20-Hydroxyeicosatetraenoic Acid (20-HETE) Metabolism in Coronary Endothelial Cells*

We have investigated the role of endothelial cells in the metabolism of 20-hydroxyeicosatetraenoic acid (20-HETE), a vasoactive mediator synthesized from arachidonic acid by cytochrome P450 ω-oxidases. Porcine coronary artery endothelial cells (PCEC) incorporated 20-[3H]HETE primarily into the sn-2 position of phospholipids through a coenzyme A-dependent process. The incorporation was reduced by equimolar amounts of arachidonic, eicosapentaenoic or 8,9-epoxyeicosatrienoic acids, but some uptake persisted even when a 10-fold excess of arachidonic acid was available. The retention of 20-[3H]HETE increased substantially when methyl arachidonoyl fluorophosphonate, but not bromoenol lactone, was added, suggesting that a Ca2+-dependent cytosolic phospholipase A2 released the 20-HETE contained in PCEC phospholipids. Addition of calcium ionophore A23187 produced a rapid release of 20-[3H]HETE from the PCEC, a finding that also is consistent with a Ca2+-dependent mobilization process. PCEC also converted 20-[3H]HETE to 20-carboxy-arachidonic acid (20-COOH-AA) and 18-, 16-, and 14-carbon β-oxidation products. 20-COOH-AA produced vasodilation in porcine coronary arterioles, but 20-HETE was inactive. These results suggest that the incorporation of 20-HETE and its subsequent conversion to 20-COOH-AA in the endothelium may be important in modulating coronary vascular function.

20-Hydroxyeicosatetraenoic acid (HETE), 1 an eicosanoid synthesized from arachidonic acid by cytochrome P450 (CYP) isoforms of the 4A and 4F classes (1)(2)(3), produces vasoconstriction of small arteries and regulates salt balance in the kidney (4 -6). There is growing interest in 20-HETE because it increases systemic blood pressure, and animal studies indicate that it plays a key role in the pathophysiology of some experimental forms of hypertension (2, 4, 6 -8). For example, the CYP4A2 gene is overexpressed during the development of hypertension in the SHR rat (9), and disruption of the CYP4a14 gene causes 20-HETE production and hypertension in male mice (10). Furthermore, treatment with an antisense oligonucleotide reduces CYP4A expression in the vasculature and lowers blood pressure (11), and treatment with a CYP4A inhibitor reduces hypertension induced by angiotensin II (12).
As opposed to its vasoconstrictor effect in the renal and cerebral circulation, 20-HETE acts as a vasodilator in the pulmonary and coronary circulations (22,23). The pulmonary vasodilator effect is dependent on the presence of endothelium (22), and the vasorelaxation produced by 20-HETE in the coronary artery is mediated by the release of prostacyclin (23). 20-HETE also is converted to a bronchodilator substance by the pulmonary endothelial cyclooxygenase (24). In an attempt to better understand the mechanism of these endothelium-dependent effects, we have investigated the utilization of 20-HETE in porcine coronary artery endothelial cells (PCEC), an experimental model that we have used previously for biochemical studies of eicosanoid metabolism (25)(26)(27)(28). To investigate the potential functional significance of these observations, we examined the vasoactivity of 20-HETE and 20-carboxy-arachidonic acid (20-COOH-AA), its main metabolite (19), in an isolated porcine coronary microvessel preparation. (16.7 Ci/nmol) was synthesized by incubating 15 nmol of arachidonic acid containing 250 Ci [5,6,8,9,11,12,14, H]arachidonic acid (60 -100 Ci/mmol, American Radiolabeled Chemicals, Inc., St. Louis, MO) with commercially available microsomes containing 30 pmol of baculovirus expressed human recombinant CYP4F3B, CYP reductase and cytochrome b 5 isolated from BTI-TN-5B1-4 insect cells (Gentest, Woburn, MA). The mixture was incubated in 1 ml of 200 mM phosphate buffered saline, pH 7.4, containing 0.1 M bovine serum albumin (BSA), 1.3 mM NADP ϩ , 3.3 mM glucose 6-phosphate, 3.3 mM MgCl 2 , 0.4 