Generation of 8-Epiprostaglandin F by Human Monocytes

F-isoprostanes are free radical-catalyzed products of arachidonic acid. One of these compounds, 8-epiprostaglandin F (8-epi-PGF), is a mitogen and vasoconstrictor. We have shown that 8-epi-PGF, unlike other F-isoprostanes, is a minor product of the prostaglandin endoperoxide synthase-1 (PG G/H S-1) expressed in human platelets (Praticó, D., Lawson, J. A., and FitzGerald, G. A.(1995) J. Biol. Chem. 270, 9800-9808). Human monocytes express PG G/H S-1 constitutively and exhibit regulated expression of PG G/H S-2. Induction of PG G/H S-2 by concanavalin A, the phorbol ester, phorbol 12-myristate 13-acetate, and bacterial lipopolysaccharide was confirmed with a specific antibody in monocytes pretreated with aspirin to inhibit PG G/H S-1. Induction of PG G/H S-2 by all three stimuli coincided with increased formation of prostaglandin E (PGE), thromboxane B (TxB), and 8-epi-PGF, but not of other F-isoprostanes. Confirmation of PG G/H S-2 as the source of 8-epi-PGF formation was obtained by down-regulating the enzyme with dexamethasone; preventing protein synthesis with cycloheximide; and preventing synthesis of PGE, TxB, and 8-epi-PGF with the specific PG G/H S-2 inhibitor, L 745,337. Monocytes also exhibit the facility to generate 8-epi-PGF in a free radical-dependent manner. Thus, stimulation with opsonized zymosan or coincubation with low density lipoprotein was unassociated with product formation. However, coincubation of low density lipoprotein with zymosan-stimulated human monocytes resulted in marked formation of 8-epi-PGF, but not of PGE or TxB. Production of 8-epi-PGF coincided with that of thiobarbituric acid-reactive substances and lipid hydroperoxides, but was unaccompanied by PG G/H S-2 induction. Pretreatment of monocytes with the antioxidant, butylated hydroxytoluene or with superoxide dismutase, but not with L 745,337, suppressed formation of 8-epi-PGF, thiobarbituric acid-reactive substances, and lipid hydroperoxides. In conclusion, human monocytes may form bioactive 8-epi-PGF either via free radical- or enzyme-catalyzed pathways. 8-Epi-PGF is a more abundant product of monocyte PG G/H S-2 than of platelet PG G/H S-1. Formation by inducible PG G/H S-2 must be considered as a source of this compound in vivo.

Monocytes are thought to play a central role in atherogenesis (1). They adhere to and transmigrate between endothelial cells (2). Monocyte-derived macrophages have high affinity receptors for oxidized low density lipoprotein (LDL), 1 and LDL ingestion results in their transformation into foam cells (3,4). Discharge of the contents of foam cells, themselves marked constituents of fatty streaks (5), is thought to contribute to formation of atherosclerotic plaque (6). Monocytes and macrophages exhibit the ability to transform arachidonic acid to prostaglandins and related compounds via the enzyme prostaglandin endoperoxide synthase (PG G/H S). There are two forms of this enzyme (7,8). Monocytes express PG G/H S-1 constitutively (9), and PG G/H S-2 is up-regulated in response to cytokines, growth factors, and bacterial lipopolysaccharide (LPS) (9 -11). It is unknown what role, if any, prostanoids may play in atherogenesis, although they have been implicated in regulating cellular proliferation (12,13), vascular tone and permeability (14,15), and aspects of cholesterol metabolism (16). Additional to its susceptibility to enzyme-catalyzed metabolism to biologically active compounds, arachidonic acid may be subject to free radical attack, leading to formation of prostaglandin isomers in situ in the cell membrane phospholipid (17,18). These isoprostanes may exert biological effects intra-or extracellularly. A prostaglandin F 2 isomer, 8-epi-PGF 2␣ , induces vasoconstriction and cellular proliferation, effects that are prevented by thromboxane A 2 receptor antagonists (19,20). Recently, we have shown that this compound is also a minor product of PG G/H S-1 in human platelets (21).
