Metabolism of Arachidonic Acid to Epoxyeicosatrienoic Acids. Hydroxyeicosatetraenoic Acids, and Prostaglandins in Cultured Rat Hippocampal Astrocytes

Abstract: We have recently shown that brain slices are capable of metabolizing arachidonic acid by the epoxy-genase pathway. The purpose of this study was to begin to determine the ability of individual brain cell types to form epoxygenase metabolites. We have examined the astrocyte epoxygenase pathway and have also confirmed metabolism by the cyclooxygenase and lipoxygenase enzyme systems. Cultured rat hippocampal astrocyte homogenate, when incubated with radiolabeled [³H]-arachidonic acid, formed products that eluted in four major groups designated as R_17–30, R_42–50, R_51–82, and R_83–90 based on their retention times in reverse-phase HPLC. These fractions were further segregated into as many as 13 peaks by normal-phase HPLC and a second reverse-phase HPLC system. The principal components in each peak were structurally characterized by gas chromatography/electron impact-mass spectrometry. Based on HPLC retention times and gas chromatography/electron impactmass spectrometry analysis, the more polar fractions (R_17–30) contained prostaglandin D₂ as the major cyclooxygenase product. Minor products included 6-keto prostaglandin F_1α, prostaglandin E₂, prostaglandin F_2α, and thromboxane B₂. Fractions R_42–50, R_51–82. and R_83–90 contained epoxygenase and lipoxygenase-like products. The major metabolite in fractions R_83–90 was 5, 6-epoxyeicosatrienoic acid (EET). Fractions R_51–82 contained 14, 15-and 8, 9-EETs, 12-and 5-hydroxyeicosatetraenoic acids, and 8, 9-and 5, 6-dihydroxyeicosatrienoic acids (DHETs). In fractions R_42–50, 14, 15-DHET was the major product. When radiolabeled [³H]14, 15-EET was incubated with astrocyte homogenate, it was rapidly metabolized to [³H]14, 15-DHET. The metabolism was inhibited by submicromolar concentration of 4-phenylchalcone oxide, a potent inhibitor of epoxide hydrolase activity. Formation of other polar metabolites such as triols or epoxyalcohols from 14, 15-DHET was not observed. In conclusion, astro-cytes readily metabolize arachidonic acid to 14, 15-EET, 5, 6-EET, and their vicinal-diols. Previous studies suggest these products may affect neuronal function and cerebral blood flow.     

Phloretin as an Antagonist of Prostaglandin F2α Receptor in Cultured Rat Astrocytes

Abstract: The effect of phloretin on prostaglandin (PG) F_2α-induced phosphoinositide hydrolysis and elevation of intracellular Ca²⁺ concentration was examined in cultured rat astrocytes. Phloretin inhibited PGF_2α (1 μ M )-induced phosphoinositide hydrolysis in a concentration-dependent manner with an IC₅₀ value of 16 μ M . The inhibitory action of phloretin was specific for PGs. The addition of increasing concentrations of phloretin caused progressive shifts of the dose-response curves of PGF_2α to the right. In digitoninpermeabilized astrocytes, phloretin (100 μ M ) inhibited the stimulation induced by PGF_2α (1 μ M ) plus GTPγS (50 μ M ) without affecting that induced by GTPγS alone. PGF_2α at 1 μ M transiently increased astrocytic intracellular Ca²⁺ concentration in 39% of the cells tested. The response was completely blocked by 100 μ M phloretin and the calcium response recovered again after washing out phloretin. These results suggest that phloretin is an antagonist of PGF_2α receptor linked to phospholipase C in astrocytes.     