units of glucose 6-phosphate dehydrogenase, and 0.05 mM sodium citrate. After 1 h at 37°C, the reaction was terminated by addition of 1 ml H 2 O and formic acid to lower the pH to 4, and the lipids were extracted three-times with 8 ml of H 2 O-saturated ethyl acetate. The ethyl acetate extracts were combined, dried under N 2 , suspended in 3 ml of acetonitrile, and separated by reverse-phase high-performance liquid chromatography (HPLC) on a 5-m 4.6 ϫ 250 mm Discovery C 18 column (Supelco, Bellefonte, PA) and a dual pump gradient HPLC system (Gilson Medical Electronics, Inc., Middleton, WI) equipped with Model 306 pumps, a Model 117 Dual wavelength UV detector, a Model 231 XL automatic sample injector. The solvent system contained H 2 O adjusted to pH 4.0 with formic acid and an acetonitrile gradient that began at 30% and was increased to 57% acetonitrile at 30 min, to 65% over the next 20 min, to 100% over 5 min, and then maintained at 100% acetonitrile for 15 min. Fig. 1 shows the analysis of the radiolabeled material. The main product eluted from the reverse phase HPLC column 24 min before [ 3 H]arachidonic acid (Fig. 1A). As shown in Fig. 1B Product containing no radioactivity also was synthesized and isolated by HPLC. The HPLC eluate containing the main product was collected in 0.5 ml fractions and added to 2.5 ml chloroform-methanol (2:1, v/v), and the fractions were pooled. Phosphate-buffered saline was added, and the chloroform phase was separated and collected. After removal of the solvent under N 2 , the lipid extract was dissolved in acetonitrile and further separated by reverse phase HPLC with a solvent system consisting of H 2 O adjusted to pH 4.0 with formic acid and an acetonitrile gradient that increased from 30 to 100% over 70 min (29). The product was collected, dried under N 2 , and analyzed by liquid chromatography combined with mass spectrometry (27). The mass spectrum was obtained by using an Agilent 1100 MSD liquid chromatography-mass spectrometry system with source collision-induced decomposition (27). As shown in Fig. 1C, the mass spectrum contained a molecular ion at m/z 319 and ions at m/z 301 (M-H 2 O) and m/z 275 (M-CO 2 ), consistent with the structure of 20-HETE.
The efflux of 20-HETE previously incorporated by the PCEC was investigated using a pulse-chase experimental design. PCEC cultures were incubated initially for 1 h with 20-[ 3 H]HETE and after removal of this medium, the cells were washed and incubated for 20 min to 16 h in a culture medium containing BSA. Following incubation, the medium and cells were separated, and the lipid-soluble material was extracted and assayed for total radioactivity content. In some experiments, ionophore A23187 was added to determine whether activation of a calciummediated signaling pathway increased the efflux of 20-[ 3 H]HETE from the PCEC (25)(26)(27)(28). In other experiments, either methyl arachidonoyl fluorophosphonate (MAFP) or bromoenol lactone (BEL), both obtained from Biomol (Plymouth Meeting, PA), was added to investigate the role of phospholipase A 2 (PLA 2 ) in mediating the efflux of 20-[ 3 H]HETE from the cells.
Assay of Incubation Medium-The amount of radioactivity remaining in the medium after incubation was measured by liquid scintillation counting. Lipids present in the medium were extracted twice with 4 ml of ice-cold ethyl acetate saturated with water. Extracts were combined, the ethyl acetate was evaporated under N 2 , and the lipids were dissolved in acetonitrile and separated by reverse-phase HPLC with the dual pump gradient system and Discovery C 18 column. The solvent system contained H 2 O adjusted to pH 4.0 with formic acid and an acetonitrile gradient that increased from 30 to 100% over 70 min at a flow rate of 0.7 ml/min (27). Radioactivity was measured by combining the column effluent with scintillator solution and then passing the mixture through a flow scintillation detector (IN/US Systems, Inc., Tampa, FL).
