C-Fos brain mapping of global and subunit-specific NMDA receptor antagonists: Relevance for their potential use as antidepressants

Introduction

NMDA receptor antagonists as ketamine represent fast-acting alternatives to monoaminergic-based antidepressants. Major drawbacks of these drugs are psychosis-like states and cortical neurotoxicity, effects correlating with potent activation of the cingulated and retrosplenial cortex. The molecular mechanisms underlying these side-effects have not been deciphered yet.

Aims

We aimed to determine potential molecular components of the NMDA receptor implicated in their psychotomimetic action and investigated whether subunit-specific NMDA receptor antagonists also induce similar neurotoxic changes as ketamine.

Method

To investigate deleterious effects of NMDA receptor antagonists, we used brain mapping with the immediate early gene c-Fos. We analyzed the expression pattern of c-Fos in brain areas responsible for deleterious adverse events, after treatment with ketamine and the NR2B subunit-specific antagonist Ro 25-6981, both in wild-type and knockout mice, lacking either the entire NR2A subunit (NR2A ko mice) or its intracellular C-terminus (NR2A deltaC mice).

Results

In contrast to ketamine (10mg/kg), Ro 25-6981, even at high dosages (50mg/kg) does not induce any c-Fos expression in the cingulated and retrosplenial cortex of wildtype mice. However, Ro 65-2981 evokes, both in NR2A ko mice and NR2A deltaC mice, strong c-Fos expression in these areas.

Conclusions

Our data indicate that blockade of both NR2A and NR2B subunits is necessary to induce deleterious effects specific for ketamine. Deletion of the C-terminus of the NR2A subunit is sufficient to disinhibit, together with pharmacological NR2B blockade, neuronal networks associated with psychosis. Therefore, NR2B antagonists may represent safer alternatives to ketamine as potential antidepressants.

Spike-timing-dependent plasticity in hippocampal CA3 neurons

Synaptic plasticity of different inputs converging onto CA3 pyramidal neurons is central to theories of hippocampal function. The mossy fibre (MF) input to these neurons is thought to exhibit plasticity that is in nearly all aspects fundamentally different from plasticity in other brain regions: in particular, when induced by high frequency presynaptic stimulation, plasticity at these synapses is independent of NMDA receptor (NMDAR) activation and presynaptically expressed. Here, we show that different stimulation protocols that depend on the close timing of MF activity and postsynaptic spikes induce bidirectional plasticity in CA3 neurons in 3-week-old rats. Long-term potentiation (LTP) is observed when an excitatory postsynaptic potential (EPSP), evoked by MF stimulation, precedes a single postsynaptic action potential (AP) or a brief AP burst by 10 ms. Instead, timing-dependent long-term depression (LTD) requires the pairing of a single AP to an EPSP with a delay of 30 ms. The pairing of APs to synaptic activity is required for plasticity induction, since the application of unpaired APs or EPSPs did not alter synaptic strength. Furthermore, our results demonstrate that both timing-dependent LTP and LTD critically depend on the activation of NMDARs. Specifically blocking postsynaptic NMDARs prevents plasticity, demonstrating that NMDARs important to spike-timing-dependent plasticity in CA3 neurons are required at postsynaptic sites. In summary, this study shows that the close timing of APs to MF excitatory synaptic input can alter synaptic efficacy in CA3 neurons in a bidirectional manner.

Dominance of the lurcher mutation in heteromeric kainate and AMPA receptor channels

Homomeric glutamate receptor (GluR) channels become spontaneously active when the last alanine residue within the invariant SYTANLAAF-motif in the third membrane segment is substituted by threonine. The same mutation in the orphan GluRdelta2 channel is responsible for neurodegeneration in "Lurcher" (Lc) mice. Since most native GluRs are composed of different subunits, we investigated the effect of an Lc-mutated subunit in heteromeric kainate and AMPA receptors expressed in HEK293 cells. Kainate receptor KA2 subunits, either wild type or carrying the Lc mutation (KA2(Lc)), are retained inside the cell but are surface-expressed when assembled with GluR6 subunits. Importantly, KA2(Lc) dominates the gating of KA2(Lc)/GluR6(WT) channels, as revealed by spontaneous activation and by slowed desensitization and deactivation kinetics of ligand-activated whole-cell currents. Moreover, the AMPA receptor subunit GluR-B(Lc)(Q) which forms spontaneously active homomeric channels with rectifying current-voltage relationships, dominates the gating of heteromeric GluR-B(Lc)(Q)/GluR-A(R) channels. The spontaneous currents of these heteromeric AMPAR channels show linear current-voltage relationships, and the ligand-activated whole-cell currents display slower deactivation and desensitization kinetics than the respective wild-type channels. For heteromeric Lc-mutated kainate and AMPA receptors, the effects on kinetics were reduced relative to the homomeric Lc-mutated forms. Thus, an Lc-mutated subunit can potentially influence heteromeric channel function in vivo, and the severity of the phenotype will critically depend on the levels of homomeric GluR(Lc) and heteromeric GluR(Lc)/GluR(WT) channels.

