Calcium Influx Factor, Further Evidence It Is 5,6-Epoxyeicosatrienoic Acid*

We present evidence in astrocytes that 5,6-epoxyeicosatrienoic acid, a cytochrome P450 epoxygenase metabolite of arachidonic acid, may be a component of calcium influx factor, the elusive link between release of Ca2+ from intracellular stores and capacitative Ca2+ influx. Capacitative influx of extracellular Ca2+ was inhibited by blockade of the two critical steps in epoxyeicosatrienoic acid synthesis: release of arachidonic acid from phospholipid stores by cytosolic phospholipase A2 and cytochrome P450 metabolism of arachidonic acid. AAOCF3, which inhibits cytosolic phospholipase A2, blocked thapsigargin-stimulated release of arachidonic acid as well as thapsigargin-stimulated elevation of intracellular free calcium. Inhibition of P450 arachidonic acid metabolism with SKF525A, econazole, orN-methylsulfonyl-6-(2-propargyloxyphenyl)hexanamide, a substrate inhibitor of P450 arachidonic acid metabolism, also blocked thapsigargin-stimulated Ca2+ influx. Nano- to picomolar 5,6-epoxyeicosatrienoic acid induced [Ca2+] i elevation consistent with capacitative Ca2+ influx. We have previously shown that 5,6-epoxyeicosatrienoic acid is synthesized and released by astrocytes. When 5,6-epoxyeicosatrienoic acid was applied to the rat brain surface, it induced vasodilation, suggesting that calcium influx factor may also serve a paracrine function. In summary, our results suggest that 5,6-epoxyeicosatrienoic acid may be a component of calcium influx factor and may participate in regulation of cerebral vascular tone.

One of the biochemical pathways for metabolism of arachidonic acid is the cytochrome P450 monooxygenase pathway, which results in formation of 4 regio-and stereoisomeric products; cis-5, 6-, 8,9-, 11,12-, and 14,15-epoxyeicosatrienoic acids (EETs). 1 Compared with our knowledge of the lipoxygenase and cyclooxygenase pathways for arachidonic acid metabolism, relatively little is understood about the epoxygenase pathway.
Numerous physiological roles have been suggested for the EETs and collectively, the EETs appear to have potent effects on ion channels (1)(2)(3).
One of the mechanisms for regulation of intracellular calcium dynamics in response to hormones and other agonists is through the "capacitative pathway" as originally described by Putney (4,5). Activation of this pathway occurs through G protein receptor-mediated activation of phospholipase C, which catalyzes the hydrolysis of phosphatidylinositol bisphosphate yielding inositol trisphosphate (Ins(1,4,5)P 3 ) and diacylglycerol. Ins(1,4,5)P 3 binds to its receptor on the intracellular calcium stores, initiating release of stored calcium (6). As the stores are depleted of calcium, a second messenger termed "calcium influx factor" (CIF) is released (7). CIF induces influx of extracellular calcium through second messenger operated channels (SMOC) in the plasma membrane, thereby coupling calcium entry to depletion of internal stores. To date, the identity of CIF remains unknown.
In previous work from our laboratory (8), we reported that capacitative calcium influx was linked to arachidonic acid release by activation of the 85-kDa cytosolic phospholipase A 2 (cPLA 2 ) in human U937 lymphoma cells and in rat cortical astrocytes (9), suggesting that arachidonic acid, or a metabolite thereof, was a component of CIF. Recent work from several other groups has shown that the actions of cytochrome P450 monooxygenase may also be coupled to capacitative calcium influx (10,11). Hoebel et al. (11) recently reported that functional P450 activity was critical to regulation of store-operated calcium influx and proposed that an EET may constitute the driving force for capacitative calcium entry in endothelial cells. Graier et al. (10) have presented evidence suggesting that 5,6-EET stimulates capacitative calcium influx in endothelial cells, consistent with CIF. However, if 5,6-EET is CIF or a component thereof, then it should have similar effects in all cells lines which signal through Ins(1,4,5)P 3 .
