Interleukin-1 enhances pancreatic islet arachidonic acid 12-lipoxygenase product generation by increasing substrate availability through a nitric oxide-dependent mechanism.

Interleukin-1 (IL-1) impairs insulin secretion from pancreatic islets and may contribute to the pathogenesis of insulin-dependent diabetes mellitus. IL-1 increases islet expression of nitric oxide (NO) synthase, and the resultant overproduction of NO participates in inhibition of insulin secretion because NO synthase inhibitors, e.g.NG-monomethyl-arginine (NMMA), prevent this inhibition. While exploring effects of IL-1 on islet arachidonic acid metabolism, we found that IL-1 increases islet production of the 12-lipoxygenase product 12-hydroxyeicosatetraenoic acid 12-(HETE). This effect requires NO production and is prevented by NMMA. Exploration of the mechanism of this effect indicates that it involves increased availabilty of the substrate arachidonic acid rather than enhanced expression of 12-lipoxygenase. Evidence supporting this conclusion includes the facts that IL-1 does not increase islet 12-lipoxygenase protein or mRNA levels and does not enhance islet conversion of exogenous arachidonate to 12-HETE. Mass spectrometric stereochemical analyses nonetheless indicate that 12-HETE produced by IL-1-treated islets consists only of the S-enantiomer and thus arises from enzyme action. IL-1 does enhance release of nonesterified arachidonate from islets, as measured by isotope dilution mass spectrometry, and this effect is suppressed by NMMA and mimicked by the NO-releasing compound 3-morpholinosydnonimine. Although IL-1 increases neither islet phospholipase A2 (PLA2) activities nor mRNA levels for cytosolic or secretory PLA2, a suicide substrate which inhibits an islet Ca2+-independent PLA2 prevents enhancement of islet arachidonate release by IL-1. IL-1 also impairs esterification of [3H8]arachidonate into islet phospholipids, and this effect is prevented by NMMA and mimicked by the mitochondrial ATP-synthase inhibitor oligomycin. Experiments with exogenous substrates indicate that NMMA does not inhibit and that the NO-releasing compound does not activate islet 12-lipoxygenase or PLA2 activities. These results indicate that a novel action of NO is to increase levels of nonesterified arachidonic acid in islets.

Interleukin-1 (IL-1) impairs insulin secretion from pancreatic islets and may contribute to the pathogenesis of insulin-dependent diabetes mellitus. IL-1 increases islet expression of nitric oxide (NO) synthase, and the resultant overproduction of NO participates in inhibition of insulin secretion because NO synthase inhibitors, e.g. N G -monomethyl-arginine (NMMA), prevent this inhibition. While exploring effects of IL-1 on islet arachidonic acid metabolism, we found that IL-1 increases islet production of the 12-lipoxygenase product 12-hydroxyeicosatetraenoic acid 12-(HETE). This effect requires NO production and is prevented by NMMA. Exploration of the mechanism of this effect indicates that it involves increased availabilty of the substrate arachidonic acid rather than enhanced expression of 12-lipoxygenase. Evidence supporting this conclusion includes the facts that IL-1 does not increase islet 12lipoxygenase protein or mRNA levels and does not enhance islet conversion of exogenous arachidonate to 12-HETE. Mass spectrometric stereochemical analyses nonetheless indicate that 12-HETE produced by IL-1treated islets consists only of the S-enantiomer and thus arises from enzyme action. IL-1 does enhance release of nonesterified arachidonate from islets, as measured by isotope dilution mass spectrometry, and this effect is suppressed by NMMA and mimicked by the NO-releasing compound 3-morpholinosydnonimine. Although IL-1 increases neither islet phospholipase A 2 (PLA 2 ) activities nor mRNA levels for cytosolic or secretory PLA 2 , a suicide substrate which inhibits an islet Ca 2؉ -independent PLA 2 prevents enhancement of islet arachidonate release by IL-1. IL-1 also impairs esterification of [ 3 H 8 ]arachidonate into islet phospholipids, and this effect is prevented by NMMA and mimicked by the mitochondrial ATP-synthase inhibitor oligomycin. Experiments with exogenous substrates indicate that NMMA does not inhibit and that the NO-releasing compound does not activate islet 12-lipoxygenase or PLA 2 activities. These results indicate that a novel action of NO is to increase levels of nonesterified arachidonic acid in islets.
Under the conditions of these studies (up to 24 h incubation with 5 units/ml IL-1), IL-1 is not cytotoxic to islets, and effects on insulin secretion are fully reversible upon removal of IL-1 from medium and continued culture (29). Lack of cytotoxicity in our experiments is also reflected by normal islet morphology, preservation of islet number, preservation of islet total acid-ethanol-extractable insulin content, and constant glyceraldehyde-3-phosphate dehydrogenase expression during incubation with IL-1 and other additives for up to 24 h.
Influence of the Nitric Oxide-releasing Compound SIN-1 on Islet Arachidonate Release-Islets (about 400/condition) were incubated in cCMRL (3 h, 37°C) after additions were complete, and medium content of arachidonate was then quantitated. Control islets were incubated without further additions. SIN-1 (3-morpholinosydnonimine) was added (100 M) to the remaining conditions. For conditions involving agents (500 M NMMA or 50 M hemoglobin) in addition to SIN-1, these agents were added 5 min before addition of SIN-1.
