had a novel effect on angiotensin I conversion. Pulmonary artery endothelial cells converted 1 nmol/dish of 125I-angiotensin I to angiotensin II in the absence of PAF. ACE activity was increased to 2.5 nmol/dish by the addition of 1 x 10(-6) M of PAF. To clarify the mechanism of this stimulatory effect of PAF on ACE, Ca2+ influx and inositol 1,4,5-trisphosphate (IP3) release in pulmonary artery endothelial cells were determined. PAF stimulated Ca2+ influx in a dose-dependent manner. PAF also stimulated phospholipase C (PLC) activity and released IP3. To study the relationship between PLC activity and ACE activity, neomycin was added. The Ca2+ influx and IP3 release stimulated by 10(-6) M of PAF were suppressed by about 60-70%. ACE activity was also inhibited up to 70% in the presence of PAF (10(-10) - 10(-6) M) by 50 M of neomycin. These results suggest that ACE was stimulated by PAF, and that its activity in endothelial cells may be mediated by the PI-turnover pathway via changes in PLC activity and IP3-mediated Ca2+ release from intracellular stores.

Gamma-glutamyl-transpeptidase (GGTP) and NA+K(+)-ATPase activities in different subpopulations of cloned cerebral endothelial cells: responses to glial stimulation

Glial stimulation of Na+K(+)-ATPase and gamma-glutamyl-transpeptidase was taken as parameter for blood brain barrier function in cloned cerebral endothelial cells of different phenotypes. In type I cells ("cobblestone" phenotypus) gamma-glutamyl-transpeptidase activity increased 10-12 fold and Na+K(+)-ATPase activity was 2-fold increased after glial stimulation. In type II cells ("spindle-form" phenotype) gamma-glutamyl-transpeptidase was only 2-fold increased, whereas Na+K(+)-ATPase was even depressed. K(+)-(86Rb) uptake was twice as high in type I cells. These data indicate that type I cells are involved in blood brain barrier function.

The T cell triggering molecule Tp103 is associated with dipeptidyl aminopeptidase IV activity

Tp103 is a 103-kDa T cell activation molecule that defines an alternative activation signal for human T lymphocytes. It is absent from or present in only low amounts on resting T cells but is expressed strongly after activation. Cross-linking of Tp103 via a mAb CB.1 leads to triggering of functional activities in preactivated CD3+ T lymphocytes. By using mAb CB.1 in immunohistology we found that Tp103 is expressed in epithelial cells of various tissues, such as kidney, prostate, epidermis and on endothelia of liver, spleen, lungs, and most vessels, and in bile duct canaliculi in the liver. We found a carcinoma cell line expressing Tp103 and could precipitate a 110-kDa molecule from the surface of these cells. We considered several known molecules of similar distribution and molecular mass for identity with Tp103 and show here that Tp103 is probably identical to the proteolytic enzyme dipeptidyl aminopeptidase IV. When we purified Tp103 by affinity chromatography, typical dipeptidyl aminopeptidase IV activity copurified with Tp103. On activated T cells the enzymatic activity associated with Tp103 is expressed on the outside of the cell. We show that mAb CB.1 recognizes the same molecule as the anti-CD26 mAb anti-Ta1. The anti-Ta1 mAb was found to have T cell-activating activity too, but to differ in its requirements for triggering of T lymphocytes.

Protein kinase C-regulated production of prostacyclin by rat endothelium is increased in the presence of lipoxin A4

Prostacyclin is generated by cultured rat endothelial cells. Compound blocking activity of protein kinase C and cyclic nucleotide-dependent protein kinases (H7) and compound blocking interaction between Ca2+ and calmodulin (W7) diminish generation of prostacyclin in rat endothelial cells. These compounds give a synergistic effect when they are introduced to the endothelial cell cultures simultaneously. Compound HA1004, an inhibitor of cAMP- and cGMP-dependent protein kinases has no effect on prostacyclin generation. Lipoxin A4, a potent direct stimulator of protein kinase C, rapidly induces prostacyclin generation in rat endothelium in a dose- and time-dependent fashion. Lipoxin A4-induced generation of prostacyclin can be inhibited by H7 and W7 but not by HA1004. Lipoxin B4 has no significant effect on prostacyclin generation in rat endothelium. In conclusion, our results demonstrate that generation of prostacyclin by rat endothelial cells is regulated via a pathway involving protein kinase C and Ca2+.