Although CIF is hypothesized to have its primary actions at an intracellular level, it may also be released into the extracellular environment, acting as a paracrine signal for SMOC calcium influx independent of release of calcium from intracellular stores. In Jurkat cells, CIF was released to the extracellular medium upon stimulation with phytohemmaglutinin (7). Endothelium-derived hyperpolarizing factor, which induces NO-and prostaglandin I 2 -independent relaxation of vascular smooth muscle, is released in response to agonists operating through Ins(1,4,5)P 3 -dependent signaling (12)(13)(14). Harder et al. (15) and Gebremedhin et al. (16) reported that cat brain converted arachidonic acid to EETs, which dilated cerebral arteries, implying that endothelium-derived hyperpolarizing factor may be an EET. Hecker et al. (17) reported an NO-and cyclooxygenase-independent relaxation of porcine aortic rings which was mediated by calcium-activated K ϩ channels (K Ca ) and required the combined actions of PLA 2 and P450, precisely those systems which may mediate CIF. Additional reports indicate that K Ca are closely coupled to capacitative calcium influx (18,19). Taken together, these reports lead to the speculation that endothelium-derived hyperpolarizing factor may represent a paracrine function of CIF. In the present report we present evidence suggesting that 5,6-EET is a CIF in astrocytes and when released into the extracellular environment may participate in regulation of local cerebrovascular tone.

Methods
Cell Culture-Astrocytes were prepared from 1-2-day-old rat pups as described previously (20). In brief, cortices were isolated, cleaned of white matter and meninges, minced, and trypsin-digested for 10 min. The dissociated tissue was diluted into Dulbecco's modified essential medium (supplemented with 10% fetal calf serum and 2 mM glutamine) and seeded into 75-cm 2 flasks at an initial density of 1-2 ϫ 10 6 cells per flask. Flasks were cultured to confluency (10 -14 days) in 5% CO 2 /air at 37°C. Medium was changed every 2-3 days. During medium changes, flasks were forcefully shaken to remove microglia, which were decanted with the medium. On reaching confluency, flasks were trypsinized (0.25% trypsin, 0.02% EDTA in saline) for 1 min. The trypsin/EDTA solution was aspirated and the cell monolayer (still adherent to the flask) was covered with 10 ml of Dulbecco's modified essential medium and incubated at 37°C for 10 min, sufficient time to lift the cells from the flask bottom. Lifted cells were washed and plated onto collagencoated 35 ϫ 100-mm tissue culture dishes at a density of 3.2 ϫ 10 5 cells per dish. For arachidonic acid release assays, 0.5 ϫ 10 5 cells were plated into 25-mm diameter tissue culture wells (6 wells/plate). Cells were then cultured to confluency (2 weeks). Astrocyte cultures were characterized for purity as described previously (20) and found to be Ͼ98% pure as assessed by the presence of glial fibrillary acidic protein. Cells were used for experiments at a total of 4 weeks after removal from the rat.
Fura-2 AM Loading-[Ca 2ϩ ] i was measured with the ratiometric dye, Fura-2, as described previously (21). Astrocytes were washed 3 times with DPBS supplemented with 2% fatty acid-free bovine serum albumin and 1 mM glucose and placed in 1.5 ml of this media for loading. Cells were loaded with 5 M Fura-2 AM for 50 min. at room temperature. This loading procedure resulted in complete dye hydrolysis as determined by scanning the excitation spectra of loaded cells, with negligible sequestration of dye in subcellular organelles (22). The intracellular concentration of Fura-2 was monitored as described previously by Poenie et al. (23).

Measurement of [Ca 2ϩ ] i -[Ca 2ϩ
] i was measured at room temperature using a Ratiomaster (Photon Technologies Int., South Brunswick, NJ) microspectrophotometry system as described previously (24). Excitation light was provided by a xenon arc lamp coupled to a scanning monochrometer which alternated excitation light between 350 and 380 nm. Bandpass was set at Ϯ2 nm. Excitation light was delivered to the cells via fiber optics through the epifluorescence port of a Zeiss Standard 16 microscope coupled to a Zeiss Achroplan 20ϫ water immersion lens. Use of the water immersion lens allowed in situ measurements of the cells while in the culture dish and easy manipulation. However, the optical properties of the lens necessitated ratioing of Fura-2 at 350 and 380 nm. Emission was measured at 510 nm, via a microphotometer. The entire system for data collection and analysis was computer driven.
After loading, astrocytes were washed twice and placed in 2 ml of DPBS supplemented with 1 mM glucose and 1% fatty acid-free bovine serum albumin. For measurement of [Ca 2ϩ ] i , a field of 1-2 astrocytes was selected using slit width adjustments on the microphotometer. Basal [Ca 2ϩ ] i was recorded for several seconds, at a sampling rate of 1 ratio every 0.5 s. For experiments using agonists (arachidonic acid, 5,6-EET, and thapsigargin) or inhibitors (econazole, SKF96365, SKF525A, AAOCF 3 , and MS-PPOH), these agents were added to the tissue culture dish in a volume of 200 l. This volume assured almost immediate mixing with the buffer, as determined by dye diffusion. For experiments using indomethacin, Fura-2-loaded astrocytes were incubated with 10 M indomethacin for 15 min prior to measurement of [Ca 2ϩ ] i . Each experiment was performed on a separate culture of astrocytes.
