Kainate receptor agonists, antagonists and allosteric modulators.

Interest in kainate receptors has increased over the past few years. Our understanding of their physiology and pharmacology has improved markedly since their original cloning and expression in the early 1990s. For example, agonist profiles at recombinant kainate receptors have been used to identify and distinguish kainate receptors in neurons. Furthermore, the development of selective antagonists for kainate receptor subtypes has increased our understanding of the functional roles of kainate receptors in neurons and synaptic transmission. In this review we described the activity of agonists and antagonists at kainate receptors and their selectivity profiles at NMDA and non-NMDA receptors.

LY339434, a GluR5 kainate receptor agonist.

The activity of a gamma-substituted glutamate analogue, (2S, 4R, 6E)-2-amino-4-carboxy-7-(2-naphthyl)hept-6-enoic acid (LY339434) and (2S,4R)-4-methylglutamic acid at ionotropic glutamate receptors has been examined. Ligand binding studies were performed using [3H] AMPA binding to membranes expressing either homomeric recombinant GluR1, GluR2, GluR4 receptors, and [3H] kainate binding to GluR5 and GluR6 kainate receptors. LY339434 and (2S,4R)-4-methylglutamic acid showed selectivity in ligand binding studies for kainate receptors over AMPA receptors. Within the kainate class of glutamate receptors, LY339434 showed selectivity for GluR5 over GluR6 whereas (2S,4R)-4-methylglutamic acid showed high affinity for both GluR5 and GluR6 kainate receptors. Examination of the functional activity of LY339434 and (2S,4R)-4-methylglutamic acid showed that both compounds evoked inward currents in dorsal root ganglion neurons (DRG) with estimated EC50 values of 0.8 +/- 0.2 microM and 0.17 +/- 0.04 microM, respectively. In GluR5 expressing HEK 293 cells, LY339434 evoked inward currents with an estimated EC50 value of 2.5 +/- 0.9 microM but had little effect on GluR6 expressing cells at concentrations less than 100 microM. LY339434 was a weak AMPA receptor agonist (EC50 values > 300 microM) as determined by activity in acutely isolated cerebellar Purkinje neurons. LY339434 and (2S,4R)-4-methylglutamic acid had agonist activity at NMDA receptors studied in cultured hippocampal neurons with EC50s of 2.5 microM and 11.7 microM, respectively. These results indicate that both LY339434 and (2S,4R)-4-methyl glutamic acid may be useful pharmacological tools for the examination of kainate receptors.

Pharmacological characterization of glutamatergic agonists and antagonists at recombinant human homomeric and heteromeric kainate receptors in vitro.

An increasing body of evidence suggests that native kainate receptors form ion channels from homomeric and heteromeric combinations of five receptor subunits: GluR5, GluR6, GluR7, KA1 and KA2. We have examined the activity of agonists and antagonists at recombinant human kainate receptors expressed in HEK293 cells, using both whole-cell electrophysiological recording and 96-well plate fluo-3 based calcium microfluorimetry (FLIPR). Both homomeric (GluR5 and GluR6) and heteromeric (GluR5/6, GluR5/KA2 and GluR6/KA2) receptors were examined. Heteromeric receptor assemblies showed electrophysiological and pharmacological profiles which were distinct from homomeric channels. Several agonists, including AMPA, ATPA and (S)-5-iodowillardiine, and antagonists, including gamma-D-glutamylaminomethylsulphonic acid (GAMS) and the decahydroisoquinoline compounds LY293558, LY377770 and LY382884, were found to act at GluR5-containing channels while having no effect at GluR6 homomers. AMPA, ATPA and (S)-5-iodowillardiine did activate GluR6/KA2 heteromers, but only as partial agonists. Additionally, ATPA was shown to act as an antagonist at homomeric GluR6 receptors at high concentrations (IC50 approximately 2 mM). Kynurenic acid was also found to differentiate between GluR6 and GluR6/KA2 receptors, antagonizing glutamate at GluR6 (IC50 = 0.4 mM), while having no effect at GluR6/KA2 channels. The results of the current study provide a broad pharmacological characterization of both homomeric and heteromeric recombinant human kainate receptors, and identify which compounds are likely to be useful tools for studying these various receptor subtypes.

Conformationally restrained, chiral (phenylisopropyl)amino-substituted pyrazolo[3,4-d]pyrimidines and purines with selectivity for adenosine A1 and A2 receptors.