Ganglioside biosynthesis in rat liver: different distribution of ganglioside synthases in hepatocytes, Kupffer cells, and sinusoidal endothelial cells

The activities of five glycolipid-glycosyltransferases, GL2, GM3, GM2, GM1, and GD1a synthase, were determined in a cell-free system with homogenate protein of total rat liver, isolated hepatocytes, Kupffer cells, and sinusoidal endothelial cells. In rat liver parenchymal and nonparenchymal cells ganglioside synthases were distributed differently. Compared to hepatocytes, Kupffer cells expressed a nearly sevenfold greater activity of GM3 synthase, but only 14% of GM2, 19% of GM1, and 67% of GD1a synthase activity. Sinusoidal endothelial cells expressed a pattern of enzyme activities quite similar to that of Kupffer cells with the exception of higher GM2 synthase activity. Activity of GL2 synthase was distributed unifromly in parenchymal and nonparenchymal cells of rat liver, but differed by sex. It was 1 to 2 orders of magnitude below that of all the other ganglioside synthases investigated. The results indicate GL2 synthase regulates the total hepatic ganglioside content, and hepatocytes but not nonparenchymal liver cells have high enzymatic capacities to form a-series gangliosides more complex than GM3.

14, 15-Epoxyeicosatrienoic acid promotes endothelial cell dependent adhesion of human monocytic tumor U937 cells

Arachidonic acid (AA) can be metabolized in endothelial cells (EC) to a series of epoxides via cytochrome P-450 epoxygenase with 14,15 epoxyeicosatrienoic acid (14,15-EET) as the major product. In this communication we report that 14,15-EET significantly enhances U937 cell attachment to EC with maximal cell attachment at 2.5 to 5 x 10(-7) M 14,15-EET. Thus, 14,15-EET may play a substantial role in inflammation and/or atherogenesis by inducing monocyte attachment to EC.

Expression of extracellular superoxide dismutase by human cell lines

Extracellular superoxide dismutase (EC-SOD) is the major SOD isoenzyme in extracellular fluids, but occurs also in tissues. The sites and characteristics of the synthesis of the enzyme are unknown. The occurrence of EC-SOD in cultures of a large panel of human cell lines was assayed by means of an e.l.i.s.a. Unlike the situation for the intracellular isoenzymes CuZn-SOD and Mn-SOD, expression of EC-SOD occurs in only a few cell types. None of the ten investigated suspension-growing cell lines produced EC-SOD. Among normal diploid anchorage-dependent cell lines, expression was found in all 25 investigated fibroblast cell lines, in the two glia-cell lines, but not in six endothelial-cell lines, two epithelial-cell lines or in two amnion-derived lines. Among neoplastic anchorage-dependent cell lines expression was found in 13 out of 29. EC-SOD was secreted into the culture medium by cell lines expressing the enzyme. The rate of EC-SOD synthesis varied by nearly 100-fold among the fibroblast lines and remained essentially constant in the individual lines during long-term culture. In the nine investigated cases, the secreted EC-SOD was of the high-heparin-affinity C type. It is suggested that tissue EC-SOD is secreted by a few well-dispersed cell types, such as fibroblasts and glia cells, to diffuse subsequently around and reversibly bind to heparan sulphate proteoglycan ligands in the glycocalyx of the surface of most tissue cell types and in the interstitial matrix.

Induction of peroxisomal enzymes in cultured porcine endothelial cells by the hypolipidemic drug ciprofibrate

The purpose of this study was to determine if the hypolipidemic peroxisome proliferator ciprofibrate, which induces peroxisomes in the liver, can induce peroxisomes in cultured porcine pulmonary endothelial cells. Ciprofibrate was added at three concentrations to cell cultures for a 6-day period. The induction of peroxisomes in the cells was detected by determining total peroxisomal beta-oxidation and peroxisomal catalase activity. The addition of ciprofibrate was found to increase peroxisomal enzyme activities in a dose-dependent manner, with the highest activity being reached at 1000 microM ciprofibrate. Ciprofibrate also caused an increased transfer of albumin across endothelial cells cultured on micropore filters. This study shows that peroxisomal enzyme activities can be induced by ciprofibrate in endothelial cells, which may have implications in diseases mediated by vascular injury.