Prostaglandin E 2 and Insulin Measurement-PGE 2 was extracted from acidified medium with octadecylsilicic acid columns (1 ml, Baker Scientific, Phillipsburg, NJ) and, after column washing, eluted with methyl formate, as described (10,30). Eluant was concentrated to dryness, reconstituted in buffer, and PGE 2 quantified by enzyme immunoassay, as described (30). Insulin content of islet incubation medium or acid-ethanol extracts of islet pellets was measured by radioimmunoassay (31).
Quantitation of Arachidonic Acid-Arachidonic acid was quantitated by stable isotope dilution GC-NICI-MS (27). To aliquots of medium, [ 2 H 8 ]arachidonic acid (100 ng), and [ 3 H 8 ]arachidonic acid (50 nCi) were added as internal standards. Medium was acidified (pH 3.0, 1 N HCl) and extracted twice (CH 2 Cl 2 ). Extracts were concentrated to dryness, reconstituted in HPLC mobile phase, and purified by RP-HPLC on an Ultrasphere ODS column described above in the solvent system (flow 2 ml/min, column temperature 37°C), acetonitrile/water/acetic acid (80/ 20/0.1). The arachidonate peak was located by liquid scintillation counting of column eluant (retention volume about 50 ml), extracted (CH 2 Cl 2 ), concentrated to dryness, converted to a PFBE derivative, and analyzed by GC-NICI-MS. Quantitation of arachidonate was performed relative to [ 2 H 8 ]arachidonate internal standard by reference to a standard curve (27).
Gas Chromatography-Derivatized samples of 12-HETE-(PFBE, TMS) or arachidonate-PFBE were introduced in heptane into a Hewlett-Packard 5890 gas chromatograph (GC) via a Grob-type injector (temperature, 225°C) operated in the splitless mode and analyzed on an HP Ultraperformance capillary column (8 m length, cross-linked methylsilicone, inner diameter, 0.31 mm; film thickness, 0.17 m) interfaced with a Hewlett-Packard 5988B mass spectrometer (35,36). Helium was carrier gas (total flow, 10 ml/min; head pressure, 4 lb/in 2 ). Initial oven temperature was 85°C. Injector and interface temperatures were 225°C. At 0.5 min after injection, oven temperature was increased (40°C/min) to a final temperature of 215°C for analysis of 12-HETE-(PFBE, TMS) (retention time was about 5.78 min). For analysis of arachidonate-PFBE, initial oven temperature was 85°C, and 0.5 min after injection oven temperature was increased (30°C/min) to a final temperature of 200°C (retention time about 5.49 min).
Phospholipase A 2 Activity Measurements-PLA 2 activity in subcellular fractions (150 l of cytosolic or 100 l of membranous fractions, average protein content 70 g) was assayed (45) by ethanolic injection (5 l) of 2.5 M (final concentration) radiolabeled phospholipid substrate in assay buffer (final conditions 400 l of total volume, 200 mM Tris, pH 7.5, and either 10 mM EGTA or 10 mM CaCl 2 ). In experiments examining effects of NMMA (0.5 mM) or SIN-1 (0.1 mM) on PLA 2 activity, test compounds were added 5 min before addition of substrate. In experiments examining effects of prior incubation with IL-1 on islet PLA 2 activity, islets were preincubated without or with IL-1 (5 units/ml, 24 h, 37°C) before preparation of subcellular fractions. Radiolabeled substrates were 1-0-(Z)-hexadec-1Ј-enyl-2-(9,10- Assay mixtures were incubated (2 min, 37°C) and reactions terminated by adding butanol (100 l) (46) and vortexing. The organic phase was separated by centrifugation (2000 ϫ g, 2 min) and a 25-l aliquot applied to channeled Silica Gel G TLC plates, which were then developed with petroleum ether/ethyl ether/acetic acid (70/30/1) to resolve fatty acids (R F 0.58) from diglycerides (R F 0.21-0.24). The fatty acid region was scraped into a scintillation vial, and liquid scintillation spectrometry was performed after addition of Universol (3 ml). PLA 2 specific activity was calculated as [R/(PxT)], where R is fatty acid released in picomoles, P assay tube protein content (mg), and T assay duration (2 min). Parameter R is [4 ϫ D ϫ S], where the factor 4 accounts for the fraction (25%) of butanol extracts analyzed; S is specific radioactivity (dpm/ pmol) of phospholipid substrate; and D is dpm of released fatty acid. with IL-1 (5 units/ml) alone (24 h, 37°C) in cCMRL. In other experiments, islets were first incubated without or with oligomycin (10 g/ml) (1 h, 37°C). Islets were then removed from medium and resuspended (cCMRL) in Petri dishes (10 ϫ 35 mm) to which were added 10 Ci (for IL-1 experiments) or 0.5 Ci (for oligomycin experiments) of [ 3 H 8 ]arachidonic acid and unlabeled arachidonic acid (10 M). Dishes were then incubated (15 or 90 min, 37°C) and islets transferred to 5-ml conical test tubes and washed 3 times by suspension in and centrifugation from PBS plus 0.1% BSA to remove unincorporated radiolabel. Islet lipids were then Bligh-Dyer extracted (47). Extracts were concentrated to dryness in scintillation vials, scintillant added, and 3 H-content determined by liquid scintillation spectrometry.
