Docosatetraenoic acid in endothelial cells: Formation, retroconversion to arachidonic acid, and effect on prostacyclin production

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 244, No. 2, February 1, pp. 813-823,1986

Docosatetraenoic Acid in Endothelial Cells: Formation, Retroconversion Arachidonic Acid, and Effect on Prostacyclin Production’** CRAIG

J. MANN,3 AND

Department

TERRY L. KADUCE, PAUL ARTHUR A. SPECTOR4

of Biochemistry, Received

of Iowa,

University August

Iowa

1

H. FIGARD,

City,

Iowa

52242

26,1985

Cultured bovine aortic endothelial cells convert arachidonic acid to docosatetraenoic acid and also take up docosatetraenoic acid from the extracellular fluid. After a 24-h incubation with biosynthetically prepared [3H]docosatetraenoic acid, about 20% of the cellular fatty acid radioactivity was converted to arachidonic acid. Furthermore, in pulse-chase experiments, the decrease in phospholipid docosatetraenoic acid content was accompanied by an increase in arachidonic acid, providing additional evidence for retroconversion. These findings suggest that one possible function of docosatetraenoic acid in endothelial cells is to serve as a source of arachidonic acid. The endothelial cells can release docosatetraenoic acid when they are stimulated with ionophore A23187, but they do not form appreciable amounts of eicosanoids from docosatetraenoic acid. Enrichment of the endothelial cells with docosatetraenoic acid reduced their capacity to produce prostacyclin (PGIa) in response to ionophore A23187. This may be related to the fact that docosatetraenoic acid enrichment caused a 40% reduction in the arachidonic acid content of the inositol phosphoglycerides. In addition, less prostacyclin was formed when the enriched cells were incubated with arachidonic acid, suggesting that docosatetraenoic acid also may act as an inhibitor of prostaglandin synthesis in endothelial Celk..

0 1986 Academic

Press, Inc.

‘7,10,13,16-Docosatetraenoic acid (22:4), the chain-elongation product of arachidonic acid (20:4), is present in the adrenal, testes, brain, renal medulla, and fatty streak lesions of the aorta (l-3). Although a number of interesting observations have been made regarding docosatetraenoic

acid, its metabolic role has not been clearly defined. Docosatetraenoic acid accumulates in the ether phospholipids of chick embryo and Ehrlich ascites cells, especially in the alkyl ether and plasmalogen forms of the ethanolamine phosphoglycerides (4, 5). Rabbit renomedullary microsomes can convert docosatetraenoic acid to 22-carbonatom lipoxygenase and cyclooxygenase products, including 13-hydroxydocosatetraenoic acid, dihomoprostaglandins, and dihomothromboxanes (6). Likewise, human platelets convert docosatetraenoic acid to dihomothromboxane, 14-hydroxy-7,10,12nonadecatrienoic acid, and several hydroxydocosatetraenoic acids (7). However, a number of other tissues do not convert docosatetraenoic acid into these products (6), suggesting that these reactions are localized to only a few types of cells. In ad-

’ This paper is dedicated to the memory of Dr. Edward C. Heath. a Supported by Grant HL-14230 from the National Heart, Lung and Blood Institute, National Institutes of Health. ‘Present address: Department of Biochemistry, University of Texas Health Science Center, San Antonio, Tex. 78284. ’ To whom reprint requests should be addressed. ’ Abbreviations used: 224, 7,10,13-16-docosatetraenoic acid; 20:4, arachidonic acid; PGIa, prostaglandin Ia or prostacyclin; 6-keto-PGFi,, 6-keto-prostaglandin Fi, 813

0003-9861/86 Copyright All rights

$3.00

0 1986 by Academic Press, Inc. of reproduction in any form resewed.

814

MANN

dition, docosatetraenoic acid inhibits the conversion of arachidonic acid to prostaglandins in renal medullary homogenates (8) and to thromboxane in platelets (7). It also inhibits the proliferation of smooth muscle cells and skin fibroblasts (3, 9). Small amounts of docosatetraenoic acid normally are found in endothelial cell phospholipids, and the endothelium can convert arachidonic acid to docosatetraenoic acid (10-12). Arachidonic acid is used by the endothelial cells for prostaglandin synthesis, primarily prostacyclin (PG12), which prevents platelet aggregation and arterial vasoconstriction (13-16). Because of its relationship to arachidonic acid, we have examined the metabolism of docosatetraenoic acid in cultured endothelial cells and determined whether it has any role in endothelial prostaglandin formation. MATERIALS