Light microscopic localization of glycosyltransferase activities in cells and tissues

We describe an assay for light microscopic visualization of specific glycosyltransferases on tissue sections or on cells. The assay uses a sequence of enzyme reactions that yields two moles of NADH for each mole of the uridine-5'-diphosphate (UDP) released during transfer of a monosaccharide from a UDP sugar to an acceptor. When diaphorase and tetrazolium salts are present in the incubation mixture, the tetrazolium salts are reduced to colored diformazans, which precipitate at the sites of glycosyltransferase activity. The validity of the assay was established by applying the technique to spermatozoa and liver, in which some glycosyltransferases have previously been localized. When suspensions of mouse spermatozoa were assayed for galactosyltransferase (GalTase) activity, diformazan precipitates appeared on the plasma membranes overlying the anterior heads of the spermatozoa, in agreement with immunochemical localizations. In mouse liver slices assayed with bilirubin as acceptor for glucuronyltransferase (GluTase) activity, dense diformazan deposits appeared on the hepatocytes but not on endothelial cells, also in agreement with immunochemical data. In the absence of acceptor or UDP sugar donor, diformazan deposits were minimal and random in all tissues tested. The assay's versatility was tested by incubating tissues with different sugar donors and acceptors to localize other sites of transferase activity. In mouse frozen liver sections, GalTase activity occurred in both hepatocytes and endothelial cells; in sections of rat submaxillary glands, GalTase activity was detected in mast cells. In liver sections, GlcuTase activity with o-aminophenol as acceptor was located primarily on the endothelial cells. With the appropriate sugar donor and acceptor, this assay should detect any transferase, other than the glucosyltransferases, that utilizes UDP sugars.

The distribution, induction and isoenzyme profile of glutathione S-transferase and glutathione peroxidase in isolated rat liver parenchymal, Kupffer and endothelial cells

The distribution and inducibility of cytosolic glutathione S-transferase (EC 2.5.1.18) and glutathione peroxidase (EC 1.11.1.19) activities in rat liver parenchymal, Kupffer and endothelial cells were studied. In untreated rats glutathione S-transferase activity with 1-chloro-2,4-dinitrobenzene and 4-hydroxynon-2-trans-enal as substrates was 1.7-2.2-fold higher in parenchymal cells than in Kupffer and endothelial cells, whereas total, selenium-dependent and non-selenium-dependent glutathione peroxidase activities were similar in all three cell types. Glutathione S-transferase isoenzymes in parenchymal and non-parenchymal cells isolated from untreated rats were separated by chromatofocusing in an f.p.l.c. system: all glutathione S-transferase isoenzymes observed in the sinusoidal lining cells were also detected in the parenchymal cells, whereas Kupffer and endothelial cells lacked several glutathione S-transferase isoenzymes present in parenchymal cells. At 5 days after administration of Arocolor 1254 glutathione S-transferase activity was only enhanced in parenchymal cells; furthermore, selenium-dependent glutathione peroxidase activity decreased in parenchymal and non-parenchymal cells. At 13 days after a single injection of Aroclor 1254 a strong induction of glutathione S-transferase had taken place in all three cell types, whereas selenium-dependent glutathione peroxidase activity remained unchanged (endothelial cells) or was depressed (parenchymal and Kupffer cells). Hence these results clearly establish that glutathione S-transferase and glutathione peroxidase are differentially regulated in rat liver parenchymal as well as non-parenchymal cells. The presence of glutathione peroxidase and several glutathione S-transferase isoenzymes capable of detoxifying a variety of compounds in Kupffer and endothelial cells might be crucial to protect the liver from damage by potentially hepatotoxic substances.

Cytochrome P450IA1 induction and localization in endothelium of vertebrate (teleost) heart

Previous studies have shown that high levels of cytochrome P450 can occur in cardiac microsomes of vertebrates [Mol. Pharmacol. 21:517-526, (1982)]. Here we identify the dominant cardiac P450 in the marine fish scup as P450E, a teleost representative of P450IA1, and we describe its restricted cellular localization in the heart. Treatment of scup with beta-naphthoflavone produced an unusually strong (10-fold) induction of spectrally measured P450 in cardiac microsomes, with specific content reaching levels (0.5 nmol/mg) similar to those induced in scup liver. Microsomal ethoxyresorufin O-deethylase and aryl hydrocarbon hydroxylase activities, catalytic functions of scup P450E, were induced in parallel with P450 content. Similar induction was seen in both atrium and ventricle. Immunoblot analysis with monoclonal antibody 1-12-3, specific to scup P450E and other vertebrate P450IA1 proteins, showed that this hydrocarbon-inducible P450 is the dominant and possibly sole P450 form in heart microsomes of experimentally induced animals. Immunohistochemical analysis of scup heart sections (2-4-microns) with monoclonal antibody 1-12-3 revealed that P450E was detectable only in endothelial cells of the endocardium and of the coronary vasculature. A similar endothelial cell localization of the monoclonal antibody 1-12-3 epitope was observed in heart of rainbow trout, induced with beta-naphthoflavone, indicating a general nature for the endothelial localization of induced cardiac P450. Morphometric analysis showed that endothelium could constitute 8-9% of the volume of teleost heart, from which we calculate that P450IA1 could account for as much as 25% of the endothelial cell microsomal protein. Heart microsomes of untreated animals from contaminated environments also contained high levels of P450E, indicating that induction like that caused by beta-naphthoflavone could occur with chemicals in the environment. Strongly induced P450E (P450IA1) in endothelium could play a critical role in chemical-biological interactions involving xenobiotics affecting the vasculature of the heart or other organs.