Reverse Phase HPLC Analysis of [ 3 H]Arachidonate Released from Prelabeled Islets Treated with Interleukin-1 and Other Agents-In experiments described in Table I, [ 3 H]arachidonate released from islets that had previously been labeled and then incubated (24 h) without or with IL-1 in the presence or absence of NMMA at 5 or 20 mM glucose was determined by RP-HPLC. Medium was acidified (pH 3.0, 1 N HCl), extracted (CH 2 Cl 2 ), concentrated to dryness, reconstituted in CH 3 OH (0.1 ml), and analyzed on the Ultrasphere ODS column described above in the solvent system (flow 2 ml/min, column temperature 40°C) methanol/water/acetic acid (80/20/0.1). The 3 H-content of the arachidonate peak (retention volume about 50 ml) was determined by liquid scintillation spectrometry.

RESULTS
To determine whether treatment of pancreatic islets with IL-1 would suppress 12-HETE production, islets were incubated without or with IL-1 (5 units/ml) for 24 h. Islets were then incubated in the absence of IL-1 for 30 min with 3 mM D-glucose (basal condition) or with 17 mM D-glucose plus 0.5 mM carbachol (stimulatory condition). Islet production of 12-HETE was then measured by GC-NICI-MS and release of PGE 2 and insulin by immunoassay. As illustrated in Fig. 1, 17 mM Dglucose plus carbachol stimulated insulin secretion and production of PGE 2 and 12-HETE by control islets, and prior IL-1 exposure augmented PGE 2 production and suppressed insulin secretion, as previously observed (3-6, 8, 9, 13, 14, 16). Surprisingly, prior IL-1 exposure did not suppress but rather enhanced islet 12-HETE production ( Fig. 1, panel A). The majority of 12-HETE and PGE 2 produced under these conditions was released into medium, and little remained associated with islets (not shown).
The time course of production of 12-HETE and PGE 2 was examined with islets incubated without or with IL-1 for various periods (Fig. 2). Incubation of islets with IL-1 induced a timedependent rise in medium 12-HETE, PGE 2 , and nitrite contents compared to control conditions. Nitrite is an NO oxidation product formed in aqueous solutions (21), and its accumulation reflects induction of islet nitric oxide synthase by IL-1 (14). No difference in medium 12-HETE content, relative to control conditions, was observed until islets had been incubated with IL-1 for 8 h. IL-1-induced increases in medium PGE 2 , 12-HETE, and nitrite contents continued to rise for 24 h, by which time IL-1-treated islets had released nearly 10-fold more 12- HETE than control islets.
The effects of IL-1 to induce an increase in islet PGE 2 production and to suppress insulin secretion require new protein synthesis and are prevented by inhibitors of transcription (e.g. actinomycin D) or translation (e.g. CHX) (14,48). Islet proteins induced by IL-1 include inducible NO synthase and cyclooxygenase-2. (14). IL-1-induced enhancement of islet PGE 2 production is also partially suppressed by the NO synthase inhib-itor NMMA, an effect attributable to activation of cyclooxygenase by NO (14). We reasoned that IL-1-enhancement of islet 12-HETE production (Figs. 1 and 2) might reflect induction of synthesis of 12-lipoxygenase enzyme, analogous to its effects on cyclooxygenase-2, but that the increase in 12-HETE synthesis might be blunted by concomitant generation of NO, which was expected to partially inhibit islet 12-lipoxygenase by analogy with platelet 12-lipoxygenase (24). Effects of NMMA, actinomycin D, and CHX on IL-1-stimulation of islet production of eicosanoids and nitrite were therefore examined (Fig. 3). As expected, IL-1 enhancement of islet production of PGE 2 and nitrite was prevented by actinomycin D and CHX and was reduced by NMMA. Both actinomycin D and CHX also prevented IL-1-enhancement of islet 12-HETE production, consistent with a requirement for protein synthesis for this effect. Contrary to expectations, NMMA also prevented IL-1-stimulation of 12-HETE production (Fig. 3), suggesting that this effect required NO generation.
The possible involvement of NO in enhancing islet 12-HETE production by IL-1 suggested that the 12-HETE might derive from NO-dependent, nonenzymatic peroxidation of arachidonate rather than from 12-lipoxygenase action. NO interacts with superoxide anion to yield peroxynitrite, which, upon protonation, decomposes to yield hydroxyl radicals which can initiate lipid peroxidation (49). The stereochemical composition of 12-HETE generated from IL-1-treated islets was therefore determined. Islet 12-lipoxygenase produces exclusively 12-S-HETE, but nonenzymatic lipid peroxidation produces a racemic mixture of 12-S-and 12-R-HETE (8). Stereochemical analyses were performed by addition of racemic [ 18 O 2 ]12-HETE internal standard and sequential chiral-phase HPLC and then GC-NICI-MS analyses. These analyses revealed that 12-HETE released from IL-1-treated islets consisted exclusively of the Senantiomer (Fig. 4) and established that the dominant mechanism in enhancing islet 12-HETE production by IL-1 is enzymatic synthesis.
