IL-1β stimulates COX-2 dependent PGE₂ synthesis and CGRP release in rat trigeminal ganglia cells

Objective

Pro-inflammatory cytokines like Interleukin-1 beta (IL-1β) have been implicated in the pathophysiology of migraine and inflammatory pain. The trigeminal ganglion and calcitonin gene-related peptide (CGRP) are crucial components in the pathophysiology of primary headaches. 5-HT1B/D receptor agonists, which reduce CGRP release, and cyclooxygenase (COX) inhibitors can abort trigeminally mediated pain. However, the cellular source of COX and the interplay between COX and CGRP within the trigeminal ganglion have not been clearly identified.

Methods and Results

1. We used primary cultured rat trigeminal ganglia cells to assess whether IL-1β can induce the expression of COX-2 and which cells express COX-2. Stimulation with IL-1β caused a dose and time dependent induction of COX-2 but not COX-1 mRNA. Immunohistochemistry revealed expression of COX-2 protein in neuronal and glial cells. 2. Functional significance was demonstrated by prostaglandin E2 (PGE₂) release 4 hours after stimulation with IL-1β, which could be aborted by a selective COX-2 (parecoxib) and a non-selective COX-inhibitor (indomethacin). 3. Induction of CGRP release, indicating functional neuronal activation, was seen 1 hour after PGE₂ and 24 hours after IL-1β stimulation. Immunohistochemistry showed trigeminal neurons as the source of CGRP. IL-1β induced CGRP release was blocked by parecoxib and indomethacin, but the 5-HT1B/D receptor agonist sumatriptan had no effect.

Conclusion

We identified a COX-2 dependent pathway of cytokine induced CGRP release in trigeminal ganglia neurons that is not affected by 5-HT1B/D receptor activation. Activation of neuronal and glial cells in the trigeminal ganglion by IL-β leads to an elevated expression of COX-2 in these cells. Newly synthesized PGE₂ (by COX-2) in turn activates trigeminal neurons to release CGRP. These findings support a glia-neuron interaction in the trigeminal ganglion and demonstrate a sequential link between COX-2 and CGRP. The results could help to explain the mechanism of action of COX-2 inhibitors in migraine.     

Interaction between the serotoninergic and dopaminergic systems in d-fenfluramine-induced activation of c-fos and jun B genes in rat striatal neurons

To test for the relative contributions of the dopaminergic and serotoninergic systems in the striatum to the effects of d-fenfluramine, an indirect serotonin receptor agonist, we assessed the expression of Fos/Jun proteins induced by d-fenfluramine given alone or in the presence of dopaminergic or serotoninergic agents. To determine the neuronal targets of d-fenfluramine in the striatum, we identified the phenotypes of striatal neurons in which d-fenfluramine induced Fos expression. Our results demonstrated that d-fenfluramine evokes nuclear expression of Fos/Jun B proteins in the striatum, and that the Fos expression was dose-dependent and accompanied by transient induction of c-fos mRNA. Fos expression was blocked by p-chloroamphetamine, a serotoninergic neurotoxin. Pretreatment with SCH 23390, a D1-dopamine receptor antagonist, led to a marked decrease in Fos/Jun B expression in the caudoputamen, but not in the cortex, whereas pretreatment with methiothepin, a nonselective serotonin 5-HT1 receptor antagonist, blocked Fos expression completely in the cortex and only partially in the caudoputamen. The expression of Fos/Jun B in the striatum occurred mainly in dynorphin-containing neurons and in a subpopulation of striatal interneurons that exhibited NADPH-diaphorase activity. Most of the enkephalin-containing neurons of the striatum did not show Fos/Jun B staining. These results suggest that the mechanism by which d-fenfluramine induces c-fos and jun B expression in the rat caudoputamen depends at least in part on activation of the dopaminergic system by serotonin.     