AND

METHODS

Endothelial cell cultures. Bovine aorta endothelial cell cultures were initiated from vessels obtained from freshly slaughtered animals as described previously (17). Confluent monolayers were taken for study 4 to 6 days after seeding (1’7,18). The cells had a uniform appearance, demonstrated an epithelioid growth pattern, and were positive for Factor VIII antigen, angiotensin-converting enzyme, and plasminogen activator (11). For the experiments described here, the cultures were maintained at 37°C in medium-199 with Earle’s salts and 10% heat-inactivated fetal bovine serum in an atmosphere containing5% COa. This medium was enriched with fatty acids by adding a warm solution of the sodium salt (19), the pH was adjusted to 7.4 at 37°C and the medium was sterilized by filtration through 0.22-pm filters. Fatty acids were obtained from Nu-Chek Prep (Elysian, Minn.). Bovine endothelial cultures were studied between passage numbers 12 and 25. Incubutim and lipid analysis After the maintenance medium was removed, 2 ml of medium supplemented with fatty acid was added, and the incubation was carried out at 37°C with a 5% CO, atmosphere. [5,6,8,9,11,12,14,15,-nH]Arachidonic acid (135 Ci/ mmol), [1-‘*C]arachidonic acid (58 mCi/mmol), and [2-“Cldocosatetraenoic acid (40-60 mCi/mmol) were purchased from Amersham (Arlington Heights, Ill.). The cultures were incubated for varying periods with these isotopes, and the incubations were terminated by removing the medium and washing the monolayers with 2 ml of ice-cold Dulbecco’s phosphate-buffered saline solution containing 50 pM bovine serum albumin and then twice with 2 ml of the ice-cold buffer. Bovine

I
DOCOSATETRAENOIC

ACID

control incubations were done in culture dishes without cells. After incubation, the medium was removed, acidified to pH 3.5 with 1.2 M citric acid, and extracted twice with 2 vol of acetone-chloroform (l:l, v/v) (24). Separation of cyclooxygenase and lipoxygenase products was carried out by high-performance liquid chromatography on a Beckman 332 gradient system equipped with a 4.6 X 250-mm Vydac column containing Cl8 reverse-phase 5-pm spherical packing using a gradient of water adjusted to pH 3.4 with phosphoric acid and acetonitrile (25). Radioactivity was assayed by mixing the column effluent with 3a70B scintillator solution (Research Products International, Mt Prospect, Ill.) at a 1:3 ratio and passing the mixture through an on-line Radiomatic Instruments Flo-one/ Beta radioactive detector equipped with a 0.5-ml flow cell. Quenching was corrected and the data were integrated using the computer and software of Radiomatic Instruments (Tampa, Fla.). The system was standardized with a mixture of known eicosanoids and fatty acids. Background radioactivity obtained from corresponding incubations in the absence of cells was subtracted from the quantities recovered in the endothelial cell incubations. &Keto-PGF,, rodioimmunoassay. After the experimental medium was removed, the monolayer of intact, confluent endothelial cells was washed with Dulbecco’s buffer at 37°C and again with this solution containing 50 FM bovine albumin. The monolayers then were incubated for 5 min at 3’7°C with 1 ml of 50 pM bovine albumin in Dulbecco’s buffer as a control, or this solution containing either 10 pM calcium ionophore A23187 or arachidonic acid bound to bovine serum albumin. PGIa contained in the supernatant solution was measured by a radioimmunoassay for 6keto-prostaglandin Fi, (6-keto-PGFiJ, a stable catabolic product of PGIe (26). Assay detection limits were 0.3 pmol of 6-keto-PGFi,, and 50% inhibition was obtained with 3.3 pmol of 6-keto-PGF,,. This assay has negligibly small amounts of cross-reactivity with prostaglandins other than 6-keto-PGFi,.

AND

ENDOTHELIAL

815

CELLS

“C]docosatetraenoic acid was incubated in the absence of cells. Moreover, the formation of these products was not inhibited by the presence of 100 PM ibuprofen, 10 pM nordihydroguaiaretic acid, or 0.5 mM metyrapone. When the endothelial cultures were incubated with [1-14C]arachidonic acid, large quantities of radioactivity eluted with the 6-keto-PGFi, standard. Smaller amounts of radioactivity chromatographed with prostaglandin E2 and in the hydroxyeicosatetraenoic acid regions of the chromatogram eluting between 35 and 40 min. Prolonged incubations of the endothelial cultures with either arachidonic or docosatetraenoic acid also were done, followed by assay of the incubation medium for PGIz content (in the form of 6-keto-PGFi,) by radioimmunoassay. Figure 1 shows that when cells were incubated in a medium containing fetal bovine serum with supplemental arachidonic acid, PGIz was detected in the medium at the earliest time

166 -

0 2014 0 22~4

RESULTS

Prostaglandin production. Bovine aortic endothelial cultures were examined for their ability to produce prostaglandins when they were incubated for 20 min with either [2-14C]docosatetraenoic acid or [l“C]arachidonic acid. When the endothelial cultures were incubated with [2-14C]docosatetraenoic acid, small amounts of radioactivity were present in the prostaglandin and hydroxylated fatty acid regions of the high-performance liquid chromatogram. However, corresponding amounts were also recovered when [2-

0

5

10

Time

20

$

FIG. 1. PGIa accumulation in the medium during incubation of endothelial cultures with either arachidonic (20:4) or docosatetraenoic (22:4) acids. The culture medium contained 10% fetal bovine serum and 100 pM supplemental fatty acid. PGIa was measured by radioimmunoassay of the stable end product, 6keto-PGFi,. Each point is the mean of three values obtained from three separate cultures; only those SE bars that are large enough to be visible are shown.