[Ca 2ϩ ] i was calculated as described previously (25) using a correction for intracellular viscosity (23). Autofluorescence at 350 and 380 was recorded from wells of astrocytes not loaded with Fura-2 and was subtracted from all measurements. Agonist-stimulated [Ca 2ϩ ] i concentrations were normalized to the basal [Ca 2ϩ ] i level in each experiment and are expressed as a percent of basal. Leakage of Fura-2 into the medium was monitored by measurement of intracellular Fura-2 concentration at the isobestic wavelength, 362 nm (25).
For calcium-free experiments, cells were placed in calcium-free DPBS containing 1 mM glucose and 1% fatty acid-free bovine serum albumin immediately prior to the experiment. EGTA (0.5 mM) was added to chelate any residual calcium. Astrocytes were not exposed to calcium-free conditions for longer than 15 min. As described previously, this treatment did not result in lifting of cells from the monolayer.
Arachidonic Acid Release-Arachidonic acid release was measured as described previously (8,26). Briefly, 1 ϫ 10 6 cells grown in a 25-mm diameter well were radiolabeled by addition of 0.5 Ci of [ 3 H]arachidonic acid to the growth medium for 24 h prior to the assay. Incorporation of radiolabel was 95%. After labeling, the cells were washed three times and placed in DPBS supplemented with 1 mM glucose and 1% fatty acid-free bovine serum albumin. Next, cells were preincubated with 10 M econazole, 5 M AAOCF 3 , 1 M MS-PPOH, or 10 M SKF525A for 2 min in a final volume of 1.0 ml. Following preincubation, cells were stimulated with 1 M thapsigargin for 2 min. Radioactivity released into the medium after thapsigargin stimulation was determined by scintillation counting. Controls consisted of unstimulated cells, or cells treated with inhibitors alone. Econazole, AAOCF 3 , MS-PPOH, and SKF525A did not significantly alter basal arachidonic acid release as compared with untreated controls. For experiments conducted under calcium-free conditions, calcium-free DPBS supplemented with 1 mM glucose, 1% fatty acid-free bovine serum albumin, and 0.5 mM EGTA was utilized for all incubations. Results are expressed as percent of control (unstimulated) cells.
Preparation and Handling of EETs-EETs are highly labile in the aqueous environment. Concentrated EET stock solutions were stored in aliquots at Ϫ70°C in acetonitrile. A fresh aliquot was used for each experiment. Just prior to use, the acetonitrile solution was dried under N 2 and EETs were resuspended in ethanol and kept on ice. Aliquots were removed and resuspended in DPBS as required for agonist stimulation. The final concentration of ethanol in all experiments did not exceed 0.1%.
Cranial Window-The acute cranial window technique and in vivo microscopy were utilized to examine the effect of 5,6-EET on rat cerebral arteriolar diameter, as described previously (27). Briefly, young adult male Sprague-Dawley rats were anesthetized with thiopental (75 mg/kg) and supplemented with pentobarbital. After completion of a tracheotomy, each rat was ventilated with room air. The end-expiratory CO 2 of each rat was continuously monitored with a capnometer (Transverse Medical Monitors, model 2200) and was maintained at approximately 30 mm Hg by adjusting the respirator rate and volume. Arterial blood pressure was measured via a cannula inserted into the right femoral artery. Arterial samples were periodically analyzed with a Corning Blood Gas Analyzer to ensure normal P a O 2 , P a CO 2 , and blood pH. A cannula was also inserted into the right femoral vein for systemic administration of supplemental anesthetic.