Analyses of Cell Lipids-After the medium was removed and the PCEC monolayer washed with ice-cold buffer, the lipids were extracted from the cells with 6 ml of chloroform-methanol (2:1, v/v). Phosphatebuffered saline, 1.5 ml containing 0.05 N HCl was added, the phases were separated, the chloroform was removed and dried under N 2 . The lipid residue was dissolved in 200 l of chloroform-methanol, and an aliquot was dried under N 2 and assayed for radioactivity in a liquid scintillation spectrometer (31).
To determine the distribution of the radioactivity in the cell lipids, aliquots of the extract were separated by thin layer chromatography (TLC) on Whatman LK5D silica gel plates (Whatman Inc., Clifton, NJ) with a solvent system of chloroform-methanol-40% methylamine-H 2 O (60:36:1.5: (27). The radioactivity contained in the separated lipids was determined with an AR-200 Imaging Scanner (Bioscan, Inc., Washington, D. C.). Phospholipid standards (Avanti Polar Lipids, (Naperville, IL) were added to each chromatogram, and the separated lipids were visualized by staining with 1-anilinonaphthalene 8-sulfonic acid.
Studies also were done to determine the positional distribution of the radioactivity incorporated into the PCEC phospholipids (28,32). After incubation with 20-[ 3 H]HETE, the phospholipids contained in the cell lipid extract were separated by TLC as indicated above and each phospholipids fraction was extracted and suspended in 1 ml of phosphate buffer (pH 7.5) containing 2 mM CaCl 2 . Snake venom PLA 2 (obtained from Sigma, 125 units/ml) dissolved in this buffer solution was added, and the samples were sealed under N 2 and incubated at 37°C for 45 min. The hydrolysates were extracted with chloroform-methanol (2:1 v/v), dried under N 2 , and fractionated by TLC using a solvent mixture containing chloroform, methanol, 40% methylamine, water (60:36:1.5: , and the distribution of radioactivity in the chromatogram was determined with the AR-200 Imaging Scanner. To determine whether the radioactivity incorporated in the PCEC lipids remained as 20-HETE or was converted to metabolites, additional aliquots of the cell extract were hydrolyzed for 1 h at 50°C in 0.5 ml methanol containing 50 l of 2 N NaOH (32). After the pH was adjusted to 7.0 with formic acid/H 2 O (1:10, v/v) and the volume adjusted to 1 ml with H 2 O, the hydrolyzed lipid products were extracted twice with 4 ml of ice-cold ethyl acetate saturated with H 2 O. The ethyl acetate was removed under N 2 , and the lipids were dissolved in acetonitrile and separated by reverse-phase HPLC on the Discovery C 18 column with a gradient containing H 2 O adjusted with formic acid to pH 4.0 and acetonitrile (27).
Mass Spectrometry-Liquid chromatography tandem mass spectrometry (LC/MS-MS) experiments to separate and identify the metabolites formed by the PCEC from 20-HETE were performed using a ThermoFinnigan LCQ Deca quadrupole ion trap mass spectrometer interfaced with a ThermoFinnigan Surveyor liquid chromatograph. The same HPLC parameters were used as described above, except a Supelco C 18 5 m 2.1 ϫ 150 mm Discovery column was used and the flow rate was 200 l/min. Negative electrospray ionization was also used. The N 2 sheath gas and auxiliary gas were maintained at 40 and 30 arbitrary units, respectively, the capillary was heated to 300°C, and the normalized collision energy was 45% for the LC/MS-MS experiments. Ther-moFinnigan Xcalibur 1.3 software was used for data processing.