Stimulation of Phospholipase D Activity by Phorbol Esters in Cultured Astrocytes

The phorbol ester 12-O-tetradecanoylphorbol 13-acetate (TPA) was found to stimulate phospholipase D activity in cultured primary astrocytes. Both the hydrolysis and the transphosphatidylation reaction catalyzed by phospholipase D were studied in cells labeled with [3H]glycerol. Phosphatidic acid (PA) synthesis was increased after addition of 100 nM TPA. When ethanol was present in the cell culture medium, phosphatidylethanol (Peth), a product of phospholipase D-catalyzed transphosphatidylation, was formed. The half-maximum effective concentrations (EC50) of TPA were 25 nM for PA increase as well as for Peth formation. The formation of Peth in ethanol-treated cells was accompanied by an inhibition of the TPA-induced increase in labeled PA. Increasing ethanol concentrations led to an increase in [3H]Peth and a decrease in [3H]PA. A protein kinase C inhibitor, 1-(5-isoquinolinesulfonyl)-2-methylpiperazine (H7), inhibited both the synthesis of PA and the formation of Peth observed after TPA addition to the astrocytes. Dioctanoyl-glycerol (100 microM) stimulated the formation of Peth in the presence of ethanol. In addition to the induction of Peth formation in astrocytes, TPA induced Peth formation in ethanol-treated neurons. The present results indicate that phospholipase D activity is stimulated by TPA in cultured primary brain cells. Modulation of phospholipase D activity by protein kinase C is a mechanism that may be important in signal transduction cascades.     

Induction of Manganese Superoxide Dismutase by Cytokines and Lipopolysaccharide in Cultured Mouse Astrocytes

Abstract: To determine whether cytokines or lipopolysaccharide (LPS) are involved in the induction of superoxide dismutase (SOD) in the nervous system, we examined the effects of these substances on the levels of SOD in cultured mouse astrocytes. Treatment of astrocytes with 10² to 10⁴ U/ml tumor necrosis factor-α for 3 days increased the levels of Mn SOD in a dose- and time-dependent manner to as much as six times the level under nontreated conditions. Treatment with 1.0 µg/ml LPS for 3 days elicited a fourfold increase in levels of Mn SOD, and the effect of LPS was also dose dependent. Furthermore, Mn SOD in astrocytes was induced by a 3-day exposure to interleukin-1α at concentrations of 10² or 10³ U/ml. However, these stimuli had no effect on levels of copper-zinc SOD (Cu/Zn SOD) in astrocytes. By contrast, interferon-γ did not change the levels of either Mn or Cu/Zn SOD in the cells. The results indicate that the selective induction of Mn SOD by cytokines and LPS, which has been observed in nonnervous tissues, may also occur in nervous tissues. The induction of Mn SOD may represent a mechanism for protection of cells from oxidative stress.     

Insulin Binding and Effects on Pyrimidine Nucleoside Uptake and Incorporation in Cultured Mouse Astrocytes

Binding of 125I-insulin to primary cultures of differentiated mouse astrocytes was time-dependent, reaching equilibrium after 2 h at 22 degrees C, with equilibrium binding corresponding to 20.79 fmol/mg of protein, representing approximately 5,000 occupied binding sites/cell. The half-life of 125I-insulin dissociation at 22 degrees C was 2 min, with an initial dissociation rate constant of 4.12 X 10(-2) s-1. Dissociation of bound 125I-insulin was not accelerated significantly in the presence of unlabeled insulin (16.7 microM). Porcine and desoctapeptide insulins competed for specific 125I-insulin binding in a dose-dependent manner, whereas growth hormone, glucagon, and somatostatin did not. For porcine insulin, Scatchard analysis suggested multiple-affinity binding sites (high-affinity Ka = 4.92 X 10(8) M-1; low-affinity Ka = 0.95 X 10(7) M-1). After incubation with insulin (0.5 microM) for 2 h at 37 degrees C, increases above basal values of 254 +/- 23 and 189 +/- 34% for [3H]uridine uptake and incorporation, respectively, were observed. After incubation with insulin (0.5 microM) for 24 h at 37 degrees C, there were increases of 145 +/- 6% for [3H]thymidine uptake and 166 +/- 11% for thymidine incorporation. Basal and stimulated uridine and thymidine uptake and incorporation were inhibited by 50 microM dipyridamole. These studies confirm that mouse astrocytes in vitro possess specific insulin receptors and demonstrate an effect of insulin on pyrimidine nucleoside uptake and incorporation.     