Two modes of tethering a chiral (phenylisopropyl)amino substituent in pyrazolo[3,4-d]pyrimidines and purines have been explored. One mode gave (S)-2,7-dihydro-7-phenyl-2-(phenylmethyl)-5- propoxy-3H-imidazo[1,2-c]pyrazolo-[4,3-e]pyrimidine (12a) and its corresponding R-enantiomer 12b, which were selective for A2 and A1 adenosine receptors, respectively. The corresponding diimidazo[1,2-c:4',5'-e]pyrimidines 12e and 12f were analogously selective. This is the first example where a single chiral recognition unit provides enantiomers with opposite selectivities for adenosine receptors. The second mode gave (2S-trans)-2,7-dihydro-2-methyl-3,7-diphenyl-5- propoxy-3H-imidazo[1,2-c]-pyrazolo[4,3-e]pyrimidine (12c) and its corresponding R-enantiomer 12d. Compounds 12c and 12d were significantly less potent than 12a and 12b at A1 receptors, and were nonselective.

Xanthines with C8 chiral substituents as potent and selective adenosine A1 antagonists.

Several 8-substituted 1,3-dipropylxanthines were synthesized, and their receptor binding affinities at adenosine A1 and A2 receptors were measured. When enantiomeric pairs of compounds were examined, the R enantiomers were significantly more potent than the corresponding S enantiomers. The most potent compound at the A1 receptor was (R)-3,7-dihydro-8-(1-methyl-2-phenylethyl)-1,3-dipropyl-1H-purine-2,6-di one (5a; MDL 102,503), whose Ki value at the A1 receptor was 6.9 nM. However, a more selective compound was (R)-3,7-dihydro-8-(1-phenylpropyl)-1,3-dipropyl-1H-purine-2,6-dione (5d; MDL 102,234), which had a Ki value of 23.2 nM at the A1 receptor and an A2/A1 ratio of 153.

5-Aryl-3-(alkylthio)-4H-1,2,4-triazoles as selective antagonists of strychnine-induced convulsions and potential antispastic agents.

Selected examples from three series of isomeric (alkylthio)-1,2,4-triazoles were prepared and examined for anticonvulsant activity versus strychnine-, maximal-electroshock-, pentylenetetrazole-, and 3-mercaptopropionic-acid-induced seizures in mice. A number of 5-aryl-3-(alkylthio)-4H-1,2,4-triazoles were selective antagonists of strychnine-induced convulsions. The isomeric 3-aryl-5-(alkylthio)- and 5-aryl-3-(alkylthio)-1H-1,2,4-triazoles were essentially inactive as anticonvulsants. The most potent antagonist of strychnine-induced convulsions was 5-(2-fluorophenyl)-4-methyl-3-(methylthio)-4H-1,2,4-triazole (3s), while the most selective antagonist was 5-(3-fluorophenyl)-4-methyl-3-(methylsulfonyl)-4H-1,2,4-triazole (3aa). The anticonvulsant profiles of these 4H-1,2,4-triazoles suggested that they were acting functionally like glycine receptor agonists. Since it has recently been postulated that compounds possessing glycine-agonist-like properties might be useful in the treatment of spasticity, we examined 5-phenyl-4-methyl-3-(methylsulfonyl)-4H-1,2,4-triazole (3c) in an in vivo model of spasticity. In this regard, 3c reduced the occurrence of hyperreflexia in rats that had received spinal transections 5-10 weeks previously. While triazole 3c appeared to possess glycine-agonist-like properties in vivo, it did not displace [3H]strychnine binding from rat brain stem/spinal cord membranes in vitro. On the other hand, 3c enhanced muscimol-stimulated 36Cl influx in a rat cerebellar membrane preparation, indicating a possible interaction of these triazoles with the GABAA receptor.

MDL 27,531 selectively reverses strychnine-induced seizures in mice.

