12-Keto-Eicosatetraenoic Acid A Biologically Active Eicosanoid in the Nervous System of Aplysia

The lipoxygenase product 12-hydroperoxy-5,8,10,14-eicosatetraenoic acid (1 2- HPETE), simulates the synaptic responses produced by the modulatory transmitter histamine and the neuroactive peptide Phe-Met-Arg-Phe-amide (FMRFamide) in identified neurons of the marine mollusk Aplysia caIiJornica."* The 12-lipoxygenase pathway has not yet been fully characterized, but 12-HPETE is known to be metabolized further. Therefore, we began to search for other metabolites in order to investigate whether the actions of 1ZHPETE might require its conversion to other active products. We have identified 12-keto-5,8,10,14-eicosatetraenoic acid (12-KETE) as a metabolite of 12-HPETE formed by Aplysia nervous tissue. 12-KETE was identified in incubations of the tissue with arachidonic acid using HPLC, UV spectrometry, and gas-chromatography/mass spectrometry. ['HI 12-KETE is formed from endogenous lipid stores in nervous tissue, labeled with [ 'Hlarachidonic acid upon stimulation by application of histamine. In L14 and L10 cells, identified neurons in the abdominal ganglion, applications of 12-KETE elicit changes in membrane potential similar to those evoked by histamine. Another metabolite of I2-HPETE, 12(s)-hydroxy-5,8,10,14-eicosatetraenoic acid [ 12(S)-HETE], is inactive. These results support the hypothesis that 12-HPETE and its metabolite, 12-KETE, participate in transduction of histamine responses in Aplysia neurons.


INTRODUCTION
In the neural tissue of ApEysia, the neurotransmitter histamine and the peptide FMRFamide evoke release of 12-hydroxyeicosatetraenoic acid ( 1 2-HETE), a stable end product of the 12-lipoxygenase p a t h~a y . ' .~ The short-lived precursor of I2-HETE. 1 2-HPETE, simulates electrophysiological responses induced by FMRFamide in sensory cells394 and by histamine in L14 cells: suggesting that this hydroperoxide or one of its metabolites serves as intracellular signal to mediate the synaptic actions of FMRFamide and histamine. Several novel metabolites derived from 12-HPETE have recently been identified in mammalian tissue^.^" Therefore, it is possible that 12-HPETE produces its effects in Aplysia neurons only after conversion to an active metabolite.
Here, we report that a metabolitc of 12-HPETE, 12-KETE, is released when Aplysia nervous tissue is stimulated by applying histamine. We also show that application of 12-KETE, like 1 2-HPETE. produces electrophysiological responses similar to those evoked by histamine in two identified '4plysia neurons, L10 and L14.

EXPERIMENTAL PROCEDURES
Aplysia weighing 70-200 g (Howard Hughes Medical Institute Maricuiture Research Facility, Woods Hole Oceanographic Institution, Woods Hole, MA; and Marinus, Sand City, C A ) were kept in aquaria at 15 "C. Homogenates of nervous tissue and isolated neural components (cell bodies and neuropil) were prepared as described previously.' bFellow of the Louis and Rose Klosk Fund. 'Address for correspondence: Dt. Daniele Piomelli, Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University, 1230 York Avenue, New York, New York 10021.

Extraction of Lipids
Acetone (0-4OC) was added to homogenates (l:l, v/v) and the precipitate removed by low-speed centrifugation. Metabolites were extracted twice with ethyl acetate (2 vol) after acidifying the supernatant to pH 3.5. The organic layers were combined, dried over sodium sulfate, and evaporated under vacuum. Samples from experiments with prelabeled nervous tissue were extracted with ethyl acetate without prior addition of acetone.