This study demonstrates that PG G/H S-2 may also form 8-epi-PGF 2␣ . Interestingly, it is a more abundant product of this enzyme in monocytes than of the PG G/H S-1 isoform in platelets. Monocytes also retain the ability to form this compound in a free radical-catalyzed manner. Indeed, coincubation of monocytes with LDL results in a time-dependent formation of 8-epi-PGF 2␣ , coincident with LDL oxidation. Isolation and Stimulation of Human Monocytes-Mononuclear cells were obtained from fresh peripheral blood of healthy volunteers, who did not take any medication during the previous 2 weeks. Blood was subjected to Ficoll-Hypaque density gradient centrifugation as described by Boyum (24). The mononuclear cell layer was recovered and washed with Hanks' balanced salt solution. The cells were resuspended in RPMI 1640 medium containing 10% inactivated fetal calf serum and EDTA for 15 min at 37°C to remove platelets specifically adherent to the monocytes. They were washed twice in Hanks' balanced salt solution and resuspended at 1 ϫ 10 6 ml in RPMI 1640 tissue culture medium supplemented with L-glutamine (220 mg/l) and antibiotics (streptomycin and penicillin). More than 94% of the cells were estimated to be viable based on trypan blue dye exclusion.

Materials-LPS
The cells were plated in 24-well multiplates and maintained at 37°C in a temperature-controlled, humidified 95% air, 5% CO 2 incubator. The nonadherent cells were removed by washing the plates twice with Dulbecco's phosphate-buffered saline after 2 h of incubation, and the adherent cells were maintained in RPMI 1640 medium enriched with 5% inactivated fetal calf serum, L-glutamine, and antibiotics. The resultant harvested adherent cells routinely contained Ͼ92% monocytes as determined by morphology and staining for nonspecific esterase (25). The contribution of PG G/H S-1 activity to eicosanoid production in response to the different stimuli was suppressed by pretreating the blood with aspirin (10 g/ml) in the sampling syringe and at time 0. This was confirmed by product analysis. Isolated monocytes were incubated for 1, 4, 8, and 24 h in the absence and presence of LPS (10 g/ml), concanavalin A (10 g/ml), or PMA (100 nM). Fresh medium containing 10 M arachidonic acid was added at each time point after removing the incubation medium; the supernatants were harvested after 30 min.
Concentrations of PGE 2 , TxB 2 , and 8-epi-PGF 2␣ were determined by gas chromatography/mass spectrometry (GC/MS) as described previously (21,26). The contribution of induced PG G/H S-2 to formation of these products was elucidated by use of a protein synthesis inhibitor, cycloheximide (27); a specific inhibitor of this isoform of the enzyme, L 745,337 (28); or dexamethasone, which has previously been shown to down-regulate PG G/H S-2 (29).
GC/MS Analysis of PGE 2 , TxB 2 , and 8-Epi-PGF 2␣ -Briefly, a 30-m DB-1 capillary column was used for analysis of all products. The temperature program was 190 -320°C at 20°C/min. The ions monitored were m/z 614 for TxB 2  Lipoprotein Preparation-LDL was prepared according to previously described methods that minimize oxidation and exposure to endotoxin (30). Each batch of LDL was assayed for endotoxin contamination by the Limulus amebocyte lysate assay. Final endotoxin contamination was always Ͻ0.02 unit/mg of LDL cholesterol. LDL was stored in 0.5 mM EDTA. Immediately before use, LDL was dialyzed at 4°C against phosphate-buffered saline without calcium or magnesium in the dark. LDL was used at final concentration of 0.40 mg of protein/ml.