Endothelins Stimulate Expression of Cyclooxygenase 2 in Rat Cultured Astrocytes

Endothelin (ET) is one of the active endogenous substances regulating the functions of astrocytes. In the present study, we examined effects of ET on cyclooxygenase (COX) expression in cultured astrocytes. ET-3 (100 nM) caused transient increases in the expression of both COX2 mRNA and protein, but not those of COX1, in cultured astrocytes. ET-induced COX2 mRNA expression was suppressed by 5 microg/ml actinomycin D, 30 microM BAPTA/AM, inhibitors of protein kinase C (1-100 nM staurosporin and 100 microM H-7), 2 microM dexamethasone, and prolonged treatment with 100 nM phorbol 12-myristate 13-acetate. ET-3 stimulated production of prostaglandin (PG) E2 in cultured astrocytes. The effect of ET-3 on the PGE2 production was diminished by actinomycin D. Indomethacin and NS398, a selective COX2 inhibitor, comparably decreased both the basal and the ET-stimulated PGE2 production. Proliferation of cultured astrocytes was stimulated by 100 nM ET-3, and the increased proliferation was reduced by co-addition of 1 microM PGE2. Treatment with 1 microM PGE2 caused astrocytic morphological changes accompanied by disappearance of stress fibers, a prominent structure of organized cytoskeletal actin in cultured astrocytes. In the presence of 10 nM ET-3, PGE2 did not show an effect on astrocytic actin organization. The present study shows that ET is an inducer of astrocytic COX2 and suggests that ET-induced PGE2 production through COX2 may be involved in the regulation of astrocytic functions.     

Induction of Nitric Oxide Synthase Inhibits Gap Junction Permeability in Cultured Rat Astrocytes

Abstract: Nitric oxide (^?NO) synthase (NOS) was induced in cultured rat astrocytes by incubation with lipopolysaccharide (LPS) for 18 h and gap junction permeability was assessed by the scrape-loading/Lucifer yellow transfer technique. Induction of NOS was confirmed by determining either the N^G-methyl-l -arginine (NMMA)-inhibitable production of nitrites and nitrates or the conversion of l -[³H]arginine to l -[³H]citrulline. Incubation with LPS dose-dependently inhibited gap junction permeability to 63.3% at 0.05 µg/ml LPS and no further inhibition was observed on increasing the LPS concentration up to 0.5 µg/ml. LPS-mediated gap junction inhibition was irreversible but was prevented by incubation with the NOS inhibitor NMMA and with the superoxide anion (O₂^??) scavenger superoxide dismutase. Incubation of the cells with both the ^?NO donor S-nitroso-N-acetylpenicillamine and the O₂^??-generating system xanthine/xanthine oxidase inhibited gap junction permeability. These results suggest that the in situ reaction between ^?NO and O₂^??, to form the peroxynitrite anion (ONOO^?), may be responsible for the inhibition of gap junction permeability. Scavenging the ONOO^? derivative hydroxyl radical (^?OH) with either dimethyl sulfoxide or mannitol prevented the LPS-mediated inhibition of gap junction permeability. Finally, exposure of astrocytes to authentic ONOO^? caused a dose-dependent inhibition of gap junction permeability (65.7% of inhibition at 0.5 mM ONOO^?). The pathophysiological relevance of ONOO^?-mediated inhibition of gap junctional communication in astrocytes after NOS induction by LPS is discussed, stressing the possible role played by this mechanism in some neurodegenerative diseases.     

Biosynthesis of Prostaglandins D2 and E2 in Chick Dorsal Root Ganglion During Development

Newly formed prostaglandins (PGs), which are assumed to act as modulators of afferent sensory messages, were studied in chick dorsal root ganglia (DRG) during development. [1-14C]Arachidonic acid was converted by DRG homogenates from 1-week-old chickens into two major 14C-PGs: PGE2 and PGD2. The enzymatic conversion of arachidonic acid was characterized as follows: (a) Boiled preparations were inactivated; (b) synthesis of PGs was inhibited by pretreatment with aspirin or indomethacin and enhanced by esculetin, a protector of cyclooxygenase; and (c) [14C]PGE2 and [14C]PGD2 accumulation was a protein dose-dependent process. Further fractionation of crude homogenates indicated that PG endoperoxide synthetase (EC 1.14.99.1) and PGE2 synthetase (EC 5.3.99.3) were membrane-bound enzymes, whereas PGD2 synthetase (EC 5.3.99.2) was recovered in the cytosol. During development, from embryonic day 10 to day 14 after hatching, PGD2 synthetase activity remained constant; in contrast, a sharp rise in [14C]PGE2 synthesis was observed from embryonic day 14 to 18. The time curves of PGD2 and PGE2 synthetase specific activity may be related to changes taking place in the cell population of developing DRG. It is therefore suggested that arachidonic acid would be enzymatically converted early into PGD2 by maturing ganglion cells and then later into PGE2 by proliferating fibroblasts.     