Porcine Coronary Microvessel Studies-Porcine hearts were obtained from a local abattoir. The hearts were quickly harvested and immediately placed in cold (4°C), oxygenated Krebs bicarbonate buffer solution for dissection. A standard in vitro pressurized arteriole preparation was used to study coronary microvessels (33,34). Ventricular microvessels of 100 -175 m intraluminal diameter and ϳ1 mm in length were carefully removed from the myocardium, cleaned with the aid of a dissecting microscope and placed in an organ chamber. Each end of the microvessel was cannulated with a glass micropipette and secured with 10-0 ophthalmic suture. The organ chamber was placed on the stage of an inverted microscope. Attached to the microscope were a video camera, a video monitor, and a calibrated video caliper. The organ chamber was connected to a rotary pump that continuously circulated oxygenated Krebs Henseleit solution containing in mM: NaCl 120.0; KCl 4.7; CaCl 2 2.5; MgSO 4 1.2; NaHCO 3 23.0; KH 2 PO 4 1.2; glucose 11.0 and EDTA 0.025. Solutions were aerated with 20% O 2 , 5% CO 2 , and 75% N 2 and maintained at 37°C, pH 7.4. An image of the microvessel was displayed on the video monitor and intraluminal diameters were measured by manually adjusting the video micrometer. The resolution of the system allowed measurement of very small (1-2 m) changes in vessel diameter.
Microvessels were allowed to equilibrate for 30 min at a hydrostatic distending pressure of 60 mm Hg under conditions of no flow. KCl (75 mM) was added to the bath to test constrictor capacity. After washing with fresh Krebs buffer, the vessel diameter returned to baseline. Endothelin-1 (0.40 -1.2 nM, Phoenix Peninsula Laboratories, Inc., San Carlos, CA) was used to constrict the microvessels to 30 -60% of their resting diameter. Cumulative concentration-response relationships were evaluated for 20-HETE (10 Ϫ10 to 10 Ϫ6 M), 20-COOH-AA, (10 Ϫ10 to 10 Ϫ6 M), or sodium nitroprusside (10 Ϫ10 to 10 Ϫ4 M) by adding the compound directly to the organ bath. A single dose of sodium nitroprusside (10 Ϫ4 M, Sigma Chemical) was given at the end of each experiment to determine the maximal diameter. All solutions and vasoactive agents were prepared fresh on the day of the experiment.
Statistical Analysis-Data are expressed as mean Ϯ S.E. Sigmastat software (Jandel Scientific) was used for statistical analyses. All concentration response curves were evaluated for differences using 2-way repeated measures analysis of variance followed by the Fisher LSD correction for multiple comparisons. Differences with p Ͻ 0.05 are considered significant.  (Fig. 2B). However, the percentage uptake of available 20-[ 3 H]HETE decreased as the concentration in the medium increased, from 50% at 0.25 M to 17% at 7.5 M.

20-HETE
Although the capacity of the PCEC to incorporate 20-HETE is substantial, it is less than arachidonic acid under the same conditions. For example, at a concentration of 1 (Fig. 3). The major radiolabeled metabolites detected during the incubation are designated in the chromatograms as Products I-IV. More than 95% of the lipid-soluble radioactivity recovered from the medium after 1 h of incubation remained as 20-HETE, which had an RT of 32.2 min in this HPLC analysis (Fig 3A). However, 20-[ 3 H]HETE accounted for only 40% of the radioactivity remaining in the medium after 4 h, and Product III (RT, 22.5 min) was the most abundant radiolabeled metabolite (Fig. 3B). At this time, the four main metabolites accounted for 12% of the 20-[ 3 H]HETE added to the medium at the start of the incubation. After 8 h, little radiolabeled 20-HETE or Product II remained, Product I (RT, 31 min) was the major metabolite, and Product IV (RT, 17 min) increased (Fig. 3C). The only prominent metabolite remaining in the medium at the end of the 24-h incubation was Product I (Fig. 3D). Increasing amounts of radiolabeled material that eluted with the solvent front also accumulated as the incubation progressed. Preliminary analysis indicated that this material consisted of a mixture of shortchain products, and no further analysis of this material was attempted.
The main 20-HETE metabolites were identified by LC/MS-MS. Fig. 4 shows the MS-MS spectra of the precursor (M-H) Ϫ ions of 20-HETE (Fig. 4A) and the metabolites that were detected (Fig. 4, B-E). The fragmentation seen in these spectra are consistent with the following structures for the products:  (Fig. 5B). Thus, PCEC have the capacity to convert substantial amounts of 20-HETE to 20-COOH-AA.