Serotonin 5-HT2C receptor-mediated inhibition of the M-current in hypothalamic POMC neurons

Hypothalamic proopiomelanocortin (POMC) neurons are controlled by many central signals, including serotonin. Serotonin increases POMC activity and reduces feeding behavior via serotonion [5-hydroxytryptamine (5-HT)] receptors by modulating K(+) currents. A potential K(+) current is the M-current, a noninactivating, subthreshold outward K(+) current. Previously, we found that M-current activity was highly reduced in fasted vs. fed states in neuropeptide Y neurons. Because POMC neurons also respond to energy states, we hypothesized that fasting may alter the M-current and/or its modulation by serotonergic input to POMC neurons. Using visualized-patch recording in neurons from fed male enhanced green fluorescent protein-POMC transgenic mice, we established that POMC neurons expressed a robust M-current (102.1 ± 6.7 pA) that was antagonized by the selective KCNQ channel blocker XE-991 (40 μM). However, the XE-991-sensitive current in POMC neurons did not differ between fed and fasted states. To determine if serotonin suppresses the M-current via the 5-HT(2C) receptor, we examined the effects of the 5-HT(2A)/5-HT(2C) receptor agonist 2,5-dimethoxy-4-iodoamphetamine (DOI) on the M-current. Indeed, DOI attenuated the M-current by 34.5 ± 6.9% and 42.0 ± 5.3% in POMC neurons from fed and fasted male mice, respectively. In addition, the 5-HT(1B)/5-HT(2C) receptor agonist m-chlorophenylpiperazine attenuated the M-current by 42.4 ± 5.4% in POMC neurons from fed male mice. Moreover, the selective 5-HT(2C) receptor antagonist RS-102221 abrogated the actions of DOI in suppressing the M-current. Collectively, these data suggest that although M-current expression does not differ between fed and fasted states in POMC neurons, serotonin inhibits the M-current via activation of 5-HT(2C) receptors to increase POMC neuronal excitability and, subsequently, reduce food intake.     

5-Hydroxytryptophan activates colonic myenteric neurons and propulsive motor function through 5-HT4 receptors in conscious mice

Serotonin [5-hydroxytryptamine (5-HT)] acts as a modulator of colonic motility and secretion. We characterized the action of the 5-HT precursor 5-hydroxytryptophan (5-HTP) on colonic myenteric neurons and propulsive motor activity in conscious mice. Fos immunoreactivity (IR), used as a marker of neuronal activation, was monitored in longitudinal muscle/myenteric plexus whole mount preparations of the distal colon 90 min after an intraperitoneal injection of 5-HTP. Double staining of Fos IR with peripheral choline acetyltransferase (pChAT) IR or NADPH-diaphorase activity was performed. The injection of 5-HTP (0.5, 1, 5, or 10 mg/kg ip) increased fecal pellet output and fluid content in a dose-related manner, with a peak response observed within the first 15 min postinjection. 5-HTP (0.5-10 mg/kg) dose dependently increased Fos expression in myenteric neurons, with a maximal response of 9.9 +/- 1.0 cells/ganglion [P < 0.05 vs. vehicle-treated mice (2.3 +/- 0.6 cells/ganglion)]. There was a positive correlation between Fos expression and fecal output. Of Fos-positive ganglionic cells, 40 +/- 4% were also pChAT positive and 21 +/- 5% were NADPH-diaphorase positive in response to 5-HTP, respectively. 5-HTP-induced defecation and Fos expression were completely prevented by pretreatment with the selective 5-HT4 antagonist RS-39604. These results show that 5-HTP injected peripherally increases Fos expression in different populations of cholinergic and nitrergic myenteric neurons in the distal colon and stimulates propulsive colonic motor function through 5-HT4 receptors in conscious mice. These findings suggest an important role of activation of colonic myenteric neurons in the 5-HT4 receptor-mediated colonic propulsive motor response.     

Group IB secretory phospholipase A2 induces neuronal cell death via apoptosis

Group IB secretory phospholipase A2 (sPLA2-IB) mediates cell proliferation, cell migration, hormone release and eicosanoid production via its receptor in peripheral tissues. In the CNS, high-affinity binding sites of sPLA2-IB have been documented. However, it remains obscure whether sPLA2-IB causes biologic or pathologic response in the CNS. To this end, we examined effects of sPLA2-IB on neuronal survival in primary cultures of rat cortical neurons. sPLA2-IB induced neuronal cell death in a concentration-dependent manner. This death was a delayed response requiring a latent time for 6 h; sPLA2-IB-induced neuronal cell death was accompanied with apoptotic blebbing, condensed chromatin, and fragmented DNA, exhibiting apoptotic features. Before cell death, sPLA2-IB liberated arachidonic acid (AA) and generated prostaglandin D2 (PGD2) from neurons. PGD2 and its metabolite, Delta12-PGJ2, exhibited neurotoxicity. Inhibitors of sPLA2 and cyclooxygenase-2 (COX-2) significantly suppressed not only AA release, but also PGD2 generation. These inhibitors significantly prevented neurons from sPLA2-IB-induced neuronal cell death. In conclusion, we demonstrate a novel biological response, apoptosis, of sPLA2-IB in the CNS. Furthermore, the present study suggests that PGD2 metabolites, especially Delta12-PGJ2, might mediate sPLA2-IB-induced apoptosis.     