LDL Incubation with Human Monocytes-Human monocytes were washed with RPMI 1640 medium without serum, plated into 12-well tissue culture plates (1 ϫ 10 6 /ml/well), and co-cultured for 24 h with LDL at 0.40 mg of protein/ml in the presence or absence of opsonized zymosan (2 mg/ml) according to the method of Johnston (31). LDL was extracted by adding 2 volumes of Folch reagent (chloroform and methanol at a 2:1 ratio) and base-hydrolyzed (1.0 M KOH) before quantitation as described above to determine the total 8-epi-PGF 2␣ level in the supernatant.
Measurement of Lipid Peroxidation-Briefly, the presence of lipid oxidation products on LDL was detected spectrophotometrically by measuring the thiobarbituric acid-reactive substance (TBARS) levels monitored at 532 nm (32). The lipid hydroperoxide levels were measured using the FOX 2 assay at 560 nm (33).
Twenty micrograms of total cell protein were analyzed by SDS-polyacrylamide gel electrophoresis. Acrylamide (8 and 4%) was used for the separating and staking gels, respectively. The resolved proteins were transferred onto nitrocellulose membranes. Blots were saturated overnight at 4°C with a solution of 5% fat-free dried milk in Tris-buffered saline and incubated with either mouse polyclonal anti-PG G/H S-2 antibody (1 g/ml) or specific polyclonal anti-PG G/H S-1 antibody for 2 h at room temperature. The membranes were then washed extensively with phosphate-buffered saline/Tween 20 and incubated with horseradish peroxidase-conjugated anti-mouse IgG at 1:2000 for 1 h at room temperature. Chemiluminescence substrates were used to reveal positive bands according to the manufacturer's instructions, and bands were visualized after exposure to Hyperfilm ECL (Amersham Corp.).
Statistical Analysis-Results are expressed as mean Ϯ S.D. Statistical comparisons were made by using analysis of variance with subsequent application of Student's t test as appropriate.

PG G/H S-2-dependent Formation of 8-Epi-PGF 2␣ -Incuba-
tion of human monocytes with 10 g/ml LPS resulted in a time-dependent increase of prostaglandin production (Fig. 1). LPS did not significantly affect eicosanoid production at 1 h, but caused a statistically significant increase at 4, 8, and 24 h of incubation. Maximal stimulation of 8-epi-PGF 2␣ , PGE 2 , and TxB 2 at 24 h of incubation was in the concentration ratio of 1:5.3:200. To evaluate whether LPS-induced production of these compounds was dependent on induction of PG G/H S-2, we studied the effects of dexamethasone (2 M) and cycloheximide (5 g/ml). Both compounds markedly suppressed the production of the three prostaglandins (Table I), coincident with suppression of the immunoreactive 72-kDa doublet recognized by the monoclonal anti-PG G/H S-2 antibody on Western blot analysis (Fig. 2). The band recognized by the anti-PG G/H S-1 antibody was unaltered by LPS, cycloheximide, or dexamethasone (data not shown). Confirmation that product formation was dependent on PG G/H S-2 induction was indicated by the specific inhibition of that enzyme. L 745,337 dose-dependently suppressed PGE 2 , 8-epi-PGF 2␣ , and TxB 2 with the same IC 50 value (48 Ϯ 10 nM). Unlike cycloheximide and dexamethasone, the effects of L 745,337 were not accompanied by down-regulation of PG G/H S-2 protein (data not shown). We observed a selective increase in a peak at m/z 695, comigrating with the 8-epi-PGF 2␣ internal standard (m/z 699). It was unaccompanied by other peaks, presumably reflecting other F 2 -isoprostanes, and it was suppressed by cycloheximide (Fig. 3). These data are consistent with enzyme-rather than free radicalcatalyzed formation of 8-epi-PGF 2␣ under these conditions. Similar results were obtained with concanavalin A (10 g/ ml) and PMA (100 nM). Both compounds caused a time-dependent increase on 8-epi-PGF 2␣ , PGE 2 , and TxB 2 . The respective ratios of product formed at 24 h of incubation were 1:5.3:160 for concanavalin A (Fig. 4) and 1:5:130 for PMA (Fig. 5). Again, both concanavalin A (Fig. 6A) and PMA (Fig. 6B) caused induction of PG G/H S-2 protein, which was suppressed, together with product formation, by cycloheximide and dexamethasone (Tables II and III Free Radical-catalyzed Formation of 8-Epi-PGF 2␣ by Human Monocytes-Activation of human monocytes with opsonized zymosan contrasted with the other stimuli in that it did not result in formation of 8-epi-PGF 2␣ , PGE 2 , or TxB 2 . Similarly, coincubation of monocytes with LDL, a source of substrate for lipid peroxidation (30), did not induce release of any of these products into the supernatant. Moreover, no significant changes in the levels of TBARS or lipid hydroperoxides were observed in human monocytes incubated with LDL in the absence of zymosan. This is in agreement with previous observations that activation is required for monocytes to oxidize LDL (34).