Pial arteries were visualized using a cranial window implanted into the scalp through a midline incision. The skin and fascia were retracted and a 3-mm diameter craniotomy was made over the left parietal cortex using a trephine. With the aid of a surgical microscope, microscissors were used to remove the dura and expose the pial surface of the brain. A 12-mm diameter cranial window frame with a 6.5-mm diameter glass window was implanted over the craniotomy. The cranial window was equipped with three openings. Two openings were used as an inlet and outlet for filling the space under the cranial window with test solutions. The inlet and outlet valves were positioned such that the test solutions flowed over the cortical surface as viewed through the cranial window. The third opening of the cranial window was connected to a Statham pressure transducer for continuous measurement of intracranial pressure. The outlet of the window was connected to plastic tubing whose open end was placed at a fixed level to give a constant intracranial pressure of 5 mm Hg throughout the experiment. The space under the window and the plastic tubing were filled with artificial CSF. This fluid was equilibrated with gas containing 5.9% CO 2 , 6.6% O 2 , and 87.5% N 2 , which produces pH and gas tensions in a normal range for CSF. The vehicle for all agents applied under the cranial window was artificial CSF. The diameter responses of three to five arterioles were studied in each rat using a Vickers image-splitting device as described previously (28). The responses of the arterioles in a given rat were averaged, and this single number was used to compute the average for a group of rats.
After implantation of the cranial window, baseline arteriolar diameter was established by washing the window with 1 ml of artificial CSF at 5-min intervals. At the end of each 5-min interval, baseline measurements of pial arteriolar diameter were recorded. Next, 5,6-EET was infused under the window in increasing concentrations (10 Ϫ9 -10 Ϫ5 M) in a total volume of 1 ml of artificial CSF at 5-min intervals. Pial arteriolar diameters were measured at 2 and 5 min after the infusion of each test solution.
Statistics-Both one-way and repeated measures ANOVA were performed and were followed by Tukey-Kramer comparisons to determine differences between the groups using Super Anova statistical software for Macintosh. A value of p Ͻ 0.05 was considered significant. donic acid is a component of CIF, then arachidonic acid should be released in response to depletion of intracellular calcium stores. Using [ 3 H]arachidonic acid-labeled astrocytes, we investigated the total release of arachidonic acid and metabolites in response to a 2-min stimulation with thapsigargin. Activation of the capacitative pathway with thapsigargin induced release of arachidonic acid (and/or metabolites) of 399 Ϯ 43% of basal ( Fig. 2A, filled bars). In calcium-free medium supplemented with EGTA, thapsigargin-stimulated release of arachidonic acid (and/or metabolites) was reduced to 233 Ϯ 13% of basal. Similar to our observations in U937 (8), these results suggest that a portion of the thapsigargin-stimulated arachidonic acid release is coupled to influx of extracellular calcium, while a portion is also coupled to depletion of intracellular calcium stores. Thus, PLA 2 activity may be associated with release of calcium from intracellular stores. Inhibition of cytochrome P450 with econazole, SKF525A, or MS-PPOH did not inhibit thapsigargin-induced arachidonic acid release. Inhibition of cPLA 2 with AAOCF 3 inhibited thapsigargin-stimulated arachidonic acid release to control levels.

Exogenous Arachidonic Acid Stimulates Calcium Influx in
Thapsigargin-induced Elevation of [Ca 2ϩ ] i Is Blocked by Inhibitors of Cytochrome P450 and cPLA 2 -We have previously shown that capacitative calcium influx requires the action of cPLA 2 and release of arachidonic acid in U937 cells (8). The results above suggest that arachidonic acid elevates [Ca 2ϩ ] i in the astrocyte, possibly through a P450 metabolite. Therefore, we investigated whether cPLA 2 and P450 were also linked to capacitative calcium influx in the astrocyte.
Thapsigargin is a pharmacological agent that inhibits the Ca 2ϩ -ATPase of the intracellular calcium store and activates capacitative calcium influx independent of phospholipase C (29). The thapsigargin-stimulated elevation of [Ca 2ϩ ] i is dependent on 2 sources of calcium, release from intracellular stores followed by capacitative influx of extracellular calcium. If capacitative calcium influx is mediated by release of arachidonic acid and subsequent metabolism via P450, then capacitative calcium influx should be inhibited by blockade of either cPLA 2 or P450. To test this we utilized econazole, SKF525A (30), and MS-PPOH (31), inhibitors of cytochrome P450 and AAOCF 3 , a selective inhibitor of cPLA 2 (32). A summary of these experiments appears in Fig. 2A and representative tracings in Fig. 2B.
In the astrocytes, thapsigargin elevated [Ca 2ϩ ] i to 432 Ϯ 62% of basal (Fig. 2, A and B, trace 1). To further dissect this response, the relative size of thapsigargin-releasable intracellular calcium stores in astrocytes was determined by stimulation with thapsigargin in calcium-free DPBS supplemented with 0.5 mM EGTA. Under these conditions, thapsigargin elevated [Ca 2ϩ ] i by only 196 Ϯ 23% of basal (Fig. 2, A and B, trace  2), approximately half that observed in experiments where extracellular calcium was 1 mM.