To evaluate the possibility that IL-1 increases expression of islet 12-lipoxygenase, conversion of exogenous arachidonate to 12-HETE by islets incubated without or with IL-1 was examined. Previously, incubation of islets with IL-1 was found to increase PGE 2 production from maximally effective arachidonate concentrations (13), and this was the first evidence that IL-1 induces islet cyclooxygenase expression, a possibility later verified by cyclooxygenase-2 immunochemical analyses (14). Incubation of islets with IL-1 for 24 h resulted in an enhanced ability to convert exogenous arachidonate to PGE 2 (Fig. 5,  panel A). In contrast, although exogenous arachidonate induced a striking rise in islet 12-HETE production, no enhancement of this effect was observed in IL-1-treated islets (Fig. 5,  panel B), suggesting that IL-1 did not increase expression of islet 12-lipoxygenase enzyme. This was supported by immunochemical analyses. In immunoprecipitation studies, islets were metabolically labeled with [ 35 S]methionine in the absence or presence of IL-1, homogenized, and treated with anti-12-lipoxygenase antibody. Upon analysis of immunoprecipitates by SDSpolyacrylamide gel electrophoresis and fluorography, a 35 Slabeled protein of appropriate size (74 kDa) for 12-lipoxygenase was visualized, but IL-1 did not influence levels of this protein (not shown). Similar results were obtained in immunoblotting studies of islet 12-lipoxygenase (inset in panel B, Fig. 5).
In addition, RT-PCR experiments using primers specific for the sequence of rat leukocyte-type 12-lipoxygenase and islet RNA as template yielded a product of the expected size (632 base pairs) for a 12-lipoxygenase cDNA fragment before addition of IL-1 (time 0 lane of upper panel of Fig. 6), but this material did not increase in abundance after adding IL-1 to islets (time 2, 4, 8, 24, and 48 h lanes of upper panel of Fig. 6). Under these conditions, RT-PCR analyses using primers specific for sequences of iNOS or for glyceraldehyde-3-phosphate dehydrogenase revealed clear induction of iNOS mRNA and constant levels of glyceraldehyde-3-phosphate dehydrogenase mRNA (middle and lower panels of Fig. 6). Activity and protein level measurements and RT-PCR estimates of mRNA levels thus all indicate that IL-1 does not induce islet 12-lipoxygenase synthesis.
Potential explanations for suppression of IL-induced enhancement of islet 12-HETE production by NMMA are that NMMA inhibits or that NO activates islet 12-lipoxygenase. Effects of NMMA and the NO-releasing compound SIN-1 (3morpholinosydnonimine) (21) on islet conversion of exogenous arachidonate to 12-HETE were therefore examined. Preincubation of islets with NMMA or with SIN-1 before addition of arachidonate did not significantly affect islet conversion of exogenous arachidonate to 12-HETE (Fig. 7). This indicates that NMMA does not inhibit islet 12-lipoxygenase and suggests that islet 12-lipoxygenase, in contrast to the platelet isoform, is relatively insensitive to NO.
The observations that IL-1 enhances islet production of 12-HETE by 12-lipoxygenase but does not increase expression of 12-lipoxygenase enzyme suggest that IL-1 might increase substrate availability. Effects of IL-1 on release of nonesterified arachidonate from islets were therefore examined by isotope dilution GC-NICI-MS. Medium content of arachidonic acid for control islets incubated with no additions for 24 h was found to be 306 Ϯ 126 ng/ml and that for islets incubated with IL-1 rose to 1109 Ϯ 130 ng/ml (p Ͻ 0.001, Fig. 8, panel A). The NO synthase inhibitor NMMA prevented IL-1-induced accumulation of nonesterified arachidonic acid (Fig. 8, panel A), under conditions where NMMA also suppressed IL-1-induced accu- mulation of 12-HETE (Fig. 8, panel B) and PGE 2 (Fig. 8, panel  C). This suggests that enhanced 12-HETE production by IL-1treated islets reflects increased substrate availability to the 12-lipoxygenase and that this effect is mediated by an NO-dependent mechanism.
Consistent with a role for NO in accumulation of nonesterified arachidonate is the observation that the NO-releasing compound SIN-1 induced accumulation of non-esterified arachidonate in medium of islets incubated with this compound (Fig. 9). This effect was prevented by the NO scavenger hemoglobin (21) but was not influenced by the NO synthase inhibitor NMMA (Fig. 9). The lack of effect of NMMA was expected because SIN-1 releases NO by a nonenzymatic process, and NMMA is not a chemical scavenger of NO. The lack of effect of NMMA on accumulation of nonesterified arachidonate in medium of islets incubated with SIN-1 also suggests that NMMA does not inhibit islet phospholipases which release arachidonic acid esterified in phospholipids. NMMA also failed to suppress and SIN-1 failed to activate hydrolysis of radiolabeled fatty acids from synthetic phospholipid substrates catalyzed by Ca 2ϩdependent or Ca 2ϩ -independent PLA 2 activities in islet cytosolic or membranous fractions (not shown). This suggests that islet PLA 2 enzymes are neither directly inhibited by NMMA nor directly activated by NO.