816

MANN

of measurement and increased as the incubation continued. By contrast, no PGIz was detected during the first 8 h of incubation when the fetal bovine serum medium was supplemented with docosatetraenoic acid; thereafter, a small amount was produced and this increased as the incubation continued. As described below, the PGIB formed at the later times probably is due to retroconversion of the docosahexaenoic acid to arachidonic acid, which then served as the substrate for prostaglandin synthesis. Based on phase microscopic examination, there was no indication of cell injury or lysis during these or other prolonged incubations with the fatty acidsupplemented media. 3.5

3.0 0 20~4

i 3

2.5

1

2.0

kf

1.5

d x CA 1.0

0.S

0.0

J

1

1

1

0

1

1

I

1

0

1

1

I

0 2 4 6 8 10 12 14 15 12 20 22 24

Time

(h)

FIG. 2. Inhibition of PGIa production as a result of incubation of endothelial cultures with docosatetraenoic acid. The cultures were incubated for varying times with a medium containing 10% fetal bovine serum and 50 @M supplemental docosatetraenoic acid. After this medium was removed and the cultures were washed, they were incubated for 10 min with either 10 PM arachidonic acid (20:4) or 10 pM ionophore A23187. The PGIz produced was measured by radioimmunoassay of B-keto-PGFi,. Each value is the mean of results from three separate cultures + SE. Control cultures incubated for up to 24 h without supplemental docosatetraenoic acid produced amounts of PGIs within 510% of those tested at the start of the incubation period. The time value indicated on the z axis refers to the length of the initial incubation with supplemental docosatetraenoic acid.

ET

AI,.

The effect of enrichment of the endothelial cells with docosatetraenoic acid on their capacity to subsequently produce PGIz also was tested. In the experiment shown in Fig. 2, the cultures were incubated initially for 1 to 24 h with a fetal bovine serum medium containing 50 FM docosatetraenoic acid. After washing, a second incubation was done for 10 min with either 10 PM ionophore A23187 or 10 PM arachidonic acid. In every case more PG12 was produced when the cultures were incubated with arachidonic acid as opposed to ionophore A23187. As compared with control cultures, those initially incubated with supplemental docosatetraenoic acid produced much less PG12 when they were stimulated with either A23187 or arachidonic acid. Although some recovery occurred as the length of time of the initial incubation with docosatetraenoic acid increased, the amount of PG12 produced by the enriched cultures remained considerably less than in the corresponding controls. The effect of docosatetraenoic acid concentration on PGIz production was examined. Cultures were incubated for 16 h with media containing fetal bovine serum and 0 to 100 PM supplemental docosatetraenoic acid. After the 16-h incubation, these media were removed, and the cultures were washed and then stimulated with A23187. An initial incubation with 25 /IM docosatetraenoic acid produced little change in ionophore-stimulated PG12 release. Higher docosatetraenoic acid concentrations inhibited PGIz production, the inhibition reaching 45 to 50% at concentrations between 50 and 100 PM (Fig. 3). In one set of cultures the medium also contained 100 PM ibuprofen, a rapidly reversible cyclooxygenase inhibitor (27-29). The presence of ibuprofen during the incubation with docosatetraenoic acid did not prevent the decrease in PGIz production, suggesting that the inhibition is not due to cyclooxygenase inactivation. Formation arachidonic

of docosatetraenoic acid. Radioactive

acid from

docosatetraenoic acid was produced when endothelial cultures were incubated with 50 PM [5,6,8,9,11,12,14,15-3H]arachidonic acid

DOCOSATETRAENOIC

ACID

2.0

0

1.5

2 8 s.

2 1.0 b 4 3 4 (b

0.5

0.0

FIG. 3. Effect of docosatetraenoic acid concentration on endothelial PGIa production. The cultures were incubated with a medium containing 10% fetal bovine serum and 0 to 100 PM supplemental docosatetraenoic acid. In one set of cultures, the medium also contained 100 PM ibuprofen. This medium was removed after 16 h and the cultures were washed. The cultures then were incubated for 10 min with a medium containing 10 PM ionophore A23187. The amount of PGIc released into this medium was assayed by radioimmunoassay of 6-keto-PGF1,. Each point is the mean of values obtained from three cultures + SE.