To assess the effect of cPLA 2 or P450 inhibitors on capacitative calcium influx, astrocytes were pretreated with 5 M AAOCF 3 , 10 M econazole, 10 M SKF525A, or 1 M MS-PPOH for 2 min. Inhibition of cPLA 2 with AAOCF 3 or inhibition of cytochrome P450 with econazole or SKF525A had no effect on thapsigargin-stimulated elevation of [Ca 2ϩ ] i in calcium-free medium, indicating that these compounds had no effect on release of calcium from intracellular stores (data not shown). In DPBS containing 1 mM extracellular calcium (Figs. 2, A and B, trace 5) inhibition of cPLA 2 with AAOCF 3 blocked capacitative calcium influx and produced [Ca 2ϩ ] i levels consistent with release of calcium from intracellular stores alone. In calcium replete medium, inhibition of cytochrome P450 with 10 M econazole or 10 M SKF525A also inhibited capacitative calcium influx (Fig. 2, A and B, traces 3 and 4). These two P450 inhibitors blocked the maximum thapsigargin-stimulated [Ca 2ϩ ] i elevation to the same extent as calcium-free medium, suggesting that capacitative Ca 2ϩ influx was inhibited. However, the sustained phase of [Ca 2ϩ ] i elevation was only partially inhibited. The cytochrome P450 enzymes comprise a large family of isoforms (33) of which econazole and SKF5125A are generalized inhibitors which block many isoforms. We therefore used an additional inhibitor of cytochrome P450, MS-PPOH (31). MS-PPOH is a "suicide substrate" inhibitor of P450 arachidonic acid epoxygenase, designed to resemble the substrate arachidonic acid and inactivate the enzyme. In rat renal microsomes, MS-PPOH was a potent and selective inhibitor of arachidonic acid epoxygenase activity (31). In the astrocyte, 1 M MS-PPOH inhibited thapsigargin-stimulated capacitative calcium influx to a level consistent with depletion of intracellular calcium stores only (Fig. 2, A and B, trace 6). Taken together, these results suggest that formation of CIF requires the combined actions of cPLA 2 and cytochrome P450 and may involve an epoxide of arachidonic acid.

5,6-EET Elevates [Ca 2ϩ ] i in Astrocytes-
If an EET is a component of CIF, then application of exogenous EET should stimulate capacitative calcium influx directly through SMOC, without affecting intracellular calcium stores. Therefore, we examined the effect of all 4 EETs on [Ca 2ϩ ] i in astrocytes. At a concentration of 10 Ϫ6 M, 8,9-(n ϭ 3), 11,12-(n ϭ 3), and 14,15-EET (n ϭ 6) had no effect on [Ca 2ϩ ] i (data not shown). As shown in Fig. 3, 5,6-EET induced a dose-dependent increase in [Ca 2ϩ ] i in astrocytes, which was elevated to 150% of basal by 10 Ϫ10 M 5,6-EET and to a maximum of 310% of basal by 10 Ϫ7 M. The nanomolar to picomolar activity of 5,6-EET suggests that this response was not a result of lipid-induced alterations in membrane fluidity. However, the dose-response curve for 5,6-EET was bell-shaped, and at higher concentration (10 Ϫ6 M) 5,6-EET consistently produced a submaximal response in all experiments. However, this decreased response was not statistically significant. Consistent with reports by Graier et al. SMOC, and should not induce release of calcium from intracellular stores. When astrocytes were placed in calcium-free DPBS supplemented with 0.5 mM EGTA to chelate residual calcium, 5,6-EET did not elevate [Ca 2ϩ ] i , suggesting that 5,6-EET-induced influx of extracellular calcium without effect on intracellular calcium stores (Fig. 4A, broken line).
In addition to SMOC calcium influx channels, many cells also have receptor-operated channels (34). Receptor-operated channels are activated by a ligand binding to its receptor, which directly controls the opening of a calcium channel. In contrast, SMOC are controlled by a second messenger (CIF). To discriminate between receptor operated channels and SMOC we first utilized the inhibitor of SMOC, SKF96365 (35). Fig. 4A shows a typical elevation in [Ca 2ϩ ] i observed in response to 156 nM 5,6-EET in astrocytes. The 5,6-EET-induced elevation in [Ca 2ϩ ] i was completely inhibited by pretreatment with 1 M SKF96365 (Fig. 4B). Nimodipine (50 M) had no effect on 5,6-EET-stimulated elevation of [Ca 2ϩ ] i , indicating 5,6-EET did not activate voltage-gated calcium channels. Additionally, pretreatment with 100 M neomycin for 15 min also had no effect on 5,6-EET-stimulated elevation of [Ca 2ϩ ] i , suggesting that 5,6-EET did not require phospholipase C activation to elevate [Ca 2ϩ ] i (data not shown).