High-Performance Liquid Chromatography
Analytical normal-phase HPLC was performed using a silica column (250 x 4.6 mm, 5 pm; Supelco, Bellefonte, PA) eluted isocratically with hexane:isopropanol:acetic acid (98:2:0.1, v/v/v) at a flow rate of 1 mI/min. Absorbance was monitored continuously at 270 nm, and full UV spectra were obtained with a diode-array spectrophotometer (Hewlett-Packard 1090M, Palo Alto, CA); 30-sec fractions were collected and radioactivity counted by liquid scintillation. For purifications on a preparative scale, we used a Polygosil silica column (500 x 10 mm, 10 pm; Alltech, Deerfield, IL) eluted with the same solvent system at a flow rate of 3 ml/min. Reversed-phase HPLC was performed with a Nucleosil C18 column (250 x 4.6 mm, 5 pm, Alltech) eluted isocratically with methano1:water:acetic acid (65:35:0.1) a t a flow rate of 1 ml/min; absorbance was monitored a t 280 nm. In some experiments, carbonyl groups were reduced to alcohols by adding 1-2 mg of sodium borohydride to samples dissolved in ethanol (0.1 ml) and incubating for 15 min a t 0-4 "C. Samples were then filtered through glass wool and dried under nitrogen. The resulting products were separated by normal-phase HPLC as described above with the UV detector set at 235 nm.

Gas Chromatography/ Mass Spectrometry
Metabolites were purified by preparative normal-phase HPLC (see above), and converted to the methyl ester by treating the purified material with a n excess of ethereal diazomethane for 2 min. To prepare the pentafluorobenzyl (PFB) esters, we incubated samples with pentafluorobenzyl bromide (35% in 10 pl acetonitrile) and diisopropylethylamine (10 p l ) diluted with acetonitrile (30 p l ) for 10 min a t room temperature. To prepare methoxime derivatives, the esterified samples were exposed to methoxylamine hydrochloride (1% in pyridine, 20 pl) for 1 h a t 60 "C.
Analyses were performed on a Hewlett-Packard 5987A GC/MS fitted with an HP-1 capillary column (1 2 m, Hewlett-Packard, Palo Alto, CA) using helium as the carrier gas. For electron impact analyses the column temperature was programmed from 150 to 250 "C a t a rate of 30 OC/min. We kept the injector at 250 "C and the source at 200 "C. Carrier flow was regulated a t a constant head pressure of 52 kPa and the voltage kept a t 25 eV. Negative ion chemical ionization analyses were done using methane as the ionizing gas (source pressure approximately 0.8 torr). We kept the injector at 250 O C and the source at 150 "C. Oven temperature was kept a t 60 "C for 1 min and then raised to 320 "C at a rate of 30 OC/min.

Intracellular recordings
Abdominal ganglia were pinned ventral side up to Sylgard, a silicone plastic (Dow Chemical, Midland, MI), in a chamber continuously superfused with supplemented artificial seawater a t room temperature. The connective tissue sheath was removed by dissection; L14 neurons, identified as previously described:" were impaled with one or two glass recording microelectrodes filled with potassium citrate (1 --5 MQ resistance).
Compounds to be tested were ejected with pressure from a glass micropipette placed approximately 0.5 mm from the cell body. Samples from stock solutions of 12-HPETE, 12-KETE, and I2(S)-HETE (kept in hexane or ethanol at --20 "C) were dried under nitrogen, reconstituted in the seawater, and sonicated for 15 sec.