AND

ENDOTHELIAL

with supplemental arachidonic acid. Complete fatty acid compositions were obtained by gas-liquid chromatography; for brevity only results for the relevant fatty acids are presented. Table I shows the changes in phospholipid composition during a 48-h incubation with 50 PM arachidonic acid bound to 10% fetal bovine serum. Prior to incubation, the endothelial phospholipids contained 10.5% 20:4 and 1.5% 22:4. The 20: 4 content increased to 13% within 0.5 h and subsequently to 18%. The 22:4 content of the phospholipids also increased, reaching 3% in 3 h and then progressively increasing to 9%, These increases were accompanied by reductions in other unsaturated fatty acid, the largest decrease being in oleic acid (18:l) from an initial value of 29 to 20%. The effect of increasing arachidonic acid concentration on phospholipid fatty acid composition also was examined. After 48

3

Phospholipida

1g

I

-

0100

bound to 10% fetal bovine serum. As shown in Fig. 4, there was a time-dependent accumulation of radioactivity in the cell phospholipids and triglycerides. These two cell lipid fractions accounted for more than 95% of the incorporated radioactivity. During the first 8 h, most of the accumulated radioactivity remained as 2014. Subsequently, an increasing fraction of the radioactivity was recovered as 22:4; 12% after 24 h and 28% after 48 h. The amount of newly incorporated radioactive arachidonic acid that was converted to 22:4 increased when the concentration of arachidonic acid in the culture medium was raised from 25 to 200 PM. Most of the increase occurred in the triglyceride fraction. Phospholipid fat@ acyl composition The fatty acid composition of the total phospholipids obtained from the endothelial cells was measured following incubation

817

CELLS

50

-

3 E6 so-

.022E4 20:4

/ i/ 40- .

T/

i

FIG. 4. Time-dependent incorporation of [5,6,8,9, 11,12,14,15,-*H]arachidonic acid into endothelial cell lipids. The incubation medium contained 50 fiM supplemental arachidonic acid added to 10% fetal bovine serum in modified medium-199 medium. The curve labeled 20:4 represents the amount of radioactivity recovered from the cell lipids as arachidonic acid; that labeled 22:4 is the amount elongated to docosatetraenoic acid. Each point is the mean of values obtained from three separate cultures; some SE bars are omitted because they are too small to be visible.

&l‘ AL.

N. “! -? v? 1 I T I 00000000 +I +I -+I +I +I +I +I +I t-tlcqo?oq~~o! NCVNWCUWIdr+

h of incubation, the phospholipid 20:4 content increased from an unsupplemented value of 11 to 26% when 100 pM arachidonic acid was present. There was a corresponding increase in 22:4 from 2 to 15%. These increases were accompanied by decreases in the content of other unsaturated fatty acids, the largest reduction again being in 18:l which declined from 32 to 17%. The changes in the fatty acyl compositions of the four main endothelial cell phosphoglyceride fractions also were determined after supplementation for 24 h with a medium containing fetal bovine serum supplemented with 50 PM arachidonic acid. The results are given in Table II. For brevity only data for 20:4 and 22:4 are presented. The inositol phosphoglycerides, which contained the highest 20:4 and 22~4 percentages when the endothelial cells were grown in unsupplemented medium, showed relatively small increases in these fatty acids when the medium contained supplemental arachidonic acid. By contrast, appreciable increases in both 20:4 and 22:4 occurred in the choline, ethanolamine, and serine phosphoglycerides when the culture medium was supplemented with arachidonic acid, with the highest percentage of 22~4 occurring in the ethanolamine fraction. Incorporation and retroumversion of docosatetraenoic acid. In addition to forming

docosatetraenoic acid from arachidonic acid, the bovine aortic endothelial cells were able to take up docosatetraenoic acid when it was available in the culture medium as a free fatty acid. Incorporation into the phospholipids and triglycerides of the endothelial cells increased as the concentration of docosatetraenoic acid was raised from 10 to 150 PM. At the lowest concentration, equivalent amounts were incorporated into both lipid fractions. As the concentration was raised, however, much more of the uptake was recovered in triglycerides. The large capacity to incorporate docosatetraenoic acid into triglycerides has been observed previously in Ehrlich ascites cells (5). The changes in the fatty acyl composition of the main endothelial cell phospholipids also were examined after a 24-h in-

DOCOSATETRAENOIC

ACID

AND

TABLE FATTY

Phospholipid

Inositol Choline Ethanolamine Serine

25.7 5.7 8.5 3.5

f + f f

Arachidonic 2214

3.0 1.5 0.7 1.5

FRACTIONS OF ENDOTHELIAL CELLS FOLLOWING OR DOCOSATETRAENOIC ACID” fatty

Unsupplemented* 20:4 3.5 0.7 1.3 1.1

+ + f f

1.0 1.5 0.8 1.5

819

CELLS

II

ACID COMPOSITION OF INDIVIDUAL PHOSPHOGLYCERIDE SUPPLEMENTATION WITH EITHER ARACHIDONIC

Phosphoglyceride fraction

ENDOTHELIAL

20:4 31.8 16.1 15.3 10.4

f + + *

acid composition acid”