Further support for 5,6-EET acting directly on SMOC is provided in Fig. 5A. In these experiments, astrocytes were treated with 1 M thapsigargin in 1 mM calcium-containing DPBS. After the maximum elevation in [Ca 2ϩ ] i was attained in response to thapsigargin, 156 nM 5,6-EET was added. If 5,6-EET activated receptor-operated channels, it is likely that an additional increase in [Ca 2ϩ ] i would be produced (4). However, there was no further elevation in [Ca 2ϩ ] i , providing further evidence that 5,6-EET acts directly on SMOC.
Inhibition of Thapsigargin-stimulated [Ca 2ϩ ] i Elevation by Econazole Is Overcome by Addition of 5,6-EET-If 5,6-EET is a CIF, then inhibition of calcium influx with the P450 inhibitor econazole should be reversed by addition of exogenous 5,6-EET. To test this, Fura-2 loaded astrocytes were pretreated with 10 M econazole for 2 min as shown in Fig. 5B. Next, cells were stimulated with 1 M thapsigargin. As described previously, in the presence of econazole the thapsigargin-stimulated eleva-tion in [Ca 2ϩ ] i was reduced to a level consistent with depletion of intracellular calcium stores alone. Upon attaining the maximum sustained [Ca 2ϩ ] i in response to thapsigargin, 156 nM 5,6-EET was added. As can be seen in Fig. 5B 5. A, 5,6-EET-stimulated [Ca 2ϩ ] i elevation is not additive to that of thapsigargin. Fura-2 loaded astrocytes were stimulated with 1 M thapsigargin as indicated. After the maximal response to thapsigargin had been attained, 156 nM 5,6-EET was added. No further elevation in [Ca 2ϩ ] i was observed. Results shown are representative tracings from three separate experiments performed on different days. B, 5,6-EET overcomes econazole-inhibition of thapsigargin-stimulated [Ca 2ϩ ] i elevation. Fura-2 loaded astrocytes were pretreated with 10 M econazole, followed by 1 M thapsigargin. Note that the thapsigargin response is reduced to the same degree as shown in Fig. 2. After the maximal response to thapsigargin was attained, 156 nM 5,6-EET was added.
[Ca 2ϩ ] i was measured as described in the legend to Fig. 1 4. 5,6-EET-stimulated [Ca 2؉ ] i elevation is blocked by SKF96365. A, Fura-2 loaded astrocytes were stimulated with 156 nM 5,6-EET at the indicated time point in DPBS with (-) or without (---) 1 mM Ca 2ϩ . B, separate cultures of Fura-2 loaded astrocytes in DPBS containing 1 mM Ca 2ϩ were pretreated with SKF96365 followed by 5,6-EET as shown. Tracings are representative of four separate experiments.
in Astrocytes-Our previous work (20,36) and that of others (37,38) indicates that the primary EET metabolites in astrocytes are 5,6-and 14,15-EET. As described above, 14,15-EET had no effect on [Ca 2ϩ ] i in astrocytes. However, in six separate experiments, we found that pretreatment of astrocytes with 156 nM 14,15-EET blocked 5,6-EET-induced elevation of [Ca 2ϩ ] i . A representative tracing of these results is shown in Fig. 6. Subsequent addition of a second dose of 156 nM 5,6-EET counteracted the inhibitory effect of 14,15-EET.