To examine the possibility that IL-1 might increase expression of islet PLA 2 enzymes, both Ca 2ϩ -dependent and Ca 2ϩindependent PLA 2 activities were measured with exogenous phospholipid substrates added to cytosolic or membranous fractions from control islets and from islets that had been incubated with IL-1 for 24 h. IL-1 did not induce increases in any PLA 2 activity in either subcellular fraction (not shown). In addition, RT-PCR studies were performed with islet RNA as template and primers specific for cytosolic 85-kDa PLA 2 (cPLA 2 ), for a type II 14-kDa PLA 2 (TIIPLA 2 ), for iNOS, and for glyceraldehyde-3-phosphate dehydrogenase. Under conditions where IL-1 induced an increase in iNOS mRNA, there was a constant level of signal for message for cPLA 2 , TIIPLA 2 , and glyceraldehyde-3-phosphate dehydrogenase (Fig. 10). Activity measurements and RT-PCR analyses thus suggest that IL-1 does not increase islet expression of cPLA 2 or TIIPLA 2 .
In addition to cPLA 2 and TIIPLA 2 , a third PLA 2 activity expressed in islets is an ATP-stimulated, Ca 2ϩ -independent (ASCI)-PLA 2 (31,35,36,45,50). The molecular mass of the catalytic subunit of the islet enzyme (49) and of an analogous myocardial enzyme (51) is 40 kDa, and myocardial ASCI-PLA 2 is activated by a regulatory subunit which is an isoform of the glycolytic enzyme phosphofructokinase (52). Both islet and myocardial ASCI-PLA 2 are inhibited by a haloenol lactone suicide substrate (HELSS) which does not inhibit cPLA 2 or TIIPLA 2 (53,54). NO has recently been reported to activate an ASCI-PLA 2 in RAW 264.7 cells because NO augments release of [ 3 H]arachidonic acid from prelabeled cells; this effect is enhanced by increasing medium glucose concentration; and the effect is suppressed by HELSS (55). ASCI-PLA 2 activation in this system is thought to reflect stimulation of glycolytic flux by NO (55), perhaps because of impairment of mitochondrial function by NO (21, 23).
To examine the possibility that ASCI-PLA 2 activation might be responsible for IL-1-induced enhancement of islet release of nonesterified arachidonate, islets were prelabeled by a 24-h incubation with [ 3 H 8 ]arachidonate. Prelabeled islets were then incubated with no additions, with IL-1 alone, or with IL-1 plus NMMA in the presence of either 5 or 20 mM glucose for 24 h. At the end of the incubation, release of [ 3 H]arachidonic acid was measured by RP-HPLC. At 5 mM glucose, IL-1 induced more than a doubling of [ 3 H]arachidonic acid release, and this effect was prevented by NMMA (Table I, panel A). IL-1-induced release of [ 3 H]arachidonic acid was not enhanced by 20 mM glucose, however, but was attenuated (Table I,  A mechanism whereby NO might increase availability of nonesterified arachidonate without enhancing hydrolysis of arachidonate from phospholipids is by suppressing re-esterification of arachidonate released during phospholipid turnover. Evidence that incubation of islets with IL-1 suppresses esterification of arachidonic acid into phospholipids by a mechanism involving NO production was obtained from experiments in which islets were pulse-labeled (15 min) with [ 3 H 8 ]arachidonic acid after incubation for 24 h with no additions, with IL-1 alone, or with IL-1 plus the NO synthase inhibitor NMMA (Fig.  11, panel A). [ 3 H 8 ]Arachidonate was incorporated equally well into lipids of islets that had been incubated with no additions or with IL-1 plus NMMA, but incorporation was substantially reduced with islets incubated with IL-1 alone. (Under loading conditions in Fig. 11, panel A, a mean specific radioactivity of 517 dpm/fmol was achieved in the islet nonesterified arachidonate pool, and there was no significant difference in this parameter among the 3 conditions). Conversely, when islets were first pulse-labeled with [ 3 H 8 ]arachidonate and then incubated for 24 h with no additions, with IL-1 alone, or with IL-1 plus NMMA, retention of [ 3 H]arachidonate in islet lipids was indistinguishable for islets incubated with no additions or with IL-1 plus NMMA but was significantly reduced for islets incubated with IL-1 alone (Fig. 11, panel B). The decline in content of esterified [ 3 H 8 ]arachidonate in IL-1-treated islets was also associated with a decrement in esterified arachidonate mass (control 34.90 Ϯ 4.7 pmol/islet; IL-1 24.65 Ϯ 0.25 pmol/islet; IL-1 ϩ NMMA 39.38 Ϯ 3.97 pmol/islet).