Cell Lipid Analysis-A study was done to determine whether these metabolites accounted for a substantial amount of the radioactivity that accumulated in the cells. Following incubations of the PCEC for various times between 2 and 24 h with 1 M 20-[ 3 H]HETE, the cell lipids were hydrolyzed by saponification, and the radioactivity contained in the resulting lipid extract was separated and assayed by HPLC. Fig. 6 demonstrates that essentially all of the radioactivity contained in the cell lipids remained as 20-HETE at the end of the 24 h incubation. Similar chromatograms were obtained from the cells at each of the earlier time points (data not shown). Additional studies indicated that when the PCEC were incubated for 2 h with 20-[ 3 H]HETE concentrations between 0.25 and 7.5 M, between 94 and 100% of the radioactivity incorporated into the cell lipids remained as 20-HETE (data not shown). Thus, under all of the conditions tested, the 20-HETE that was retained in the PCEC was not structurally modified, and none of the radiolabeled metabolites that were observed in the culture medium were detected in the cells.  Fig. 7. Following a 1-h incubation with 1 M 20-[ 3 H]HETE, phospholipids contained 78% of the incorporated radioactivity, 16% was contained in neutral lipids (NL), and 6% was present as unesterified 20-HETE. The distribution in phospholipids was 31% in PI, 39% in choline phosphoglycerides (PC) and 30% in ethanolamine phosphoglycerides (PE). When 5 M of the acyl CoA synthase inhibitor triacsin C obtained from Biomol was added to the incubations, the total uptake of 20-[ 3 H]HETE decreased by 58%, and the incorporation in phospholipids and NL decreased by 63%. The magnitude of the decreases in the PI, PC, PE, and NL fractions were similar, indicating that the incorporation of 20-HETE into PCEC lipids is CoA-dependent. However, no decrease was observed in the amount of unesterified 20-HETE present in the cells.

The distribution of the incorporated 20-[ 3 H]HETE in the PCEC lipids, as determined by TLC analysis, is illustrated in
A study was done to determine the positional distribution of the 20-[ 3 H]HETE incorporated into the PCEC phospholipids. After a 1 h incubation with 1 M 20-[ 3 H]HETE, the PI, PC, and PE fractions were separated by TLC, and each fraction was isolated and incubated with PLA 2 . After extraction and separation of the products by TLC, all of the radioactivity originally contained in PI and PC, and 92% of the radioactivity in PE, co-migrated with the 20-HETE standard (data not shown). Therefore, almost all of the 20-[ 3 H]HETE was incorporated into the sn-2 position of the PCEC phospholipids.
Retention in Cells-The retention of 20-HETE and arachidonic acid in the PCEC was compared under basal conditions (Fig. 9)  Because more than 75% of the 20-[ 3 H]HETE present in the labeled PCEC was contained in phospholipids and almost all of this was present in the sn-2 position, we tested whether PLA 2 inhibition of would prevent the release of 20-HETE from the cells. Two inhibitors were tested: MAFP, a dual cytosolic PLA 2 (cPLA 2 ) and calcium-independent PLA 2 (iPLA 2 ) inhibitor (42,43); and BEL, a selective iPLA 2 inhibitor (43,44). The experimental conditions were similar to those described above, except that the time of incubation of the labeled cells was reduced to 3 h because of the possibility that the inhibitors might produce cytotoxicity if the exposure was prolonged. The release of 20-[ 3 H]HETE from the PCEC was reduced by 53% when 10 M MAFP was added (55 Ϯ 2 pmol in the control cultures as compared with 26 Ϯ 1 pmol in this treated with MAFP, n ϭ 3). By contrast, the amount of 20-[ 3 H]HETE that was released into the medium was not reduced appreciably by either 10 M or 25 M BEL (48 Ϯ 1 and 50 Ϯ 4 pmol, respectively, n ϭ 3). These concentrations of BEL are reported to be effective against iPLA 2 -mediated processes in other systems (43,44). The time dependence of the reduction in 20-[ 3 H]HETE release from the PCEC produced by 15 M MAFP is shown in Fig. 10. Less 20-[ 3 H]HETE accumulated in the medium throughout this period when MAFP was added, and a 62% reduction occurred at the end of the 3 h incubation. The differential effects of MAFP and BEL suggest that a substantial amount of the efflux of 20-HETE is mediated by cPLA 2 .