Docosatetraenoic

22:4 1.6 1.3 3.0 1.5

6.2 7.6 14.5 10.8

(%)

f + + *

20~4 3.5 1.3 5.8 0.8

14.8 10.6 9.7 15.2

+ f f f

acidd 22:4

4.0 2.2 1.5 4.6

12.7 12.4 14.1 12.1

f f f k

3.8 4.1 0.6 3.2

(LAfter cultured bovine endothelial cells were incubated for 24 h, the cellular phospholipids were isolated by high-performance liquid chromatography. Following transmethylation, the fatty acyl composition was determined by gas-liquid chromatography. Values for all of the fatty acids contained in each phospholipid fraction were obtained; for clarity, only those for 20:4 and 224 are listed. Each value is the mean + SE of results from four separate cell cultures. *Medium contained 10% fetal bovine serum. ‘Medium contained 10% fetal bovine serum plus 50 pM arachidonic acid. d Medium contained 10% fetal bovine serum plus 50 pM docosatetraenoic acid.

cubation with 50 PM docosatetraenoic acid. was accompanied by an increase (nonstoiin the phospholipid 20:4 conTable II shows that each of the four phos- chiometric) phoglyceride fractions exhibited a sizable tent. Because the incubation media contained increase in 22:4 content. In the inositol phosphoglycerides, the increase in 224 was 10% fetal bovine serum in these experiaccompanied by a 40% reduction in 20:4 ments, it was not possible to determine content. No appreciable change in the 20:4 with certainty that these increases in 20~4 content occurred when the ethanolamine were due to a retroconversion of docosaphosphoglycerides became enriched with tetraenoic acid to arachidonic acid. To test 22:4. In the choline and serine phosphodirectly for retroconversion, the bovine glycerides, however, the increase in 22:4 endothelial cultures were incubated with was associated with a large increase in 20: biosynthetically prepared [3H]docosatet4 percentage, suggesting the possibility of raenoic acid, the cell lipids were transretroconversion of 22:4 to 20:4. methylated, and the effluent from the gasEvidence consistent with retroconverliquid chromatograph was collected ussion also was obtained by gas-liquid chro- ing a stream splitter (9) to measure the matographic analysis of the total phos- radioactivity present in individual fatty pholipid fatty acid composition in a pulse- acids. In the first experiment, 21.7% of chase experiment. The results are shown the [7,8,10,11,13,14,16,17,-3H]docosatetrain Table III. Again, for brevity only the enoic acid incorporated into cell lipids durresults for 20:4 and 22:4 are presented. Af- ing a 48-h incubation was present in the ter an 18-h exposure of the endothelial cul- form of radioactive 20:4; 20.0% was recovtures to a medium containing fetal bovine ered as 20:4 in a second experiment. Only serum and 50 PM supplemental docosatet1.6 and 0.5% of the radioactivity was reraenoic acid, the incubations were contincovered in fatty acid eluting before 20:4 in ued without fatty acid supplementation. these experiments. The phospholipids were highly enriched in Release of fatty acid from endothelial 22:4 after the initial exposure. Subse- cells. The capacity of the bovine endothelial quently, the 22:4 content declined, and this cells to release arachidonic and docosatet-

MANN

TABLE

III

CHANGES IN ENDOTHELIAL CEIL PHOSPHOLIPID FAIT ACID COMPOSITION FOLLOWING ENRICHMENT WITH DOCOSATETRAENOIC ACID” Phospholipid fatty composition”

acid

Condition

20:4

22~4

Enriched cultures After 24 h incubation After 48 h incubation

14.2 + 0.3 14.8 + 0.3 17.2 + 0.3

17.6 f 0.5 11.5 + 0.3 10.5 f 0.4

a Twelve endothelial cultures were incubated for 18 h with a medium containing 10% fetal bovine serum and 50 jtM docosatetraenoic acid. After removal of this medium and washing, four cultures were taken for analysis of the phospholipid fatty acid composition by gas-liquid chromatography. The remaining eight cultures were incubated for either 24 or 48 h in a medium containing no supplemental fatty acid, and then taken for phospholipid fatty acid compositional analysis. Each value is the mean f SE of results obtained from four separate cultures. Values for all of the phospholipid fatty acids were obtained but, for clarity, only those for 20:4 and 224 are listed.

raenoic acids was compared. Cultures were loaded with either [l-14C]arachidonic acid or [2-14C]docosatetraenoic acid. As seen in Table IV, more radioactive arachidonic acid was incorporated under these conditions. The distribution of radioactivity among the main cellular phosphoglyceride classes, however, was similar in both cases. With [l-‘4C]arachidonic acid, the distribution in choline, inositol, ethanolamine, and serine phosphoglycerides was 68, 15, 10, and 8%; with [2-14C]docosatetraenoic acid, the distribution was 63,18, 7, and 12%. A second incubation was carried out to determine the amount of the incorporated radioactivity that the cells could release into the extracellular fluid. Except for changing the medium, the cultures were not perturbed or stimulated in any way. In the case of both fatty acids, a considerable amount of radioactivity was released during the lomin incubation with albumin alone. This presumably is due to a basal rate of turnover of the cellular phospholipids (30), especially the small pools that contain much of the newly incorporated fatty acid

ET

AL.