5,6-EET May Also Act as a Paracrine
Signal-Our present results in astrocytes suggest that 5,6-EET may be the elusive CIF of the capacitative calcium influx pathway. We have previously shown that 5,6-EET causes dilation of cerebral arterioles when applied topically under an in vivo cranial window in rabbits (36). Leffler et al. (39) have recently shown that 5,6-EET is also a potent dilator of piglet cerebral arterioles. However, the effects of EETs may differ with species. Since the astrocytes used in the above tissue culture experiments were of rat cortical origin, we tested the activity of 5,6-EET on vascular diameter in rat brain using the cranial window technique (27). As shown in Fig. 7, 5,6-EET caused dilation of cerebral arterioles when topically applied to the brain surface. Curiously, the dose-response of 5,6-EET on pial arteriolar diameter, like 5,6-EET-induced [Ca 2ϩ ] i elevation in astrocytes (Fig. 3), was bell-shaped. 5,6-EET produced a linear increase in arteriolar diameter over the concentration range 10 Ϫ9 -10 Ϫ6 M. However, a higher dose of 10 Ϫ5 M produced a very significantly reduced dilator response. A similar trend at the higher concentration of 5,6-EET has been reported by Leffler et al. (39) in the piglet cerebral microcirculation. This reduced dilator response at 10 Ϫ5 M suggests inhibition of the dilator stimulus or activation of an opposing constrictor response. DISCUSSION Since the first reports identifying store-controlled or capacitative calcium influx, the link between intracellular calcium stores and influx of calcium through the plasma membrane has remained obscure. Randriamampita and Tsien (7) isolated a CIF-like substance from stimulated Jurkat cells, which was of molecular weight less than 500 and moderately hydrophobic. Parekh et al. (40) reported that CIF formation involved a phosphatase and a diffusible second messenger. Fasolato et al. (41) have presented evidence suggesting that formation of CIF involved hydrolysis of GTP and possibly a low molecular weight G protein. It has been hypothesized that CIF represents a complex of several different biochemical components (4, 42). Our previous results suggest that the link to CIF requires the activation of cPLA 2 and release of arachidonic acid from cellular phospholipids in U937 cells (8). Other groups have reported similar correlations between arachidonic acid release and calcium influx (43,44). Cytochrome P450 activity has also been coupled to the formation of CIF (45,46) and is one of the major pathways through which arachidonic acid is metabolized. In the present report, we present evidence suggesting that the link to CIF involves 5,6-EET, a cytochrome P450 metabolite of arachidonic acid.
Several lines of evidence support a role for 5,6-EET as a component of CIF. First, CIF appears to be ephemeral in nature, having a short half-life. This property has made its identification difficult (4, 7). 5,6-EET is a relatively short-lived metabolite of arachidonic acid, which is rapidly degraded in the aqueous environment, consistent with the properties of CIF. Second, in order to function as CIF, arachidonic acid must first be released from its intracellular phospholipid storage pool through the action of PLA 2 , followed by metabolism via cytochrome P450. We have previously reported a coupling between cPLA 2 , arachidonic acid release, and depletion of calcium from intracellular stores (8). In the present report, we demonstrate that thapsigargin-stimulated arachidonic acid release is coupled to release of calcium from intracellular stores in the astrocytes. Consistent with our data from U937 cells, inhibition of cPLA 2 activity in the astrocyte effectively blocked capacitative calcium influx, but not release of intracellular calcium stores. Similar to reports from other laboratories (10,11,45), inhibition of cytochrome P450 with econazole or SKF525A also inhibited capacitative calcium influx. Furthermore, MS-PPOH, a specific inhibitor of the P450 system that metabolizes arachidonic acid to 5,6-EET, also inhibited capacitative calcium influx, to a level consistent with release of calcium from intracellular stores alone. Thus, the two enzymatic systems which produce 5,6-EET are both coupled to capacitative calcium influx.
We have also shown that 5,6-EET dose dependently activates calcium influx in the astrocyte at nano-to picomolar concentrations. The influx of calcium initiated by 5,6-EET is consistent with CIF for several reasons. First, 5,6-EET did not induce calcium influx in calcium-free medium, suggesting that it has no effect on intracellular calcium stores. Second, 5,6-EET-stimulated calcium influx was blocked by inhibition of SMOC with SKF96365. Third, the elevation in [Ca 2ϩ ] i produced by 5,6-EET was not additive to that produced by thapsigargin. Fourth, exogenous 5,6-EET could rapidly overcome econazole inhibition of thapsigargin-stimulated [Ca 2ϩ ] i elevation. Finally, the 5,6-EET-stimulated [Ca 2ϩ ] i elevation was unaffected by blockade of voltage-gated channels with nimodipine or inhibition of phospholipase C with neomycin.
Graier et al. (10) have elaborately demonstrated that 5,6-EET is consistent with CIF in endothelial cells. This group has shown that inhibition of P450 blocked capacitative calcium influx, while induction of P450 with dexamethasone/clofibrate enhanced thapsigargin-stimulated [Ca 2ϩ ] i elevation. Thus, reports from endothelial cells (10), U937 cells (8), and our present report in astrocytes all support the hypothesis that 5,6-EET is coupled to CIF.