Activation of human monocytes with opsonized zymosan in the presence of LDL caused a marked time-dependent increase in 8-epi-PGF 2␣ , but not in PGE 2 or TxB 2 . The increase in 8-epi-PGF 2␣ was associated with an increase in TBARS and hydroperoxide levels (Table IV). This phenomenon was prevented by the oxygen free radical scavengers superoxide dismutase (300 units/ml) and BHT (20 M), but not by L 745,337 (100 nM) (Fig. 7) or by the nonselective inhibitor of PG G/H S, aspirin (data not shown). The quantities of 8-epi-PGF 2␣ formed under these conditions, in the presence of an excess of lipid (LDL) substrate, are not comparable with those formed in isolated stimulated human monocytes in the earlier experiments. In contrast to the data presented in Fig. 3, an array of peaks corresponding to F 2 -isoprostanes was evident in the supernatants of zymosan-stimulated human monocytes incubated with LDL (Fig. 8, center panel). One of these corresponds to the retention time of the 8-epi-PGF 2␣ standard (m/z 699) (Fig. 8, upper panel). Pretreatment with antioxidants suppressed the peaks, reflecting free radical-catalyzed isoprostane formation under these experimental conditions (Fig. 8, lower  panel). The retention time of authentic PGF 2␣ is well away from that of the F 2 -isoprostanes. Finally, coincubation of zymosan-activated human monocytes with LDL did not result in induction of PG G/H S-2 protein (data not shown). DISCUSSION 8-Epi-PGF 2␣ is an abundant isoprostane formed in response to free radical attack on arachidonic acid (17). Urinary excretion of 8-epi-PGF 2␣ is increased in syndromes putatively asso-  /ml), and L 745,337 (100 nM) were added at the same time as LPS (10 g/ml). Prostanoid production in the supernatant was determined at the end of a 24-h incubation period (n ϭ 4). ciated with increased oxidant stress in vivo, including paracetamol and paraquat poisoning (35), cigarette smoking (36), and vascular reperfusion (37). We have previously shown that it is a minor product of the PG G/H S-1 expressed in human platelets (21). However, this route of formation appears to contribute trivially to overall 8-epi-PGF 2␣ biosynthesis, even in a setting such as chronic cigarette smoking (36), in which platelet activation (38) is thought to coincide with increased oxidant stress.