High Sensitivity of Glutamate Uptake to Extracellular Free Arachidonic Acid Levels in Rat Cortical Synaptosomes and Astrocytes

By using both synaptosomes and cultured astrocytes from rat cerebral cortex, we have investigated the inhibitory action of arachidonic acid on the high-affinity glutamate uptake systems, focusing on the possible physiological significance of this mechanism. Application of arachidonic acid (1-100 microM) to either preparation leads to fast (within 30 s) and largely reversible reduction in the uptake rate. When either melittin (0.2-1 microgram/ml), a phospholipase A2 activator, or thimerosal (50-200 microM), which inhibits fatty acid reacylation in phospholipids, is applied to astrocytes, both an enhancement in extracellular free arachidonate and a reduction in glutamate uptake are seen. The two effects display similar dose dependency and time course. In particular, 10% uptake inhibition correlates with 30% elevation in free arachidonate, whereas inhibition greater than or equal to 60% is paralleled by threefold stimulation of arachidonate release. In the presence of albumin (1-10 mg/ml), a free fatty acid-binding protein, inhibition by either melittin, thimerosal, or arachidonic acid is prevented and an enhancement of glutamate uptake above the control levels is observed. Our data show that neuronal and glial glutamate transport systems are highly sensitive to changes in extracellular free arachidonate levels and suggest that uptake inhibition may be a relevant mechanism in the action of arachidonic acid at glutamatergic synapses.     

Participation of Prostaglandin E2 in Dopamine D2 Receptor-Dependent Potentiation of Arachidonic Acid Release

Stimulation of dopamine D2 receptors potentiates Ca2+ ionophore- or ATP-induced arachidonic acid (AA) release in D2 receptor cDNA-transfected Chinese hamster ovary (CHO) cells [CHO(D2)]. By using a combination of chromatographic, biochemical, and radioimmunochemical techniques, we show here that prostaglandin (PG) E2 is a major product of AA metabolism in CHO(D2) cells stimulated with the Ca2+ ionophore A23187. Formation of this PG was markedly increased by the concomitant application of quinpirole, a D2 receptor agonist. In addition, PGE2 enhanced D2-dependent amplification of AA release, either when it was added (EC50 = 100 nM) or when it was produced endogenously, as shown by experiments carried out with the cyclooxygenase inhibitor indomethacin. The results suggest that PGE2 may participate in D2 receptor-mediated potentiation of AA release in CHO(D2) cells. They also support a functional role for this PG in the modulation of dopaminergic transmission in areas of the CNS, such as amygdala and hypothalamus, where high levels of both PGE2 and dopamine D2 receptors are found.     

Manganese Uptake and Efflux in Cultured Rat Astrocytes

Astrocytes play a central role in manganese (Mn) regulation in the CNS. Using primary astrocyte cultures from neonatal rat brains, these studies demonstrate a specific high-affinity transport system for Mn2+. Saturation kinetics are clearly indicated by both 1/v versus 1/s plots (Km = 0.30 +/- 0.03 microM; Vmax = 0.30 +/- 0.02 nmol/mg of protein/min) and plots of v versus [s]. Several divalent cations (Co2+, Zn2+, and Pb2+) failed to inhibit the initial rate of 54Mn2+ uptake. In contrast, extracellular Ca2+ at 10 microM decreased 54Mn2+ uptake. Exchange with extracellular Mn2+ was not obligatory for the efflux of 54Mn2+ into extracellular medium because efflux occurred into Mn(2+)-free extracellular medium, but efflux of 54Mn2+ was enhanced when astrocytes were equilibrated in the presence of unlabeled Mn2+. Efflux of 54Mn2+ was biphasic with both a rapid and a slow component. Efflux was most rapid during the first 10 min of incubation, with 27.5 +/- 2.2% of 54Mn2+ transported extracellularly, and 37.2 +/- 1.2% of preloaded 54Mn2+ was retained by the astrocytes at 120 min. These studies show, for the first time, that mammalian astrocytes can transport Mn via a specific transport system.     