In addition to the effects of 5,6-EET on elevation of [Ca 2ϩ ] i , we have also found that 14,15-EET may also participate in 5,6-EET-mediated [Ca 2ϩ ] i regulation. 14,15-EET effectively blocked 5,6-EET-induced elevation of [Ca 2ϩ ] i . This effect could be overcome with additional doses of 5,6-EET. However, loss of 14,15-EET's ability to prevent 5,6-EET-induced calcium influx could alternatively be explained by a time-dependent inactivation of 14,15-EET due to its metabolism to the vicinal diol by epoxide hydrolase (20). Malcolm et al. (47) have previously reported that 14,15-EET inhibited capacitative calcium entry in platelets. We (20) and others (15,48) have previously reported that 14,15-EET is synthesized and released by astrocytes. Our laboratory has also reported that 14,15-EET can be taken up and incorporated into cellular phospholipids and may act as a regulator of several isoforms of protein kinase C (49). It is tempting to speculate that 14,15-and 5,6-EET exert opposing actions on capacitative calcium influx, thereby acting as regulatory elements for [Ca 2ϩ ] i homeostasis.
Although 5,6-EET is rapidly degraded in vitro, its life span in a lipid environment is unknown. Studies by Randriamampita and Tsein (7) suggest that CIF can be released from cells and initiate capacitative calcium influx in unstimulated cell cultures. In vivo, release of 5,6-EET may act as a paracrine signal for other nearby cells. Alkayed et al. (48) have shown that EETs are released from astrocytes stimulated with glutamate and that expression of P450 2C11 protein is markedly elevated by glutamate exposure. We and others have shown that in the astrocyte, glutamate stimulation is closely coupled to activation of phospholipase C and capacitative calcium influx (21,50). According to our present results, activation of this pathway with glutamate would result in formation of 5,6-EET as CIF. Since astrocytes are known to respond to neuronal activity, we hypothesize that 5,6-EET may be released from astrocytes and act on the brain microvasculature to induce vasodilation in response to changes in neuronal activity. In support of this hypothesis, we have demonstrated that application of exogenous 5,6-EET causes dilation of cerebral vessels in the rat. We have reported similar results in the rabbit using 5,6-EET synthesized by rat cortical astrocytes (36). In the rabbit, 5,6-EETinduced vasodilation was blocked by indomethacin. Leffler (39) have recently reported that 5,6-EET induced vasodilation in newborn pig pial arterioles. They also found that 5,6-EETinduced dilation was blocked with indomethacin and restored by a low dose (10 Ϫ12 M) of iloprost, a prostaglandin I 2 mimetic. In the study by Leffler (39), 5,6-EET did not elevate 6-ketoprostaglandin F 1␣ levels in the CSF, suggesting that 5,6-EET did not directly activate prostaglandin I 2 synthesis. Leffler suggests that 5,6-EET induced vasodilation requires a "permissive" concentration of prostaglandin. Prostaglandins, as well as other agonists, often act through G-proteins coupled to regulation of cyclic AMP. It has been reported that alterations in cAMP levels may act to modulate Ins(1,4,5)P 3 -mediated signaling and capacitative calcium influx (51,52). Thus, we speculate that the permissive effect of prostaglandins on 5,6-EET-induced vasodilation are not directly due to a cyclooxygenase metabolite of 5,6-EET, but rather that a prostanoid may modulate a 5,6-EET-induced signaling component of capacitative calcium influx.
Interestingly, we found that the [Ca 2ϩ ] i elevation and arteriolar dilation induced by 5,6-EET were diminished at the highest concentrations of 5,6-EET, resulting in a bell-shaped dose response (Figs. 3 and 7). Similar results were observed by Leffler (39) for 5,6-EET-induced arteriolar dilation in the piglet. The exact reason for these bell-shaped dose-response curves is uncertain. At the cellular level we can speculate that as 5,6-EET activates SMOC calcium influx and [Ca 2ϩ ] i rises above a certain level, additional opposing intracellular mechanisms may be activated, such as reduced Ins(1,4,5)P 3 -receptor sensitivity (53)(54)(55). At the in vivo level we speculate that as intracellular calcium is initially elevated in vascular smooth muscle by 5,6-EET, calcium-activated K ϩ channels induce hyperpolarization and smooth muscle relaxation. However, as [Ca 2ϩ ] i rises above a certain point, calcium-mediated contractile mechanisms may be activated, which counteract the initial dilation.
In summary, we report that 5,6-EET may be a component of CIF in the astrocyte. Furthermore, our results suggest that 5,6-EET released by astrocytes may regulate pial arteriole diameter. Thus, Ins(1,4,5)P 3 -mediated signaling in astrocytes may be coupled to regulation of cerebral blood flow.