1. Strychnine-sensitive glycine receptors are primarily localized in the brainstem and spinal cord where they are the major mediators of postsynaptic inhibition. A compound which acts functionally like a glycine receptor agonist would be potentially useful as a pharmacological tool and as a therapeutic agent for treating disorders of glycinergic transmission. 2. MDL 27,531 (4-methyl-3-methylsulphonyl-5-phenyl-4H-1,2,4-triazole) blocked strychnine-induced tonic extensor seizures in mice following either intraperitoneal (ED50 = 12.8 mg kg-1; 30 min) or oral (ED50 = 7.3 mg kg-1; 30 min) administration. Time course studies revealed a maximal effect at 30-60 min, though significant activity was still seen after 24 h. 3. MDL 27,531 was selective in antagonizing strychnine seizures and little or no activity was seen against seizures produced by other chemical convulsants (bicuculline; quinolinic acid; mercaptopropionic acid); by electrical stimuli (maximal electroshock); or by sensory stimuli (audiogenic seizure susceptible mice). MDL 27,531 blocked pentylenetetrazol-induced seizures with an ED50 = 55 mg kg-1. This profile differed from those of the anticonvulsants diazepam, valproic acid, and diphenylhydantoin. 4. The antagonism of strychnine seizures by MDL 27,531 occurred at doses that did not produce signs of sedation (suppression of spontaneous motor activity), motor ataxia (disruption of rotarod performance), muscle relaxation (inhibition of morphine-induced Straub tail), or CNS depression (potentiation of hexobarbitone sleep time). MDL 27,531 had less side effect potential (as derived from ratios obtained from the above measures) relative to those of the known muscle relaxants diazepam and baclofen. 5. Although MDL 27,531 behaved functionally like a selective agonist at the strychnine-sensitive glycine receptor, the compound did not alter the in vitro binding of [3H]-strychnine to mice brainstem/spinal cord membranes at concentrations of up to 100 microM. In further in vitro binding assays, MDL 27,531 at concentrations of up to 100 microM, did not displace the binding of [3H]-muscimol, [3H]-flunitrazepam, or["S]-t-butylbicyclophosphorthionate to rat cortical membranes. These ligands bind to the 7y-aminobutyric acid (GABA), benzodiazepine, and picrotoxin-convulsant binding sites, respectively.6. MDL 27,531 (10-100mgkg-', i.p.) enhanced binding of the benzodiazepine antagonist [3H]-Ro15-1788 to mouse cerebral cortex in vivo without directly affecting GABA levels.7. Ro 15-1788 (16, 32 mg kg-') significantly blocked the MDL 27,531 antagonism of strychnineinduced seizures, though this antagonism was not competitive. The same doses of Ro 15-1788 produced parallel rightward shifts in the dose-response curves for diazepam inhibition of pentylenetetrazol-induced seizures, consistent with a competitive antagonism.8. Thus, MDL 27,531 acts functionally like an agonist at the strychnine-sensitive glycine receptor in its capacity to reverse selectively strychnine-induced seizures. Though the precise mechanism of action of MDL 27,531 is unknown, MDL 27,531 may act at an allosteric site on the strychnine-sensitive receptor which produces agonist-like activity.

Pharmacological characterization of a GluR6 kainate receptor in cultured hippocampal neurons.

We have examined the pharmacology of kainate receptors in cultured hippocampal neurons (6-8 days in vitro (DIV)) from embryonic rats (E17). Cultured neurons were pre-treated with concanavalin A to remove kainate receptor desensitization and whole-cell voltage clamp electrophysiology employed to record inward currents in response to glutamatergic agonists and antagonists. N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid (AMPA) receptor responses were blocked using MK801 (3 microM) and the 2,3-benzodiazepine, LY300168 (GYKI53655, 50 microM), respectively. Inward currents were recorded in hippocampal neurons upon application of kainate and the 2S,4R isomer of 4-methyl glutamic acid (SYM2081) with EC50 values of 3.4 +/- 0.4 microM and 1.6 +/- 0.5 microM, respectively (n = 6 cells). The GluR5 selective agonists, LY339434 (100 microM) and (RS)-2-amino-3-(3-hydroxy-5-tert-butyl-4-isoxazolyl) propionic acid (ATPA) (100 microM), did not evoke detectable inward currents in any cell responding to kainate. LY293558 and the selective GluR5 antagonist, LY382884, had weak antagonist effects on responses evoked by either kainate or (2S,4R)-4-methyl glutamate (IC50 > 300 microM). The quinoxalinedione, 2,3-dihyro-6-nitro-7-sulfamoyl-benzo(f)quinoxaline (NBQX), blocked both kainate and (2S,4R)-4-methyl glutamate-activated currents at much lower concentrations (IC50 approximately 10 microM). These results provide pharmacological evidence that ion channels comprised of GluR6 kainate receptor subunits mediate kainate receptor responses in hippocampal neurons cultured 6-8 DIV.