RESULTS AND DISCUSSION
Several biologically active molecules can be formed from 12-HPETE. Among the metabolites that have been identified thus far are 12-KETE,' 12-ODTE,637 and several isomeric epoxy alcohol^.^^"^'^ The possibility that 12-HPETE must be metabolized to produce its action in Aplysia neurons is suggested by an observation of Belardetti et al.:4 the increased opening of K$ channels evoked by 12-HPETE occurs only in cell-attached (but not in cell-free) patches of sensory neuron membranes. This suggests that a cytosolic component, possibly an enzyme. is required to metabolize the hydroperoxy acid further.
We describe here a novel bioactive metabolite formed in nervous tissue of Aplysia, the keto-acid 12-KETE. Identification was carried out by HPLC, U V spectromctry, and GC/MS in lipid extracts of the nervous tissue incubated with exogenous arachidonic acid or 12-HPETE. Homogenates of Aplysia nervous tissue were incubated with arachidonic acid (50 yM, 30 min), and the metabolites formed were analyzed by normal-phase HPLC (FIG. 1 A). Several unidentified components with absorption maxima at 270 nm were observed (compounds a,, a?, and h).
The UV spectra of compounds a, and a, (FIG. 1 A, inset) indicated the presence of a dienone or dienal chromophore, with maximal absorbance at 273 nm for a,, and 271 nm for a2. After they were purified by normal-phase HPLC, we also analyzed compounds a, and o2 by reversed-phase HPLC, where they eluted as a single component (FIG. 1B). UV spectral analysis (FIG. lB, inset) showed a pronounced bathochromic shift in absorbance (A max = 280 nm). normal-phase HPLC. Extracted lipids were fractionated on a silica column eluted with hexane:isopropanol:acetic acid (98:2:0.1, v/v/v) at 1 ml/min. U V absorbance was monitored at 270 nm. B reversed-phase HPLC analysis of compounds a, and ( I * after they had been purified by normal-phase HPLC. Fractions containing 12-KETE, reduced to dryness and redissolved in the mobile phase, were applied to a Nucleosil C18 column eluted with methano1:water:acetic acid (65:35:0.1, V/V/V) at 1 ml/min. UV absorbance was monitored at 280 nm. C normal-phase HPLC of the alcohols resulting from reduction of a, and a, with sodium borohydride. These alcohols were fractionated on a silica column as described above. UV absorbance was monitored at 235 nm. Insels show spectra obtained with a flow-through diode-array spectrophotometer of the compounds in the HPLC mobile phase (see EXPERIMENTAL PROCEDURES). (From Piomelli et  reduction of the compounds with sodium borohydride yields two alcohols, 12hydroxy-5,8,10,14( ZZEZ)-eicosatetraenoic acid methyl ester (12-HETE methyl ester) and its geometric isomer, 12-hydroxy-5,8,10,14( 2EEZ)-eicosatetraenoic acid.
As shown by radiolabeling experiments, compounds a, and a, are derived from arachidonic acid. Normal-phase HPLC analysis resolved two major radioactive components (FIG. 2): the first contained ['Hlarachidonate added as substrate as well as [3H] 12-HETE, and the second corresponded to compounds a, and a,. Formation of these products was inhibited (>95%, n = 2) by incubation of the homogenates with the lipoxygenase inhibitor nordihydroguaiaretic acid (NDGA) (30 pM), but not by aspirin, a cyclooxygenase blocker (0.5 mM). This suggests that a 12-lipoxygenase enzyme catalyzes the biosynthesis of this metabolite from arachidonic acid.
In accord with this idea, we found that compounds a, and a2 could also be formed when nervous tissue was incubated with 12-HPETE (50 pM, 10 min, data not shown). Boiling the tissue did not affect the conversion of exogenous 12-HPETE to these compounds, however, confirming previous reports that exogenous 12-HPETE can be converted nonenzymatically: conversion of fatty acid hydroperoxides to keto-acids and aldehydes can be catalyzed by hematin or by heme-containing protein^.^,'^.'^ Compounds a, and a2 have the HPLC retention values and UV spectra of authentic 12-KETE prepared by incubating 12-HPETE with hemoglobin or by oxidation of 12-HETE with manganese dioxide. This identification was confirmed by GC/MS. Negative-ion chemical ionization analysis of pentafluorobenzyl (PFB) esters of metabolites a, and a2 showed that these compounds eluted together and produced a mass spectrum identical to that of authentic 12-KETE, with only one prominent ion a t m / z 317 (M -181, lossof PFB) (FIG. 3A).
Additional structural analysis was carried out by electron impact GC/MS. The methyl esters of compounds a , and a2 were eluted from the GC with a carbon chain value of 21.8 (FIG. 3B), as previously reported for 12-KETE methyl ester.7 Ions of high intensity were observed at m / z 332 (M+), 314 (M' -18, loss of H20), 301 12-KETE 1 RETENTION TIME (mid

Stimulation of ['H]72-KETE Production by Neurotransmitter
Histamine stimulates the generation of [ 'H] 12-HETE in Apfysia nervous tissue prelabeled with [ 3H]arachidonic acid.' Using a similar experimental protocol, we found that application of histamine caused nearly a 10-fold increase in radioactivity associated with 12-KETE compared to controls (p < 0.05, Student's t-test) (FIG. 4).
In addition to 12-KETE, other products of 12-HPETE are formed in Apfysia nervous tissue, including the short-chain aldehyde 12-oxododecatrienoic acid (12-ODTE) and two epoxy alcohols, 8-hydroxy-11,12-epoxyeicosatrienoic and 10- The reason for the difference in metabolites released by the two physiological treatments still remains to be determined. One possibility is that activation of specific receptors may result in the release of specific metabolites because the receptors activated by histamine might be a different subset of histaminergic receptors from those activated by the transmitter released endogenously by L32. Alternatively, while all the known actions of L32 cells are simulated by histamine and L32 PSPs are sensitive to cimetidine, a histamine antagonist in A p l y s i~, " * '~ it is still uncertain whether L 32 cells are definitively histaminergic.'' In preliminary experiments, we found that intracellular stimulation of C2, an identified histaminergic neuron in the cerebral ganglion of A p l y~i u , '~~~' results in release of ['HI 12-KETE. Stimulating C2 did not evoke the formation of ['HI 12-ODTE or of the epoxy alcohols, however. These results further support the idea that activation of specific histamine receptors at some synapses leads to the formation of 12-KETE.