The magnitude of the effect of prior IL-1-treatment to suppress esterification of [ 3 H]arachidonate into islet lipids became more prominent as the labeling period was extended from 15 to 90 min, and this effect was prevented at both time points if exposure to IL-1 had occurred in the presence of NMMA (Fig.  12). After the 90-min labeling period, islet phospholipids were extracted and analyzed by NP-HPLC to separate phospholipid head group classes. Under these short-term labeling conditions, the majority of radiolablel is incorporated into phosphatidylcholine (Fig. 13), as previously reported (56). Prior treatment with IL-1 reduced incorporation of [ 3 H]arachidonic acid into all phospholipids except phosphatidylethanolamine, and the largest decrement occurred in phosphatidylcholine (Fig.   FIG. 7. Influence of  13). (Total phospholipid amounts recovered from control and IL-1-treated islets under these conditions, as estimated by NP-HPLC-UV absorbance tracings, were not different (not shown).) The fact that the effect of IL-1 to suppress esterification of [ 3 H]arachidonic acid into islet phospholipids was prevented by NMMA indicated that NO production is required for this effect. Because arachidonate requires ATP-dependent conversion to a coenzyme A thioester before esterification into phospholipids and because NO inhibits mitochondrial ATP generation (21,23), it is possible that suppression of arachidonate esterification in IL-1-treated islets is attributable to induction of islet NO synthase by IL-1 with resultant NO overproduction and inhibition of mitochondrial function. Consistent with this possibility, treatment of islets with the mitochondrial  10. Examination of islet RNA for content of species encoding a cytosolic 85-kDa PLA2, a 14-kDa type II PLA2, nitric oxide synthase, and glyceraldehyde-3-phosphate dehydrogenase after incubation with interleukin-1 without or with NMMA. Isolated islets were incubated (6 h, 37°C) with no additions (lanes 1), with IL-1 (5 units/ml) (lanes 2), or with IL-1 plus NMMA (0.5 mM) (lanes 3). At the end of incubations, total RNA was isolated as described under ''Experimental Procedures.'' First strand cDNA was transcribed from the RNA template with avian myeloblastosis virus reverse transcriptase. PCR was performed, as described under ''Experimental Procedures,'' with primer sets designed to amplify a fragment of sequences encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH, first set of 3 lanes), inducible nitric oxide synthase (iNOS, second set of 3 lanes), 85-kDa cytosolic PLA 2 (cPLA 2 , third set of 3 lanes), or a 14-kDa type II PLA 2 (TIIPLA 2 , fourth set of 3 lanes). PCR products were analyzed by 1% agarose gel electrophoresis and visualized with ethidium bromide. Similar results were obtained for all of these PCR products in time course studies similar to that in Fig. 6 in which incubation time was varied from 2 to 24 h. No induction of cPLA 2 or TIIPLA 2 was apparent at any time point, and signal for glyceraldehyde-3-phosphate dehydrogenase remained constant throughout this period.
ATP-synthase inhibitor oligomycin (57) was also found to impair esterification of [ 3 H]arachidonic acid into islet lipids (Fig. 14). DISCUSSION Interleukin-1 impairs insulin secretion by pancreatic islets and induces islet expression of cyclooxygenase and NO synthase, resulting in overproduction of PGE 2 and NO (13,14). Inhibition of insulin secretion by IL-1 is attributable in part to islet NO production because this effect is prevented by NO synthase inhibitors, e.g. NMMA (58). NO is thought to inhibit insulin secretion by inactivating iron-sulfur enzymes including mitochondrial aconitase (23,34). Because lipoxygenases are iron-sulfur enzymes that may complex with NO (59) and because islet 12-lipoxygenase products may promote insulin secretion (1-9, 18), we have characterized effects of IL-1 on islet production of the 12-lipoxygenase product 12-HETE to determine whether inhibition of islet 12-lipoxygenase might be among the mechanisms whereby IL-1 inhibits insulin secretion.
Surprisingly, IL-1 enhanced rather than suppressed islet 12-HETE production. This effect required several hours to develop and was prevented by inhibitors of transcription and translation, suggesting a requirement for new protein synthesis, and was prevented by the NO synthase inhibitor NMMA. Stereochemical analysis of 12-HETE from IL-1-treated islets indicated that it consisted only of the S-enantiomer, reflecting enzymatic synthesis and not nonenzymatic lipid peroxidation. Immunochemical and RT-PCR studies indicated that IL-1 did not increase 12-lipoxygenase protein or mRNA, and IL-1 failed to enhance islet conversion of exogenous arachidonate to 12-HETE. NMMA did not impair islet conversion of arachidonate to 12-HETE, indicating that NMMA does not inhibit islet 12lipoxygenase. IL-1 did enhance release of nonesterified arachidonate from islets and this effect was also suppressed by NMMA. These observations suggest that enhanced 12-HETE production by IL-1-treated islets reflects increased substrate availability to the 12-lipoxygenase and that this effect occurs by an NO-dependent mechanism. This possibility is supported  (24 h, 37°C), washed to remove unincorporated label, divided into aliquots for experimental incubations (about 375 islets/condition), suspended in cCMRL with or without additives, and incubated (18 h, 37°C). In A, cCMRL for experimental incubations was supplemented with no added glucose (final concentration 5 mM) or with sufficient glucose to achieve a concentration of 20 mM. Experimental medium was also supplemented with no IL-1 or NMMA, with IL-1 alone (5 units/ml), or with IL-1 plus NMMA (0.5 mM). At the end of the incubation, medium was removed and 3 H-content of an aliquot determined. Islets were then Bligh-Dyer extracted, and 3 H-content of the lipid extract determined. The sum of 3 H-contents of supernatant and lipid extract reflects total [ 3 H]arachidonate initially incorporated. Residual supernatant was extracted and analyzed by RP-HPLC to separate arachidonate from its metabolites, and by the fact that incubation of islets with an NO-releasing compound enhanced accumulation of nonesterified arachidonate and that this effect was prevented by an NO scavenger. That this effect was not suppressed by NMMA suggests that NMMA does not inhibit islet phospholipases, as does the failure of NMMA to inhibit hydrolysis of fatty acids from phospholipid substrates catalyzed by phospholipases in islet cytosol and membranes.