(31). A larger amount of fatty acid was released when the medium contained ionophore A2318’7. The increment over the release occurring in the presence of albumin alone is considered as the stimulated release; this amounted to 15.6% of the initial uptake in the cells labeled with arachidonic acid, and 12.9% in the cells labeled with docosatetraenoic acid. Analysis by thinlayer chromatography revealed that, in both cases, at least 80% of the radioactivity released to the medium when the ionophore was present was in the form of free fatty acid. DISCUSSION

These results demonstrate that endothelial cells can obtain docosatetraenoic acid either by synthesis from arachidonic acid or by uptake from an extracellular source. Unlike platelets and the renal medulla (7, 8), the endothelial cells do not convert appreciable amounts of [2-14C]docosatetraenoic acid to any radioactive dihomoprostaglandins or hydroxylated derivatives. Although some PGIz was formed during long-term incubations with docosatetraenoic acid, its appearance was delayed for 16 h (Fig. 1). By contrast, PGIB was detected within 1 h, the earliest time tested, during a corresponding incubation with arachidonic acid. Some retroconversion of docosatetraenoic acid to arachidonic acid already occurs by 16 h, suggesting that this is the source of the arachidonic acid from which the PGIz was formed during the prolonged incubation with docosatetraenoic acid. Our failure to observe dihomoeicosanoid formation in the endothelium is consistent with studies in the rabbit, indicating that the conversion of docosatetraenoic acid to these compounds is largely restricted to the kidney (6). The fact that it can be formed from arachidonic acid by a cell that does not produce dihomoeicosanoids suggests that docosatetraenoic acid may have other metabolic roles. One such role in endothelium may be to serve as a supplementary source of arachidonic acid. Through retroconversion, a portion of the docosatetraenoic acid present in endothelial lipids can be converted back to arachidonic acid. This capacity may

DOCOSATETRAENOIC

ACID TABLE

AND

ENDOTHELIAL

821

CELLS

IV

RELEASEOFRADIOACTIVE FATTYACIDSFROM ENDOTHELIAL CELLS' Radioactivity

Parameter Incorporation

Units

into cell lipids

Release to the medium Albumin Albumin plus A23187 Stimulated release Calculated Relative to initial content

[1-“ClArachidonic acid

[2-“ClDocosatetraenoic acid

dpm

409,500

k 47,600

120,000

dpm dpm

40,200 104,000

f 4600 + 11,000

27,500 43,000

f 10,800

f f

dpm

63,800

15,500

%

15.6

12.9

1700 4,800

cell

a Cultured endothelial cells were labeled for 15 min at 37°C with 0.75 pCi of either [1-%]arachidonic acid or [2-%]docosatetraenoic acid and then washed. Five cultures from each group were taken for analysis of the amount of radioactivity incorporated into cell lipids. The remaining cultures were incubated for 10 min at 37°C with a medium containing either 50 pM albumin or 50 pM albumin plus 10 HIMionophore A23187, and the amount of radioactivity released to the medium was determined. Each value is the mean f SE of results from four separate cultures.

be especially important to the endothelium because it has a low capacity to convert linoleic acid, the usual precursor, to arachidonic acid (10, 11,19). Docosatetraenoic acid retroconversion has been observed previously in three other systems. Rat liver retroconverts docosatetraenoic acid to arachidonic acid (32), while smooth muscle cells and fibroblasts retroconvert docosatetraenoic acid (22:4) to both 20:4 and octedecatrienoic acid (l&3) (33). Smooth muscle cells also retroconvert eicosatrienoic acid (20:3) to l&3 (33). Since only a small amount of 22:4 radioactivity was recovered in a fraction eluting before 20:4 it appears that retroconversion to l&3 does not occur to any appreciable extent in the endothelium. A number of other tissues also can carry out retroconversion; for example, rat testis retroconverts docosapentaenoic acid (225) to 20:4 (34,35), rat liver retroconverts docosahexaenoic acid (22:6) to 22:5 and eicosapentaenoic acid (20:5) (36), and human retinoblastoma cells retroconvert 22:6 to 20:5 (37). It is not known whether any of these additional retroconversions can take place in endothelial cells. The capacity of the endothelial cells to

produce PGIB was reduced when the cultures were incubated with docosatetraenoic acid, a finding similar to that observed in rabbit kidney (6) and human platelets (7). When endothelial cells are incubated with high concentrations of arachidonic acid under conditions similar to these, the cyclooxygenase is inactivated (29). Therefore, one possibility that was considered is that docosatetraenoic acid also inhibits PGIz formation by inactivating the cyclooxygenase reaction. Inactivation by arachidonic acid is prevented if ibuprofen, a reversible cyclooxygenase inhibitor (27, 28), is present during the incubation (20). As opposed to what was observed previously with arachidonic acid (29), the presence of ibuprofen did not prevent the inhibition produced by exposure to high concentrations of docosatetraenoic acid, indicating that the mechanism of inhibition probably does not involve cyclooxygenase inactivation. Since ibuprofen may have additional effects in endothelium, this interpretation is somewhat tenuous. It was not possible to use a more specific cyclooxygenase inhibitor such as aspirin, however, because the experiments required that the inhibi-