Rapid down regulation of beta-adrenoceptors by co-administration of desipramine and fluoxetine.

Co-administration of desipramine and fluoxetine resulted in a 27% decline in cerebral cortical beta-adrenoceptor density after four days - a time point at which neither agent alone was effective. After 14 days, desipramine- and desipramine + fluoxetine-treated rats showed decreased receptor levels, with a greater decrement seen with the combined treatment. Fluoxetine, alone, had no affect on beta-adrenoceptor density at any time point examined. These effects are attributable to central serotonergic action since they were prevented by prior treatment with p-chlorophenylalanine. Cyproheptadine, a 5-HT2 antagonist, did not block these effects. Independent administration of fluoxetine and desipramine produced approximately 20% decrement in isoproterenol-stimulated cyclic AMP accumulation after four days of treatment. Co-administration of desipramine and fluoxetine resulted in a 35% decrement in cyclic AMP accumulation which was nearly additive with that produced by either drug alone. Consequently, the combination of a norepinephrine and serotonin uptake inhibitor may be an advantageous and rapid treatment for the alleviation of certain forms of depression.

A low dose of xylamine produces sustained and selective decreases in rat brain norepinephrine without evidence of neuronal degeneration.

The effects of a chronic partial depletion of rat cortical NE by a single dose of xylamine (20 mg/kg i.p.) on pre- and postsynaptic noradrenergic functionality were studied 4 hr, 14, 21 and 35 days after treatment. This dose of xylamine resulted in a 40 to 50% selective decrease in cortical levels of NE and the major metabolites of NE, 3,4-dihydroxyphenylethyleneglycol and 3-methoxy-4-hydroxyphenylethyleneglycol and, when measured after 35 days, [3H]desipramine binding and dopamine-beta-hydroxylase activity were at control levels, which would indicate that the NE nerve terminals in the cortex were intact. The 21- or 35-day deficit of NE did not affect alpha-1, alpha-2, beta, dopamine2, 5-hydroxytryptamine, or gamma-aminobutyric acidB receptor densities, or the beta receptor mediated adenylate cyclase activity. In addition, desipramine (10 mg/kg i.p.) administration for 14 days (days 20 through 34) was able to down-regulate beta receptor number (16% decrease) and reduce NE-stimulated adenylate cyclase activity (22% decrease), indicating that postsynaptic plasticity was still maintained. Affective disorders do not appear to be associated with a substantial (or readily measurable) decrease in brain NE concentrations and there is no consistent evidence of an altered beta receptor responsiveness. Thus, partial depletion of NE with xylamine might represent a biochemical model reflecting the involvement of NE in depression which could be used to investigate more sensitive markers of altered noradrenergic function.

Inflammatory cytokines enhance muscarinic-mediated arachidonic acid release through p38 mitogen-activated protein kinase in A2058 cells.

The human melanoma cell line A2058 expresses the Gq-coupled M5 subtype of muscarinic receptor. Stimulation with the cholinergic agonist, carbachol, induces a dose-dependent increase in arachidonic acid release. The carbachol-induced arachidonate release is potentiated two- to threefold by pretreatment of A2058 cells with either of the inflammatory cytokines, tumor necrosis factor-alpha or interleukin-1beta . Cytokine-induced enhancement of muscarinic-mediated arachidonic acid release peaks near 1 h. Western analysis suggests that both cytokines are capable of activating the nuclear factor-kappaB (NF-kappaB) and p38 mitogen-activated protein kinase (MAPK) pathways. Anisomycin (1 microM) treatment mimics the cytokine-induced enhancement of arachidonic acid production and activates the p38 MAPK pathway, but does not activate the NF-kappaB pathway. Furthermore, pre-treatment of A2058 cells with the putative p38 MAPK inhibitor, SB202190, ablates the cytokine-dependent augmentation without interfering with the muscarinic-mediated arachidonic acid release in untreated cells. Moreover, cytokine treatment does not affect other M5-coupled pathways (e.g., phospholipase C activity or intracellular Ca2+ mobilization), suggesting that p38 MAPK activation principally modulates muscarinic-mediated phospholipase A2 activity. Finally, in primary cultures of cells taken from rat cerebellum, key aspects of this finding are repeated in cultures enriched for glia, but not in cultures enriched for granule neurons.