IL-1 induces synthesis of PLA 2 enzymes in some cells (60 -62), and overexpression of such enzymes in islets could result in increased levels of nonesterified arachidonate. An 85-kDa cytosolic, Ca 2ϩ -activated PLA 2 (cPLA 2 ) participates in arachidonate release in some cells (63)(64)(65)(66)(67) and is expressed in human islets (68). Immunochemical evidence indicates that cPLA 2 is expressed at very low levels in rat islets (69). Our RT-PCR data indicate that cPLA 2 mRNA is expressed in rat islets but is not induced by IL-1. Low molecular mass (14 kDa) Ca 2ϩ -dependent PLA 2 enzymes also exist in islets (70). Our RT-PCR data indicate that islets express mRNA for a 14-kDa type II PLA 2 , but IL-1 treatment does not increase its abundance. Activity measurements also failed to demonstrate increased expression of Ca 2ϩ -dependent PLA 2 activities in subcellular fractions from IL-1-treated islets. Several enzymes may contribute to activity in such assays (71)(72)(73)(74)(75)(76)(77)(78)(79)(80), however, which could obscure induction of a specific enzyme. Activation of PLA 2 by reversible phosphorylation (67,81) or by NO-amplified Ca 2ϩ entry (82) might also not be demonstrable with broken cell assays.
Islets express an ATP-stimulated, Ca 2ϩ -independent PLA 2 (ASCI-PLA 2 ) which participates in secretagogue-induced hydrolysis of arachidonate from islet phospholipids (31,35,36,45,50), and, like an analagous myocardial enzyme (52), islet ASCI-PLA 2 may be regulated by interaction with an isoform of the glycolytic enzyme phosphofructokinase. A similar PLA 2 has been reported to be activated by NO in a RAW 264.7 cells because NO enhances arachidonate release from these cells; increasing flux through glycolysis by increasing medium glucose concentration increases NO-induced arachidonate release; and this effect is prevented by the ASCI-PLA 2 suicide substrate HELSS (55). Although incubation of islets with IL-1 did not increase expression of islet ASCI-PLA 2 activity in broken cell assays, pretreatment of islets with HELSS did suppress IL-1induced release of arachidonate from islets, suggesting that ASCI-PLA 2 may play at least a permissive role in this phenomenon. In contrast to the case for RAW 264.7 cells, however, increasing medium glucose concentration did not enhance IL-1-induced NO-dependent release of arachidonate from islets. The difference between these two systems may be related to the fact that NO inhibits islet phosphofructokinase activity (83) and that IL-1 induces a specific decline in islet mRNA for the muscle isoform of phosphofructokinase. 2 Both the time required for IL-1 to enhance islet 12-HETE production and the requirement for new protein synthesis may be attributable to induction of NO synthase in islets by IL-1 2 Z. Ma and J. Turk, manuscript in preparation.

FIG. 11. Influence of prior incubation with interleukin-1 with or without NMMA on incorporation of [ 3 H]arachidonic acid into islet lipids and on retention of previously incorporated [ 3 H]arachidonic acid in islet lipids.