An ELISA method to measure inhibition of the COX enzymes

The cyclooxygenase (COX) reaction can be monitored by measurement of oxygen consumption, peroxidase co-substrate oxidation or prostaglandin (PG) detection. This protocol describes a procedure measuring cyclooxygenase activity by quantifying PGE2 produced by enzymatic conversion of arachidonic acid, in the presence or absence of potential inhibitors. This high-throughput method has the advantage that it directly measures cyclooxygenase activity and requires little enzyme. The first part of the assay consists of incubating arachidonic acid, cyclooxygenase and the test samples to generate prostaglandins. The second part uses an ELISA method to quantify the amount of PGE2 produced by the enzymatic reaction. The isolation of COX-1 and COX-2 enzymes is also described. This protocol can be completed in approximately 23 h, including 16-h and 4-h incubation phases. This does not include enzyme preparation (3 h for COX-1 and 24 h for COX-2) or preparation of ELISA plates (23 h, including incubation).     

Arachidonic acid can function as a signaling modulator by activating the TRPM5 cation channel in taste receptor cells

Vertebrate sensory cells such as vomeronasal neurons and Drosophila photoreceptor cells use TRP channels to respond to exogenous stimuli. In mammalian taste cells, bitter and sweet substances as well as some amino acids are received by G protein-coupled receptors (T2Rs or T1Rs). As a result of activation of G protein and phospholipase Cbeta2, the TRPM5 channel is activated. Intracellular Ca(2+) is known to be a TRPM5 activator, but the participation of lipid activators remains unreported. To clarify the effect of arachidonic acid on TRPM5 in taste cells, we investigated the expression profile of a series of enzymes involved in controlling the intracellular free arachidonic acid level, with the result that in a subset of taste bud cells, monoglyceride lipase (MGL) and cyclooxygenase-2 (COX-2) are expressed as well as the previously reported group IIA phospholipase A(2) (PLA(2)-IIA). Double-labeling analysis revealed that MGL, COX-2 and PLA(2)-IIA are co-expressed in some cells that express TRPM5. We then investigated whether arachidonic acid activates TRPM5 via a heterologous expression system in HEK293 cells, and found that its activation occurred at 10 microM arachidonic acid. These results strongly suggest the possibility that arachidonic acid acts as a modulator of TRPM5 in taste signaling pathways.     

The influence of repeated administration of clozapine and haloperidol on the effects of the activation of 5-HT(1A), 5-HT(2) and 5-HT(4) receptors in rat frontal cortex.

The effects of a repeated treatment with antipsychotic drugs, clozapine and haloperidol, on the modulation of network activity ex vivo by 5-HT receptors were examined in rat frontal cortical slices using extracellular recording. Rats were treated for 21 days with clozapine (30 mg/kg p.o.), or haloperidol (1 mg/kg p.o.). Spontaneous bursting activity was induced in slices prepared 3 days after the last drug administration by perfusion with a medium devoid of Mg(2+) ions and with added picrotoxin (30 mM). The application of 2-3 microM 8-OH-DPAT, acting through 5-HT(1A) receptors, resulted in a reversible decrease of bursting frequency. In the presence of 1 microM DOI, the 5-HT(2) agonist, or 5 microM zacopride, the 5-HT(4) agonist, bursting frequency increased. Chronic clozapine treatment resulted in an attenuation of the effect of the activation of 5-HT(2) receptors, while the effects related to 5-HT(1A) and 5-HT(4) receptor activation were unchanged. Treatment with haloperiol did not influence the reactivity to the activation of any of the three 5-HT receptor subtypes. These data are consistent with earlier findings demonstrating a selective downregulation of 5-HT(2A) receptors by clozapine and indicate that chronic clozapine selectively attenuates the 5-HT-mediated excitation in neuronal circuitry of the frontal cortex while leaving the 5-HT-mediated inhibition intact.     

Regulation of prostaglandin synthesis by thyrotropin, insulin or insulin-like growth factor-I, and serum in FRTL-5 rat thyroid cells.