Interrelationships of Leucine and Glutamate Metabolism in Cultured Astrocytes

Abstract: The aim was to study the extent to which leu-cine furnishes α-NH₂ groups for glutamate synthesis via branched-chain amino acid aminotransferase. The transfer of N from leucine to glutamate was determined by incubating astrocytes in a medium containing [¹⁵N]leucine and 15 unlabeled amino acids; isotopic abundance was measured with gas chromatography-mass spectrometry. The ratio of labeling in both [¹⁵N]glutamate/[¹⁵N]leucine and [2-¹⁵N]glutamine/[¹⁵N]leucine suggested that at least one-fifth of all glutamate N had been derived from leucine nitrogen. At the same time, enrichment in [¹⁵N]leucine declined, reflecting dilution of the ¹⁶N label by the unlabeled amino acids that were in the medium. Isotopic abundance in [¹⁶N]-isoleucine increased very quickly, suggesting the rapidity of transamination between these amino acids. The appearance of ¹⁵N in valine was more gradual. Measurement of branched-chain amino acid transaminase showed that the reaction from leucine to glutamate was approximately six times more active than from glutamate to leucine (8.72 vs. 1.46 nmol/min/mg of protein). However, when the medium was supplemented with α-ketoisocaproate (1 mM), the ketoacid of leucine, the reaction readily ran in the “reverse” direction and intraastrocytic [glutamate] was reduced by ～50% in only 5 min. Extracellular concentrations of α-ketoisocaproate as low as 0.05 mM significantly lowered intracellular [glutamate]. The relative efficiency of branched-chain amino acid transamination was studied by incubating astrocytes with 15 unlabeled amino acids (0.1 mM each) and [¹⁵N]glutamate. After 45 min, the most highly labeled amino acid was [¹⁵N]alanine, which was closely followed by [¹⁵N]leucine and [¹⁵N]isoleucine. Relatively little ¹⁵N was detected in any other amino acids, except for [¹⁵N]serine. The transamination of leucine was ～17 times greater than the rate of [1-¹⁴C]leucine oxidation. These data indicate that leucine is a major source of glutamate nitrogen. Conversely, reamination of a-ketoisocaproate, the ketoacid of leucine, affords a mechanism for the temporary “buffering” of intracellular glutamate.     

High concentrations of arachidonic acid induce platelet aggregation and serotonin release independent of prostagladin endoperoxides and thromboxane A2

We examined platelet aggregation and serotonin release, induced by less than 60 μM arachidonic acid, using washed platelet suspensions in the absense of albumin. The concentration of arachidonic acid use did not cause platelet lysis. Platelet responses induced by less than 20 μM arachidonic acid were inhibited by aspirin, whereas those induced by above 30 μM arachidonic acid were not inhibited, even by both aspirin and 5,8,11,14-eicosatetraynoic acid. Although phosphatidic acid and 1,2-diacylglcerol increased after the addition of arachidonic acid in aspirin-treated platelets, the amounts were not parallel to platelet aggregation. Oleic, linoleic and linolenic acids also induced platelet responses, while palmitic, stearic and arachidic acids did not. EDTA, dibutyryl cyclic AMP, apyrase and creatine phosphate / creatin phosphokinase brought about almost the same effects in platelet responses induced by the unsaturated fatty acids, other than arachodinic acid, as those induced by 40 μM arachodonic acid. These results suggest that the mechanism of the actions of more than 30 μM arachodinic acid on platelets is the same as that of the other unsaturated fatty acids and is independent of prostaglandin endoperoxides, thromboxane A₂ and, perhaps, phosphatidic acid and 1,2-diacylglycerol.     

Glucocorticoids Modify Differentially Dopamine- and Prostaglandin E1-Mediated Cyclic AMP Formation by the Cultured Human Astrocytoma Clone D384

The effects of steroid hormones on the cyclic AMP responses to stimulation of human astrocytoma cells (D384) by dopamine, prostaglandin E1 (PGE1), and isoprenaline were investigated. Incubation of D384 cells with dexamethasone resulted in a potentiation of the PGE1 and isoprenaline responses and a marked attenuation of the dopamine response. The time courses of the effects of dexamethasone on dopamine and PGE1 responses were similar, requiring long-term (at least 18 h) incubation of cells with the steroid. Concentration-response curves of dexamethasone effects on dopamine and PGE1 responses yielded similar Ka apparent values, suggesting a common mechanism. Cycloheximide, a protein synthesis inhibitor, prevented the effects of dexamethasone. Only steroids with glucocorticoid activity reproduced the dexamethasone effects. Direct stimulation of Gs with 5-guanylylimidodiphosphate and adenylate cyclase with forskolin revealed no significant differences in their activities in dexamethasone-treated and untreated cells. Furthermore, a comparison of the dopamine and PGE1 concentration-response curves obtained from dexamethasone-treated and untreated cells suggested that the affinity of the receptors for their agonists remained unchanged. These results suggest that glucocorticoids may alter protein synthesis and thereby the number of receptors expressed by D384 cells.     