822

MANN

tion be rapidly reversible so that PGIz production could be tested immediately after it was removed. A more likely mechanism for the decreased PGIz output is that because of its structural similarity to arachidonic acid, docosatetraenoic acid may act as an inhibitor of cyclooxygenase. Other polyunsaturated fatty acids have been shown to inhibit cyclooxygenase activity (38) and thereby reduce the formation of prostaglandin Ez (39), thromboxane A2 (40-43), and PGIz (44, 45). Like arachidonic acid, docosatetraenoic acid was released when the endothelial cells were incubated with A23187 (Table IV), probably contributing to the inhibition that occurred when the cells were stimulated with this ionophore (Figs. 2 and 3). However, less PGIz formation also occurred when the cells were incubated with arachidonic acid (Fig. 2). No stimulus was added to release docosatetraenoic acid from cellular lipids under these conditions, but the data in Table IV suggest that some continuing basal release occurs, presumably due to phospholipid remodeling and turnover (30, 31). Enough unesterified docosatetraenoic acid apparently is generated to act as an inhibitor in this situation. Another factor that could play a role in decreasing the ionophore-stimulated PGIz output by the docosatetraenoic acidtreated cells is the reduced arachidonic acid content of the inositol phosphoglycerides (Table II). The fact that it is the only phospholipid class containing less arachidonic acid strengthens this possible association. Recent evidence suggests that the inositol phosphoglycerides may be the source of the arachidonic acid utilized for prostaglandin production (46, 4’7), including PGIz formation in endothelial cells (48,49). This is a controversial point, however, for other phospholipid classes also appear to be involved (50), such as the choline phosphoglycerides in endothelial cells (51). When docosatetraenoic acid was formed intracellularly, no decline in the arachidonic acid content of the inositol phosphoglycerides occurred (Table II). The ability to maintain a high arachidonic acid content in the inositol fraction, which facilitates

ET

AL.

the operation of the phosphatidylinositol cycle (52), apparently is lost when the endothelial cell is faced with a large influx of preformed docosatetraenoic acid. REFERENCES 1. VAHOUNY, G. W., HODGES, V. A., AND TREADWELL, C. R. (1979) J. Lipid Res. 20,154-161. 2. COMAI, K., FARBER, S. J., AND PALJLSRUD, J. R. (1975) Lipids 10,555-561. 3. GAVINO, V. W., MILLER, J. S., DILLMAN, J. M., MILO, G. E., AND CORNWELL, D. G. (1981) Prog. Lipid Res. 20,323-325. 4. NAKAGAWA, Y., WAKU, K., AND ISHIMA, Y. (1982) B&him. Biophys. Acta 712,667-676. 5. MASUZAWA, Y., NAKAGAWA, Y., WAKU, K., AND LANDS, W. E. M. (1982) Biochim Biophys. Acta 713,185-192. 6. SPRECHER, H., VAN ROLLINS, M., SUN, I., WYCHE, A., AND NEEDLEMAN, P. (1982) J. Biol. Chem. 257,3912-3918. 7. VAN ROLLINS, M., HORROCKS, L., AND SPRECHER, H. (1985) Biochim Biophys. Acta 833,272-280. 8. CAGEN, L. M., AND BAER, P. G. (1980) Life Sci. 26, 765-770. 9. MORISAKI, N., SPRECHER, H., MILO, G. E., AND CORNWELL, D. G. (1982) Lipids 17,893-899. 10. SPECTOR, A. A., KADUCE, T. L., HOAK, J. C., AND FRY, G. L. (1981) J. Clin. Invest. 68,1003-1011. 11. KADUCE, T. L., SPECTOR, A. A., AND BAR, R. S. (1982) Arteriosclerosis 2,380-389. 12. ROSENTHAL, M. D., AND WHITEHURST, M. C. (1983) B&him. Biophys. Acta 750,490-496. 13. MONCADA, S., GRYGLEWSKI, R., BUNTING, S., AND VANE, J. R. (1976) Nature (London) 263, 663665. 14. WEKSLER, B. B., MARCUS, A. J., AND JAFFE, E. A. (1977) Proc. Natl. Acad. Sci USA 74,3922-3926. 15. MARCUS, A. J., WEKSLER, B. B., AND JAFFE, E. A. (1978) J. BioL Chem. 253,7138-7141. 16. MONCADA, S. (1982) Arteriosclerosis 2,193-207. 17. GOLDSMITH, J. C., JAFVERT, C. T., LOLLAR, P., OWEN, W. G., AND HOAK, J. C. (1981) Lab. Invest. 45,191-197. 18. PEACOCK, M. L., BAR, R. S., AND GOLDSMITH, J. C. (1982) Metabolism 31,52-56. 19. SPECTOR, A. A., HOAK, J. C., FRY, G. L., DENNING, G. M., STOLL, L. L., AND SMITH, J. G. (1980) J. Clin Invest. 65,1003-1012. 20. LEES, M. B., AND PAXMAN, S. (1972) Anal Biochem 47,184-192. 21. FOLCH, J. M., LEES, M., AND SLOANE STANLEY, G. H. (1957) J. Biol Chem 226,497-509. 22. KADUCE, T. L., NORTON, K. C., AND SPECTOR, A. A. (1983) J. Lipid Res. 24,1398-1403. 23. MORRISON, W. R., AND SMITH, L. M. (1964) J. Lipid Res. 5.600-608. 24. SALMON, J. A., AND FLOWERS, R. J. (1982) in Meth-