In panel A, islets were incubated (24 h, 37°C, cCMRL) with no additions (control, left column), with IL-1 (5 units/ml) alone (IL-1, middle column), or with IL-1 (5 units/ml) plus NMMA (500 M) (IL-1 ϩ NMMA, right column). Islets were then removed from media, resuspended in fresh cCMRL, and counted into Petri dishes. To each dish was then added [ 3 H 8 ]arachidonate (10 Ci) mixed with unlabeled arachidonate (final concentration 10 M). Dishes were then incubated (15 min, 37°C). Islets were then washed with PBS ϩ 0.1% BSA to remove unincorporated radiolabel, and islet lipids were extracted and their 3 H-content determined. Error bars indicate S.E. (n ϭ 6). The double asterisk over the middle column indicates a significant difference (p ϭ 0.006) between the 3 H-lipid content of islets that had been incubated with IL-1 compared to that of control islets. There was no significant difference (p ϭ 0.269) between the 3 H-lipid content of control islets compared to that of islets that had been incubated with IL-1 ϩ NMMA. In panel B, islets were first prelabeled with [ 3 H 8 ]arachidionic acid in cCMRL (10 Ci/ml) for 30 min at 37°C and then washed with cCRML to remove unincorporated radiolabel. Islets (700/condition) were then resuspended in fresh cCRMRL and incubated (24 h, 37°C) with no additions (control, left column), with IL-1 (5 units/ml) alone (middle column), or with IL-1 (5 units/ml) plus NMMA (500 M, right column). At the end of incubations, islets were washed in PBS plus 0.1% BSA to remove unincorporated radiolabel, and islet lipids were extracted and their 3 H-content determined. Error bars represent S.E. (n ϭ 3). The double asterisk over the second column indicates a significant difference ( p ϭ 0.001) between the 3 H-lipid content of islets that had been incubated with IL-1 compared to that of control islets. There was no significant difference ( p ϭ 0.89) between the 3 H-lipid content of control islets compared to that of islets that had been incubated with IL-1 ϩ NMMA. (14,23,34,58) because NO production is required for enhanced 12-HETE production. Among the mechanisms whereby NO might increase 12-HETE production is by increasing 12-lipoxygenase substrate availability by suppressing re-esterification of arachidonate released during phospholipid turnover. Such reesterification requires formation of arachidonyl-CoA in a reaction that requires both ATP and CoASH. NO inhibits mitochondrial iron-sulfur enzymes including aconitase, inhibits mitochondrial respiration, and impairs ATP generation (21,23). NO can also nitrosylate thiol groups (84). Prior incubation of islets with IL-1 impaired islet incorporation of [ 3 H 8 ]arachidonate into lipids, and this effect was prevented by the NO synthase inhibitor NMMA. When islets were prelabeled with [ 3 H 8 ]arachidonate, subsequent incubation with IL-1 also impaired retention of incorporated radiolabel, and this effect was also prevented by NMMA. These observations are consistent with the possibility that IL-1 suppresses re-esterification of arachidonate into phospholipids through an NO-dependent mechanism. That this mechanism may involve inhibition of mitochondrial function by NO is suggested by the fact that the mitochondrial ATP-synthase inhibitor oligomycin (57) also suppressed esterification of [ 3 H]arachidonate into islet phospholipids.
A model consistent with our observations is that there is an ongoing cycle of deacylation/reacylation of arachidonate-con-taining phospholipids in islets. The deacylation component of this cycle may be mediated by the HELSS-sensitive enzyme ASCI-PLA 2 . A HELSS-sensitive Ca 2ϩ -independent PLA 2 in FIG. 14. Influence of oligomycin on incorporation of [ 3 H] arachidonic acid into islet lipids. Isolated islets were incubated (cCMRL, 1 h, 37°C) with no additions (left column) or with oligomycin (10 g/ml) (right column). Islets were then removed from media, resuspended in fresh cCMRL, and counted into Petri dishes. To each dish was then added [ 3 H 8 ]arachidonate (0.5 Ci) and unlabeled arachidonate (final concentration 10 M). Dishes were then incubated (15 min, 37°C). Islets were then washed with PBS ϩ 0.1% BSA to remove unincorporated radiolabel, and islet lipids were extracted and their 3 H-content determined. Error bars indicate S.E. (n ϭ 3). The asterisk over the second column indicates that the p value for the difference from the control condition is 0.028 by Student's t test. P388D1 cells has recently been reported to participate in the deacylation process involved in remodeling of arachidonatecontaining phospholipids (85). Ordinarily, deacylation of arachidonate is rapidly followed by ATP-dependent conversion to arachidonoyl-CoA and reacylation into phospholipids. The relative rapidity of reacylation compared to deacylation in unstimulated cells maintains concentrations of nonesterified arachidonate at low levels. IL-1-induced NO production may reduce the reacylation rate and lead to accumulation of nonesterified arachidonate. Blocking either the deacylation component of the pathway with HELSS or preventing the NO-induced defect in reacylation with NMMA abolishes effects of IL-1 to enhance accumulation of nonesterified arachidonate in islets and to increase flux of this substrate through the 12-lipoxygenase.
The effect of IL-1 to increase islet 12-HETE production may contribute to deleterious actions of IL-1 on islets. 12-HETE is generated from the hydroperoxy precursor 12-HPETE; 12-HPETE induces apoptosis of cells with reduced levels of glutathione peroxidase (86); and islets have low levels of glutathione peroxidase (87). Substrate-induced lipoxygenase activity also induces superoxide generation in the presence of reduced pyridine nucleotides, and this has been suggested to contribute to cellular oxidative stress (88). Islets also have low levels of superoxide dismutase (89), which reduces cellular superoxide levels. Superoxide and NO react to yield peroxynitrite, which, when protonated, yields hydroxyl radical (49), a potent oxidant postulated to participate in cytokine-induced islet injury (15). The effects of IL-1 to increase substrate flux through the lipoxygenase and to augment NO production may therefore interact cooperatively to inflict injury on beta cells.