Ca 2ϩ -stimulated 20-HETE Release-To further explore the Ca 2ϩ -dependence of 20-HETE efflux, the effect of a Ca 2ϩ ionophore was determined. The results are shown in Fig. 11. PCEC cultures were incubated for 1 h with 20-[ 3 H]HETE. After the medium was removed and the cells washed, the cultures were incubated for 20 min in a fresh medium containing 5 M BSA, with or without 4 M ionophore A23187. The cultures incubated with A23187 released 4.5 times more radioactivity than the corresponding controls (Fig 11A). HPLC analysis of the incubation medium indicated that radioactivity contained in the medium of the cultures exposed to A23187 was almost entirely 20-HETE (Fig. 11B). These data demonstrate that a rapid release of some 20-HETE contained in the PCEC can occur and that this process is Ca 2ϩ -activated.
Vascular Reactivity-The effects of 20-HETE and 20-COOH-AA on vascular function in porcine coronary arterioles is illustrated in Fig. 12. In vessels preconstricted with endothelin, 20-COOH-AA produced concentration-dependent vasodilation at concentrations as low as 100 nM. The magnitude of vasodilation was significantly less than that produced by the potent vasodilator sodium nitroprusside. In contrast, 20-HETE produced no demonstrable vasodilation at the concentrations studied. Thus, our findings suggest that under these defined experimental conditions, 20-COOH-AA produced moderately potent vasodilation, while the same concentrations of 20-HETE did not alter microvascular diameter. DISCUSSION These results indicate that the endothelium has the potential to modulate the function of 20-HETE in the vascular wall. We find that cultured endothelial cells have a large capacity to take up and metabolize 20-HETE in the concentration range that produces functional responses in vascular and renal tissues (6,14,16,45). When 1 M 20-[ 3 H]HETE was initially available, the cells took up a 45% in 1 h. After 4 h, 9.5% of the initially available 20-[ 3 H]HETE was converted to the radiolabeled metabolites detected in the medium. As observed in the rabbit kidney (20), the PCEC incorporated 20-HETE primarily in phospholipids and to a small extent in neutral lipids. The 20-HETE was distributed in PC, PI, and PE and was present almost entirely in the sn-2 position of these phospholipids, a generally similar pattern to what has been observed in the cortex and medulla of the rabbit kidney (20). A small amount of unesterified 20-HETE also was detected in the PCEC, possibly due in part to binding to a cytosolic fatty acid-binding protein (46).
Triacsin C, an acylcoenzyme A synthase inhibitor (47), reduced the amount of 20-HETE present into phospholipids and neutral lipids, indicating that the incorporation into these endothelial lipids is coenzyme A-dependent. Importantly, we show that the 20-HETE incorporated into endothelial cell lipids can be released in response to agonist-induced phospholipase activation, a mechanism that has the potential to modulate endothelium-dependent vasomotor responses.
Previous studies have shown that angiotensin II stimulates 20-HETE release from rat renal microvessels and that 20-HETE is the second messenger for endothelin-1 in the rat proximal tubule (48,49). Based on these findings, it has been proposed that 20-HETE, like EETs, is stored in the kidney and is released by peptide hormone activation of an acylhydrolase (20). These physiological agonists were not tested in the PCEC cultures. However, the fact that 20-HETE is present in the sn-2 position of the PCEC phospholipids and that its release was stimulated by the addition of a Ca 2ϩ ionophore suggests that a similar signal transduction mechanism may be operative in the endothelium. This hypothesis also is consistent with the observation that 20-HETE release from the PCEC was reduced substantially by MAFP, a cPLA 2 inhibitor (42,43).