Ascorbic Acid Uptake by a High-Affinity Sodium-Dependent Mechanism in Cultured Rat Astrocytes

Ascorbic acid (vitamin C) is synthesized in rodent liver, circulates in the blood, and is concentrated in the brain. Experiments were performed to characterize the mechanism of ascorbate uptake by rat cerebral astrocytes in primary culture. Astroglial uptake of L-[14C]ascorbate was observed to be both saturable and stereoselective. In addition, uptake was dependent on both the incubation temperature and the concentration of Na+ because it was largely inhibited by cooling to 4 degrees C, by treatment with ouabain to increase intracellular Na+, and by the substitution of K+, Li+, or N-methyl-D-glucamine for extracellular Na+. The affinity for ascorbate was relatively high in cells incubated with a physiological concentration of extracellular Na+, because the apparent Km was 32 microM in 138 mM Na+. However, the affinity for ascorbate was significantly decreased when the extracellular Na+ concentration was lowered. Treatment of astrocytes with dibutyryl cyclic AMP induced stellation and increased the maximum rate of ascorbate uptake by 53%. We conclude that astrocytes possess a stereoselective, high-affinity, and Na+-dependent uptake system for ascorbate. This system may regulate the cerebral ascorbate concentration and consequently modulate neuronal function.     

Glucose Transport in Astrocytes: Regulation by Thyroid Hormone

Primary cultures of astrocytes from newborn rat brain showed evidence of a substrate-saturable process for glucose transport. The system shows a relatively high affinity for the substrate, with an apparent Km of approximately 1 mM. Maintenance of the cells in medium containing thyroid-hormone-free serum for 3, 6, or 9 days resulted in significantly reduced rates of hexose transport. Addition of exogenous triiodothyronine to the transport incubation medium of these "hypothyroid" cells markedly increased the net rate of 2-deoxyglucose uptake within 60 s to values equal to or above those of control cultures (cells maintained in normal serum). These findings support a key role for thyroid hormone in the transport of glucose across plasma membranes of brain cells and demonstrate the presence of this regulatory system in astrocytes.     

Intracellular Ascorbic Acid Inhibits the Na+-Ca2+ Exchanger in Cultured Rat Astrocytes

Abstract: The effect of ascorbic acid on Ca²⁺ uptake in cultured rat astrocytes was examined in the presence of ouabain and monensin, which are considered to drive the Na⁺-Ca²⁺ exchanger in the reverse mode. Ascorbic acid at 0.1–1 m M inhibited Na⁺-dependent Ca²⁺ uptake significantly but not Na⁺-dependent glutamate uptake in the cells, although the inhibition required pretreatment for more than 30 min. The effect of ascorbic acid on the Ca²⁺ uptake was blocked by simultaneous addition of ascorbate oxidase (10 U/ml). Na⁺-dependent Ca²⁺ uptake was also inhibited by isoascorbate at 1 m M but not by ascorbate 2-sulfate, dehydroascorbate, and sulfhydryl-reducing reagents such as glutathione and 2-mercaptoethanol. The inhibitory effect of ascorbic acid was observed even in the presence of an inhibitor of lipid peroxidation, o -phenanthroline, or a radical scavenger, mannitol, and the degrading enzymes such as catalase and superoxide dismutase. On the other hand, the inhibitory effect was not observed under the Na⁺-free conditions that inhibited the uptake of ascorbic acid in astrocytes. When astrocytes were cultured for 2 weeks in a medium containing ascorbic acid, the content of ascorbic acid in the cells was increased and conversely Na⁺-dependent Ca²⁺ uptake was decreased. These results suggest that an increase in intracellular ascorbic acid results in a decrease of Na⁺-Ca²⁺ exchange activity in cultured astrocytes and the mechanism is not related to lipid peroxidation.