The present report shows that thyrotropin (TSH) regulates all three steps involved in prostaglandin synthesis in FRTL-5 rat thyroid cells, i.e. arachidonic acid release from membrane phospholipids, cyclooxygenase (prostaglandin H synthase) action, and individual prostaglandin formation; however, its action at specific steps may require the presence of, or can be duplicated by, insulin, insulin-like growth factor-I (IGF-I), and/or a serum factor. Thus, TSH releases free arachidonic acid from rat FRTL-5 thyroid cells whose phospholipid fraction is radiolabeled with [3H]arachidonic acid; this action involves a pertussis toxin-sensitive G protein, is not cAMP mediated, and does not require insulin or 5% serum. To quantitate TSH effects on cyclooxygenase activity and on individual prostaglandin formation, a homogenate system and a rapid reversed-phase high pressure liquid chromatography procedure have been developed to measure cyclooxygenase metabolites. TSH increased cyclooxygenase activity in homogenates only if the cells were also exposed to insulin, IGF-I, and/or 5% calf serum; TSH alone had no apparent effect on the activity. Maximal activation, 4-fold over basal/micrograms of DNA, took 36 h to achieve and reflected, at least in part, an increase in cyclooxygenase gene expression. Like cyclooxygenase activity, induction of prostaglandin E2 production required 2 or more factors, i.e. TSH plus insulin/IGF-I or TSH plus insulin/IGF-I plus serum. Increased production of prostaglandin D2, could, however, be detected if cells were treated with TSH alone and the TSH activity could be duplicated by insulin, IGF-I, or calf serum alone.     

The stimulation of arachidonic acid metabolism in human platelets by hydrodynamic stresses

Even though shear-induced platelet activation and aggregation have been studied for about 20 years, there remains some controversy concerning the arachidonic acid metabolites formed during stress activation and the role of thromboxane A2 in shear-induced platelet aggregation. In this study, platelets were labelled with [1-14C]arachidonic acid to follow the metabolism of arachidonic acid in stimulated platelets using HPLC and scintillation counting. Platelets activated by thrombin formed principally thromboxane A2, 12-hydroxy-5,8,10-heptadecatrienoic acid (HHT) and 12-hydroxy-5,8,10,14-eicosatetraenoic acid (12-HETE). In contrast, for platelets activated by shear--though arachidonic acid metabolism was stimulated--only 12-HETE was formed and essentially no cyclooxygenase metabolites were detected. This indicates that physical forces may initiate a different pathway for eicosanoid metabolism than most commonly used chemical stimuli and perhaps also implies that regulation of the cyclooxygenase activity may be a secondary level of regulation in eicosanoid metabolism.     

Amino acid determinants in cyclooxygenase-2 oxygenation of the endocannabinoid 2-arachidonylglycerol

The endocannabinoid, 2-arachidonylglycerol (2-AG), is an endogenous ligand for the central (CB1) and peripheral (CB2) cannabinoid receptors and has been shown to be efficiently and selectively oxygenated by cyclooxygenase (COX)-2. We have investigated 2-AG/COX-2 interactions through site-directed mutagenesis. An evaluation of more than 20 site-directed mutants of murine COX-2 has allowed for the development of a model of 2-AG binding within the COX-2 active site. Most strikingly, these studies have identified Arg-513 as a critical determinant in the ability of COX-2 to efficiently generate prostaglandin H(2) glycerol ester, explaining, in part, the observed isoform selectivity for this substrate. Mutational analysis of Leu-531, an amino acid located directly across from Arg-513 in the COX-2 active site, suggests that 2-AG is shifted in the active site away from this hydrophobic residue and toward Arg-513 relative to arachidonic acid. Despite this difference, aspirin-treated COX-2 oxygenates 2-AG to afford 15-hydroxyeicosatetraenoic acid glycerol ester in a reaction analogous to the C-15 oxygenation of arachidonic acid observed with acetylated COX-2. Finally, the differences in substrate binding do not alter the stereospecificity of the cyclooxygenase reaction; 2-AG-derived and arachidonic acid-derived products share identical stereochemistry.     