In addition to being incorporated into endothelial lipids, 20-HETE was metabolized by the endothelial cells through two oxidative processes, -oxidation that formed 20-COOH-AA, and ␤-oxidation. The conversion of 20-HETE to 20-COOH-AA was first detected in rabbit kidney cells isolated from the thick ascending loop of Henle (19). We find, as reported previously in the kidney studies (19), that the difference in the reverse-phase HPLC RTs of 20-COOH-AA and 20-HETE is less than 2 min. Therefore, it is likely that the formation of 20-COOH-AA may be overlooked in work with 20-HETE unless the RTs of compounds detected by HPLC are carefully compared with standards and the structure is confirmed by tandem mass spectrometry. The formation of 20-COOH-AA by the PCEC occurred slowly as compared with 20-HETE incorporation into endothelial lipids. However, 20-COOH-AA continued to accumulate in the medium during relatively long incubations, indicating that endothelial cells do not rapidly degrade this metabolite. Previous studies by Escalante et al. (19) demonstrated that 20-COOH-AA was as effective as 20-HETE in inhibiting 86 Rb uptake by the kidney tubular cells, demonstrating that it is a biologically active metabolite and not simply a degradation product of 20-HETE. We also find that 20-COOH-AA is a moderately potent vasodilator in the porcine coronary microvessels, providing additional evidence that this metabolite has biological activity. As opposed to 20-COOH AA, 20-HETE failed to produce vasodilation in this preparation. Therefore, 20-COOH-AA may have functional effects that are distinct from those of 20-HETE. Taken together, these findings suggest that the endothelium may be a site of 20-COOH-AA production, and that the vascular effects of 20-HETE may be modulated by endothelial conversion to 20-COOH-AA.
Partial ␤-oxidation metabolites also were produced from 20-HETE by the PCEC. As opposed to 20-COOH-AA, these products appeared sequentially in the medium and then were converted to chain-shortened derivatives. Endothelial cells synthesize partial ␤-oxidation products from other eicosanoids, including intra-chain HETEs and EETs (27,39,50,51), and studies with mutant human skin fibroblasts indicate that this occurs through peroxisomal ␤-oxidation (52,53). While it seems likely that the ␤-oxidation pathway functions to degrade 20-HETE, the possibility that one or more of the chain-shortened metabolites may have functional activity cannot be excluded. 20-HETE is converted to 20-hydroxy-prostaglandin G 2 and H 2 by cyclooxygenase in rat aortic endothelium and ram seminal vesicle microsomes (54). These prostaglandin endoperoxide derivatives produce vasoconstriction of the rat aorta. Lung microsomes also convert 20-HETE to a prostanoid product (22), and rabbit airway epithelial cells convert 20-HETE to a cyclooxygenase product that causes bronchial relaxation (24). We did not detect appreciable amounts of prostaglandin-like products, as determined by either HPLC or mass spectrometry, following incubation of the PCEC with 20-HETE. In this regard it should be noted that more recent results with bovine coronary arteries suggest an alternative mechanism for the cyclooxygenase-dependent vasorelaxant effect of 20-HETE (23). This work indicates that 20-HETE increases the production of prostacyclin in the coronary artery endothelium. The rapid and substantial incorporation of 20-HETE into PCEC phospholipids that we observed is consistent with such a mechanism. Incorporation of 20-HETE would be expected to displace some endogenous arachidonic acid from the sn-2 position of phospholipids and thereby make increased amounts available to the endothelial cyclooxygenase.
In conclusion, we suggest that through uptake and ␤-oxidation, the endothelium can reduce the availability of 20-HETE to the vascular smooth muscle. We also suggest that through incorporation into phospholipids, 20-HETE can affect the functional properties of the endothelium. In particular, the release of incorporated 20-HETE during agonist-induced activation of endothelial cells could modulate endothelium-dependent vasomotor responses, and the formation of 20-COOH-AA may further affect this process. Therefore, we propose the hypothesis that the endothelium can have an important influence on the vasoactive effects of 20-HETE.