The testicular transcript of the angiotensin I-converting enzyme encodes for the ancestral, non-duplicated form of the enzyme

The endothelial angiotensin I-converting enzyme (ACE) is organized in two large homologous domains, each bearing a putative active site. However, only one of these sites is probably involved in catalyzing the conversion of angiotensin I into angiotensin II. The testicular form of ACE is equally active, encoded by the same gene, but translated from a shorter mRNA. Molecular cloning of the human testicular ACE cDNA indicates that the mRNA codes for 732 residues (vs 1306 in endothelium). The testicular transcript corresponds to the 3' half of the endothelial transcript and encodes one of the two homologous domains of endothelial ACE, preceded by a short specific sequence. This 5' specific sequence contains 228 nucleotides and encodes 67 amino acids, including the putative signal peptide followed by a serine/threonine-enriched region, presumably glycosylated. The testicular transcript corresponds to the ancestral, non-duplicated form of the ACE gene. Since the carboxyl-terminal domain of the endothelial ACE is expressed in the testicular enzyme, it is likely that it bears the active site in both forms.

Plasma protein binding and endothelial enzyme interactions in the lung

The influence of plasma albumin binding of the synthetic angiotensin-converting enzyme (ACE) substrate [3H]benzoyl-phenylalanyl-alanyl-proline (BPAP) on BPAP hydrolysis by pulmonary endothelial ACE was studied in isolated rabbit lungs perfused with a salt solution containing either 5% bovine serum albumin (BSA) or 5% dextran. The single-pass indicator-dilution method was used to measure the fraction (M) of [3H]BPAP hydrolyzed. Lung M was greater with albumin-free perfusate than when BSA was present. M decreased as the time (ti) that the BPAP was in contact with the BSA before reaching the lung was increased, suggesting that some BSA binding sites for BPAP were not in equilibrium during bolus transit through the lungs. The M vs. ti data were correlated using a model incorporating both rapid and slow binding kinetics of BPAP and BSA. For the slow BPAP-BSA interaction, the dissociation rate constant was approximately 0.015 s-1, and the fraction of the BPAP bound to these slowly equilibrating sites at equilibrium was approximately 22%. The results indicate that transient plasma protein binding kinetics can affect lung BPAP hydrolysis.

Steroid regulation of monoamine oxidase activity in the adrenal medulla

Administration of different steroid hormones in vivo has distinct and specific effects on the MAO activity of the adrenal medulla. In an effort to reconstitute these effects in defined cells, we have isolated endothelial cells and chromaffin cells from the bovine adrenal medulla and tested each cell type for sensitivity to these steroids. As in the intact animal, we found that endothelial cell MAO activity was stimulated 1.5- 2.5-fold by 10 microM progesterone, hydrocortisone, and dexamethasone, inhibited by ca. 50% by 17-alpha-estradiol, but unaffected by testosterone. The type of MAO in the endothelial cells was found to be exclusively of the A type. The chromaffin cells had MAO B exclusively and were inert to treatment with dexamethasone. The mode of action of the various steroids on MAO A activity in endothelial cells seemed to be that of affecting the number of MAO molecules, as binding of [3H]pargyline, an MAO inhibitor, changed in proportion to changes in enzyme activity. Consistently, the kinetic parameters for MAO A showed changes in Vmax but not Km under all conditions. The specificity of steroid action on MAO A activity was also supported by the fact that steroid-induced changes in total cell division ([14C]thymidine incorporation) and total protein synthesis ([14C]leucine incorporation) were seen after changes in MAO A. We conclude that the differential effects of steroids on MAO activity in the intact adrenal medulla can be reproduced in cultured adrenal medullary endothelial cells but not in chromaffin cells. Therefore we suggest that the action of these steroid hormones on the intact adrenal medulla may be restricted to the endothelial cell component of this tissue.