Chronic Administration of Lamotrigine Downregulates COX-2 mRNA and Protein in Rat Frontal Cortex

Chronic administration to rats of mood-stabilizers that are effective against mania in bipolar disorder, is reported to downregulate markers of the brain arachidonic acid cascade. We hypothesized that chronic administration of lamotrigine, which is used to treat depression and rapid cycling in bipolar disorder, might do so as well. Male CDF rats were administered a therapeutically relevant dose of lamotrigine (10 mg/kg) or vehicle intragastrically once daily for 42 days. Protein levels of isoforms of phospholipase A₂ (PLA₂) and of cyclooxygenase (COX), and the mRNA level of COX-2, were quantified in the frontal cortex using immunoblotting and RT-PCR, respectively. Compared to vehicle-treated rats, chronic lamotrigine significantly decreased frontal cortex protein and mRNA levels of COX-2 without altering protein levels of the PLA₂ isoforms. Consistent with the hypothesis, lamotrigine and other mood-stabilizers have a common downregulatory action on COX-2 expression in rat brain, which may account in part for their efficacy in bipolar disorder.     

Cortical and septal responses to dorsal raphe nucleus stimulation in the rat: long-term clomipramine actions.

In cerebral cortex and lateral septal nuclei different serotonergic receptor subtypes coexist, thus a different action on neuronal firing may be expected depending on the receptor activated. Dorsal raphe nucleus stimulation produced an increased rate of firing in cortical layer V, and in lateral septal nuclei. However, firing rate in cortical layer VI remained unchanged after stimulating the dorsal raphe nucleus. Clomipramine is a tricyclic which exerts its main actions on serotonergic receptors, and long-term treatment with this antidepressant produced a selective increased firing rate in lateral septal neurons, but not in cortical neurons. From an electrophysiological point of view, it is concluded that the excitatory actions on firing rate elicited by dorsal raphe nucleus stimulation or clomipramine treatment are mediated by 5-HT2 receptor subtype activation which is likely to be acting as a 5-HT1A modulator in such places where both receptor subtypes coexist.     

Regulation of prostaglandin E2 biosynthesis by inducible membrane-associated prostaglandin E2 synthase that acts in concert with cyclooxygenase-2

Here we report the molecular identification of membrane-bound glutathione (GSH)-dependent prostaglandin (PG) E(2) synthase (mPGES), a terminal enzyme of the cyclooxygenase (COX)-2-mediated PGE(2) biosynthetic pathway. The activity of mPGES was increased markedly in macrophages and osteoblasts following proinflammatory stimuli. cDNA for mouse and rat mPGESs encoded functional proteins that showed high homology with the human ortholog (microsomal glutathione S-transferase-like 1). mPGES expression was markedly induced by proinflammatory stimuli in various tissues and cells and was down-regulated by dexamethasone, accompanied by changes in COX-2 expression and delayed PGE(2) generation. Arg(110), a residue well conserved in the microsomal GSH S-transferase family, was essential for catalytic function. mPGES was functionally coupled with COX-2 in marked preference to COX-1, particularly when the supply of arachidonic acid was limited. Increased supply of arachidonic acid by explosive activation of cytosolic phospholipase A(2) allowed mPGES to be coupled with COX-1. mPGES colocalized with both COX isozymes in the perinuclear envelope. Moreover, cells stably cotransfected with COX-2 and mPGES grew faster, were highly aggregated, and exhibited aberrant morphology. Thus, COX-2 and mPGES are essential components for delayed PGE(2) biosynthesis, which may be linked to inflammation, fever, osteogenesis, and even cancer.     

Cyclooxygenase-1 and -2 enzymes differentially regulate the brain upstream NF-kappa B pathway and downstream enzymes involved in prostaglandin biosynthesis

We have recently reported that cyclooxygenase (COX)-2-deficiency affects brain upstream and downstream enzymes in the arachidonic acid (AA) metabolic pathway to prostaglandin E2 (PGE2), as well as enzyme activity, protein and mRNA levels of the reciprocal isozyme, COX-1. To gain a better insight into the specific roles of COX isoforms and characterize the interactions between upstream and downstream enzymes in brain AA cascade, we examined the expression and activity of COX-2 and phospholipase A2 enzymes (cPLA2 and sPLA2), as well as the expression of terminal prostaglandin E synthases (cPGES, mPGES-1, and - 2) in wild type and COX-1(-/-) mice. We found that brain PGE2 concentration was significantly increased, whereas thromboxane B2 (TXB2) concentration was decreased in COX-1(-/-) mice. There was a compensatory up-regulation of COX-2, accompanied by the activation of the NF-kappaB pathway, and also an increase in the upstream cPLA2 and sPLA2 enzymes. The mechanism of NF-kappaB activation in the COX-1(-/-) mice involved the up-regulation of protein expression of the p50 and p65 subunits of NF-kappaB, as well as the increased protein levels of phosphorylated IkappaBalpha and of phosphorylated IKKalpha/beta. Overall, our data suggest that COX-1 and COX-2 play a distinct role in brain PG biosynthesis, with basal PGE2 production being metabolically coupled with COX-2 and TXB2 production being preferentially linked to COX-1. Additionally, COX-1 deficiency can affect the expression of reciprocal and coupled enzymes, COX-2, Ca2+ -dependent PLA2, and terminal mPGES-2, to overcome defects in brain AA cascade.     