Physiological Activity of 12-KETE on Identified Aplysia Neurons
Pharmacological experiments with L14 and L10, neurons of the abdominal ganglion, suggest that 12-KETE participates in the intracellular transduction of some of the actions of histamine. Each identified cell shows different and characteristic electrophysiological responses to histamine. In L14, histamine rapidly depolarizes the membrane, which is typically followed by a longer-lasting hyperpolarization." In the majority of neurons tested, applications of 12-HPETE or 12-KETE (1-2 nmol) from an extracellular puff micropipette produced a response similar to that evoked by histamine (FIG. 5 , TABLE 1). Similar puffs of I2(S)-HETE were ineffect ive .
L10, a mixed-action neuron regulating heart and kidney function,2' responds to histamine with a slow hyperpolarization, caused by increased K' conductance and decreased Ca2+ conductance.'* A similar inhibitory response in L10 cells was produced by 12-KETE (TABLE 1). In only 30% of the cells tested was IZHPETE effective, however; and 12(S)-HETE was again ineffective (TABLE 1). A possible explanation is that, as applied by the puffing micropipette, the metabolites are not completely accessible to critical sites in L10 at the concentrations used. Further experiments, using L10 neurons in culture, would be useful to test this idea.
The biological actions of 12-KETE that we have observed are in agreement with the hypothesis that conversion of 12-HPETE to the keto-acid is necessary for some of the effects of the hydroperoxy acid. Voltage-clamp and patch-clamp studies would show whether these 12-lipoxygenase products affect the same ion channels modulated by the endogenous transmitter. 1-2 nmoles of a test substance were ejected by pressure (5 sec, 6 psi) from a glass micropipette situated about 0.5 mm from the cell body of a L14 impaled with a voltage-sensitive microelectrode. Histamine (HIST) and 12-keto-5,8,10,14-eicosatetraenoic acid (12-KETE) elicited early depolarizing responses followed by a small slow hyperpolarization. 12(S)-hydroxy-5,8,10,14eicosatetraenoic acid [ 12(S)-HETE] was ineffective in changing the membrane potential. The histamine response measured in this particular specimen was larger and longer-lasting than the response to 12-KETE (note difference in calibration of the electrophysiological traces). (From Piomelli et Reprinted by permission from the Journal of Biological Chemistry.)

SUMMARY
The lipoxygenase product 12-hydroperoxy-5,8,10,14-eicosatetraenoic acid (1 2-HPETE), simulates the synaptic responses produced by the modulatory transmitter histamine and the neuroactive peptide Phe-Met-Arg-Phe-amide (FMRFamide) in identified neurons of the marine mollusk Aplysia caIiJornica."* The 12-lipoxygenase pathway has not yet been fully characterized, but 12-HPETE is known to be metabolized further. Therefore, we began to search for other metabolites in order to investigate whether the actions of 1ZHPETE might require its conversion to other active products. We have identified 12-keto-5,8,10,14-eicosatetraenoic acid (12-KETE) as a metabolite of 12-HPETE formed by Aplysia nervous tissue. 12-KETE was identified in incubations of the tissue with arachidonic acid using HPLC, UV spectrometry, and gas-chromatography/mass spectrometry. ['HI 12-KETE is formed from endogenous lipid stores in nervous tissue, labeled with [ 'Hlarachidonic acid upon stimulation by application of histamine. In L14 and L10 cells, identified neurons in the abdominal ganglion, applications of 12-KETE elicit changes in membrane potential similar to those evoked by histamine. Another metabolite of I2-HPETE, 12(s)hydroxy-5,8,10,14-eicosatetraenoic acid [ 12(S)-HETE], is inactive. These results support the hypothesis that 12-HPETE and its metabolite, 12-KETE, participate in transduction of histamine responses in Aplysia neurons.