DOCOSATETRAENOIC

ACID

AND

ods in Enzymology (Lands, W. E. M., and Smith, W. L., eds.), Vol. 86, pp. 47’7-493, Academic Press, New York. 25. ELING, T., TAINER, B., ALLY, A., AND WARNOCK, R. (1982) in Methods in Enzymology (Lands, W. E. M., and Smith, W. L., eds.), Vol. 86, pp. 511-517, Academic Press, New York. 26. CZERVIONKE,R. L., SMITH, J. B., HOAK, J. C., FRY, G. L., AND HAYCRAFT, D. L. (1979) Thromb. Res. 14,781-786. 27. MCINTYRE, B. A., AND PHILP, R. B. (1977) Thromb. Res. 12, 67-77. 28. PARKS, W. M., HOAK, J. C., AND CZERVIONKE, R. L. (1981) J. Pharmacol. Exp. Ther. 219,415419. 29. SPECTOR,A. A., KADUCE, T. L., HOAK, J. C., AND CZERVIONKE,R. L. (1983) Arteriosclerosis 3,323331. 30. D’SONZA, C. J. M., CLARKE, J. T. R., COOK, H. W., AND SPENCE, M. W. (1983) B&him. Biophys. Acta 752, 467-473. 31. SPECTOR,A. A., AND STEINBERG,D. (1967) J. BioL Chem. 242,3057-3062. 32. STOFFEL, W., ECKER, W., ASSAD, H., AND SPRECHER,H. (1970) Hoppe-Seyler’s 2. Physiol. Chem. 35,1545-1554. 33. GAVINO, V. C., MILLER, J. S., DILLMAN, J. M., MILO, G. E., AND CORNWELL,D. G. (1981) J. Lipid Res. 22,57-62. 34. BRIDGES,R. B., AND CONIGLIO,J. G. (1970) B&him Biophys. Acta 218, 29-35. 35. SCHLENK, H., GELLERMAN, J. L., AND SAND, D. M. (1967) Biochim. Biophys. Acta 137,420-426. 36. SCHLENK, H., SAND, D. M., AND GELLERMAN,J. L. (1969) B&him. Biophys. Acta 187,201-207. 37. YOREK, M. A., BOHNKER, R., DUDLEY, D., AND

38. 39. 40.

41.

42.

43.

44. 45.

46. 47. 48. 49. 50. 51.

52.

ENDOTHELIAL

CELLS

823

SPECTOR,A. A. (1984) B&him. Biophys. Acta 795,277-285. PACE-ASCIAK, C., AND WOLFE, L. S. (1968) Biochim. Biophys. Acta 152,784-787. ZIBOH, V. A. (1973) J. Lipid Res. 14,377-384. NEEDLEMAN, P., WYCHE, A., LEDUC, L., SANKARAPPE, S. K., JAKSCHIK,B. A., AND SPRECHER,H. (1981) Prog. Lipid Res. 20,415-422. SIESS, W., SCHERER, B., BOHLIG, B., ROTH, P., KURZMANN, I., AND WEBER, P. C. (1980) Lancet 1,441-444. NEEDLEMAN, P., WHITAKER, M. O., WYCHE, A., WATERS, K., SPRECHER,H., AND RAY, A. (1980) Prostaglandins 19,165-181. HORNSTRA,G., CHRIST-HAZELHOF,E., HADDEMAN, E., TEN HOOR, F., AND NUGTEREN, D. H. (1981) Prostaglandins 21, 727-738. MORITA, I., SAITO, Y., CHANG, W. C., AND MUROTA, S. (1983) Lipids l&42-49. SPECTOR, A. A., KADUCE, T. L., FIGARD, P. H., NORTON, K. C., HOAK, J. C., AND CZERVIONKE, R. L. (1983) J. Lipid Res. 24,1595-1604. BELL, R., AND MAJERUS,P. W. (1980) J. Bid Chem. 255,1790-1792. NISHIZUKA, Y. (1984) Science 225,1365-1370. HONG, S. L., AND DEYKIN, D. (1982) J. Biol Chem. 257,7151-7154. HONG, S. L., MCLAUGHIN, N. J., TZENG, C. Y., AND PATTON, G. (1985) Thromb. Res. 38,1-10. BROCKMAN, M. J., WARD, J. W., AND MARCUS, A. J. (1981) J. BioL Chem 256,8271-8274. THOMAS, J. M. F., HULLIN, F., CHAP, H., AND DOUSTE-BLAZY,L. (1984) Thromb. Res. 34,117123. SUGIURA, T., AND WAKU, K. (1984) Biochim Biophys. Acta 796,190-198.