FIG. 2. Western blot analysis of PG G/H S-2 protein expression in monocytes treated
Human monocytes contain PG G/H S-2 additional to the isoform expressed in platelets. This form of the enzyme is highly regulated by cytokines, LPS, hormones, and mitogenic factors (9 -11). It is thought to be the source of prostanoid production in inflammatory states (39) and, perhaps, in cancer (40). The primary sequences of both human enzymes exhibit 60% similarity (8, 41); however, mutational analysis suggests that the active site of PG G/H S-2 is the more accommodating.  a Dexamethasone (2 M), cycloheximide (5 g/ml), and L 745,337 (100 nM) were added at the same time as concanavalin (10 g/ml). Prostanoid production in the supernatant was determined at the end of a 24-h incubation period (n ϭ 4). a Dexamethasone (2 M), cycloheximide (5 g/ml), and L 745,337 (100 nM) were added at the same time as PMA (100 nM). Prostanoid production was determined in the supernatant at the end of a 24-h incubation period (n ϭ 4). a Peripheral blood monocytes were incubated with LDL (400 g of protein/ml) and opsonized zymosan (2 mg/ml). Superoxide dismutase (300 units/ml), BHT (20 M), and L 745,337 (100 nM) were added at the same time as LDL and zymosan. For example, although aspirin is a relatively nondiscriminate PG G/H S inhibitor (42,43), inhibition of PG G/H S-2, but not of PG G/H S-1, by aspirin is associated with increased production of (15R)-hydroxyeicosatetraenoic acid, coincident with inhibition of prostanoid formation (44). The present studies demonstrate that 8-epi-PGF 2␣ may be formed in a PG G/H S-2dependent manner. Using three independent stimuli, formation coincided in time with the kinetics of the de novo synthesis of PG G/H S-2. Furthermore, reduction in activity of the induced enzyme, by either cycloheximide or steroids, prevented 8-epi-PGF 2␣ formation coincident with that of the conventional PG G/H S-2 products of these cells, PGE 2 and thromboxane A 2 . A selective PG G/H S-2 inhibitor suppressed generation of all three products without preventing synthesis of the enzyme. PG G/H S-1 was not induced during stimulation of 8-epi-PGF 2␣ generation in these experiments; thus, the contribution of the two enzymes to its production is clearly segregated.
Unlike the case with platelet PG G/H S-1, 8-epi-PGF 2␣ is formed in greater abundance relative to the conventional products of PG G/H S-2 in monocytes. Thus, whereas maximal 8-epi-PGF 2␣ formation in platelets is roughly 1 ⁄1000 that of TxB 2 , maximal formation by PG G/H S-2 in monocytes is 1 ⁄6 and 1 ⁄100 of the corresponding production of PGE 2 and TxB 2 , respectively, the most abundant conventional products of PG G/H S-2 in these cells. This observation raises the possibility that the mitogenic properties of 8-epi-PGF 2␣ (20) may contribute to the role of PG G/H S-2 activation in syndromes of vascular prolif-eration (45,46). Selective inhibitors of this enzyme might be used in clarifying the utility of urinary 8-epi-PGF 2␣ as an index of free radical generation in syndromes putatively associated with oxidant stress, in which PG G/H S-2 induction is possible. Although enzyme-catalyzed formation seems to be a feature of 8-epi-PGF 2␣ , but not of other F 2 -isoprostanes, similar caution might be applied to estimates of "F 2 -isoprostanes" (47), of which 8-epi-PGF 2␣ is an abundant constituent (48).
Monocytes are the first example of cells in which both enzyme-and free radical-catalyzed formation of 8-epi-PGF 2␣ have been demonstrated. Activation of monocytes with zymosan, in contrast to LPS, concanavalin A, or PMA, fails to induce PG G/H S-2 expression or 8-epi-PGF 2␣ generation. However, when coincubation with LDL affords the availability of an abundant lipid substrate, zymosan-activated monocytes catalyze the formation of a substantial amount of 8-epi-PGF 2␣ . This occurs coincident with production of TBARS and hydroperoxides, both indices of lipid oxidation, but not with either induction of PG G/H S-2 or generation of either PGE 2 or TxB 2 . Under these circumstances, selective PG G/H S-2 inhibitors are ineffective in preventing 8-epi-PGF 2␣ formation. Rather, antioxidants, such as superoxide dismutase or BHT, inhibit its production. Incubation of human monocytes with LDL in the absence of zymosan activation fails to stimulate 8-epi-PGF 2␣ production.
Previous work has demonstrated that copper-induced oxidation of LDL (49) is associated with increased 8-epi-PGF 2␣ generation. Whether it or indeed other arachidonic acid products contribute to the role of monocytes and their cellular derivatives in the process of atherogenesis remains to be determined.