[Ultracytochemical study of Ca++-ATPase activity in rabbit trabecular meshwork]

Cytochemical localization of Ca++-ATPase activity in endothelial cells of rabbit trabecular meshwork was studied by the method of Ando et al. Ca++-ATPase activity was localized on the plasma membrane, mitochondria, endoplasmic reticulum, and the basal cytoplasmic matrix, which includes the actin filament bundles beneath the basal plasma membrane. Ca++-ATPase activity associated with membranous organelles was considerably reduced by quercetin, an inhibitor of ATP dependent Ca++-transport. Changes in the concentrations of Ca (1-10mM) greatly influenced the activity of the cytoplasmic matrix. The most intense activity was obtained at 1mM Ca. These findings suggest that Ca++-ATPase of mitochondria and sarcoplasmic reticulum is related to the active uptake of Ca++ by these structures, and the reaction within actin filament bundles may reflect actomyosin ATPase activity. The function of cAMP and actomyosin ATPase reaction may be to provide contractility of the trabecular meshwork resulting in an alteration of outflow resistance of aqueous humor drainage.

Endothelial pump function in rabbit corneas kept under eye bank conditions

Endothelial pump activity of rabbit corneas kept under eye bank conditions was examined. Different swelling rates for medium with and without ouabain were used for the calculation of endothelial pump rate. After an initial weight increase during 24 h, the control corneas kept approximately constant weight for 1 week. The ouabain-treated corneas increased continually, and during the time interval 24-48 h the increase was 41% of start weight. This corresponds to a pump activity of 2.3% of the weight of the normal corneal stroma per h, or to 9 microns of the corneal thickness per h. Each endothelial cell pumped fluid corresponding to 200-300% of the cell volume per h between day 1 (24 h) and day 8. The pump activity was gradually reduced after 8 days. The average weight was the same in both groups after 7 weeks.

Lung injury in rabbits induced by intravenous administration of heterologous polyclonal antibodies to angiotensin converting enzyme (kininase II)

Antibody interactions with the endothelial cell membrane glycoprotein angiotensin converting enzyme (Kininase II) in vivo exhibit features of aggregation and capping with resultant shedding similar to those events described in several in vitro isolated cell systems. Requirements for divalent ligand binding, deposition of complement and participation of cytoskeletal elements are demonstrated in vivo. Persistence of antigen in immune complexes with complement interaction appear to be necessary to induce an inflammatory response. Abrogation of this response occurs when circumstances permit antigenic modulation with removal of the immune complex from the endothelial surface.

Interspecific variations of cerebral endothelial cholinesterases in rodents and carnivores

The cholinesterase equipment of cerebral microvessels was studied in some rodents and carnivores using the Koelle-Friedenwald histochemical method with 3 artificial substrates and specific inhibitors for butyrylcholinesterase or acetylcholinesterase. Our observations reveal a great heterogeneity in cholinesterase types and their distribution in each of the different species studied. Only in the rat, butyrylcholinesterase appears to be a marker for the microvessels provided with a blood-brain barrier.

Stimulation of de novo synthesis of prostaglandin G/H synthase in human endothelial cells by phorbol ester

The present study was undertaken to determine the mechanism by which phorbol ester stimulates eicosanoid synthesis in endothelial cells. We observed that phorbol 12-myristate 13-acetate (PMA) actively stimulated eicosanoid synthesis over a prolonged period of time, and the stimulatory effect was abolished by cycloheximide and actinomycin D. Western blot was employed to test the hypothesis that PMA elicited sustained eicosanoid synthesis via the stimulation of de novo synthesis of prostaglandin G/H synthase (cyclooxygenase, EC 1.14.99.1). Treatment of cultured human umbilical vein endothelial cells resulted in an enhancement of the 70-kDa immunoreactive prostaglandin G/H synthase band over the control cells treated with medium alone. The enhancement was abolished by cycloheximide. Human umbilical vein endothelial cells were then metabolically labeled with L-[35S]methionine, and the effect of PMA on methionine incorporation was evaluated by immunoblotting. PMA increased the synthetic rate of prostaglandin G/H synthase over the control cells. By pulse-chase experiments, we further showed that prostaglandin G/H synthase has a rapid turnover rate (t1/2 less than 10 min) in control cells, and PMA had no effect on the enzyme turnover. Our data indicate that PMA increases the synthesis of prostaglandin G/H synthase which is required for circumventing the autoinactivation of prostaglandin G/H synthase and hence permit sustained conversion of arachidonic acid into eicosanoids.