The blockade of the transient receptor potential vanilloid type 1 and fatty acid amide hydrolase decreases symptoms and central sequelae in the medial prefrontal cortex of neuropathic rats

Background

Neuropathic pain is a chronic disease resulting from dysfunction within the "pain matrix". The basolateral amygdala (BLA) can modulate cortical functions and interactions between this structure and the medial prefrontal cortex (mPFC) are important for integrating emotionally salient information. In this study, we have investigated the involvement of the transient receptor potential vanilloid type 1 (TRPV1) and the catabolic enzyme fatty acid amide hydrolase (FAAH) in the morphofunctional changes occurring in the pre-limbic/infra-limbic (PL/IL) cortex in neuropathic rats.

Results

The effect of N-arachidonoyl-serotonin (AA-5-HT), a hybrid FAAH inhibitor and TPRV1 channel antagonist, was tested on nociceptive behaviour associated with neuropathic pain as well as on some phenotypic changes occurring on PL/IL cortex pyramidal neurons. Those neurons were identified as belonging to the BLA-mPFC pathway by electrical stimulation of the BLA followed by hind-paw pressoceptive stimulus application. Changes in their spontaneous and evoked activity were studied in sham or spared nerve injury (SNI) rats before or after repeated treatment with AA-5-HT. Consistently with the SNI-induced changes in PL/IL cortex neurons which underwent profound phenotypic reorganization, suggesting a profound imbalance between excitatory and inhibitory responses in the mPFC neurons, we found an increase in extracellular glutamate levels, as well as the up-regulation of FAAH and TRPV1 in the PL/IL cortex of SNI rats. Daily treatment with AA-5-HT restored cortical neuronal activity, normalizing the electrophysiological changes associated with the peripheral injury of the sciatic nerve. Finally, a single acute intra-PL/IL cortex microinjection of AA-5-HT transiently decreased allodynia more effectively than URB597 or I-RTX, a selective FAAH inhibitor or a TRPV1 blocker, respectively.

Conclusion

These data suggest a possible involvement of endovanilloids in the cortical plastic changes associated with peripheral nerve injury and indicate that therapies able to normalize endovanilloid transmission may prove useful in ameliorating the symptoms and central sequelae associated with neuropathic pain.     

Cyclooxygenase-2-dependent neuronal death proceeds via superoxide anion generation

Cyclooxygenase-2 (COX-2) expression is induced in the neurons of the pathologic brain and elevated COX-2 expressions can lead to neuronal death. Here, we report that COX-2 induction in cortical neurons induced by LPS pretreatment for more than 12 h increased the neurotoxic effects of low doses of Fe2+ by more than 2.5-fold. Moreover, the neurotoxicity induced by 30 muM Fe2+ in LPS-pretreated cells exceeded that induced by 100 microM Fe2+ in LPS-untreated cells. LPS pretreatment also similarly aggravated the neurotoxic effects of low doses of H2O2, Zn2+, and sodium nitroprusside. This LPS-induced Fe2+ -toxicity enhancement was blocked by trolox, vitamin C, the SOD mimetic MnTBAP, and by the COX-2-specific inhibitor NS398, but not by inhibitors of xanthine oxidase, NADPH oxidase, NOS, and monoamine oxidase. Cortical neurons with enhanced COX-2 expression showed superoxide generation, GSH depletion, and lipid peroxidation in response to low doses of Fe2+, and all of these changes were repressed by MnTBAP or NS398. Consistent with this pharmacological data, cortical neurons prepared from COX-2 knockout mice showed marked reductions in LPS-induced Fe2+ -toxicity enhancement and superoxide generation. These results suggest that COX-2 functions as a cellular factor which induces superoxide-mediated cell death in primary cortical neurons.