Mutagenesis reveals a role for ABP/GRIP binding to GluR2 in synaptic surface accumulation of the AMPA receptor

We studied the role of PDZ proteins GRIP, ABP, and PICK1 in GluR2 AMPA receptor trafficking. An epitope-tagged MycGluR2 subunit, when expressed in hippocampal cultured neurons, was specifically targeted to the synaptic surface. With the mutant MycGluR2delta1-10, which lacks the PDZ binding site, the overall dendritic intracellular transport and the synaptic surface targeting were not affected. However, over time, Myc-GluR2delta1-10 accumulated at synapses significantly less than MycGluR2. Notably, a single residue substitution, S880A, which blocks binding to ABP/GRIP but not to PICK1, reduced synaptic accumulation to the same extent as the PDZ site truncation. We conclude that the association of GluR2 with ABP and/or GRIP but not PICK1 is essential for maintaining the synaptic surface accumulation of the receptor, possibly by limiting its endocytotic rate.

C-Terminal truncation of NR2A subunits impairs synaptic but not extrasynaptic localization of NMDA receptors

NMDA receptors interact via the extended intracellular C-terminal domain of the NR2 subunits with constituents of the postsynaptic density for purposes of retention, clustering, and functional regulation at central excitatory synapses. To examine the role of the C-terminal domain of NR2A in the synaptic localization and function of NR2A-containing NMDA receptors in hippocampal Schaffer collateral-CA1 pyramidal cell synapses, we analyzed mice which express NR2A only in its C-terminally truncated form. In CA1 cell somata, the levels, activation, and deactivation kinetics of extrasynaptic NMDA receptor channels were comparable in wild-type and mutant NR2A(Delta)(C/)(Delta)(C) mice. At CA1 cell synapses, however, the truncated receptors were less concentrated than their full-length counterparts, as indicated by immunodetection in cultured neurons, synaptosomes, and postsynaptic densities. In the mutant, the NMDA component of evoked EPSCs was reduced in a developmentally progressing manner and was even more reduced in miniature EPSCs (mEPSCs) elicited by spontaneous glutamate release. Moreover, pharmacologically isolated NMDA currents evoked by synaptic stimulation had longer latencies and displayed slower rise and decay times, even in the presence of an NR2B-specific antagonist. These data strongly suggest that the C-terminal domain of NR2A subunits is important for the precise synaptic arrangement of NMDA receptors.

Role of heteromer formation in GABAB receptor function

Recently, GBR1, a seven-transmembrane domain protein with high affinity for gamma-aminobutyric acid (GABA)B receptor antagonists, was identified. Here, a GBR1-related protein, GBR2, was shown to be coexpressed with GBR1 in many brain regions and to interact with it through a short domain in the carboxyl-terminal cytoplasmic tail. Heterologously expressed GBR2 mediated inhibition of adenylyl cyclase; however, inwardly rectifying potassium channels were activated by GABAB receptor agonists only upon coexpression with GBR1 and GBR2. Thus, the interaction of these receptors appears to be crucial for important physiological effects of GABA and provides a mechanism in receptor signaling pathways that involve a heterotrimeric GTP-binding protein.

Candidate editases for GluR channels in single neurons of rat hippocampus and cerebellum

RNA editing by site selective adenosine deamination changes codons in several nuclear transcripts in the mammalian brain and affects critical properties of the encoded proteins, as exemplified by the calcium permeability of AMPA receptor channels. The recently cloned RNA dependent adenosine deaminases ADAR1, ADAR2 and ADAR3 form a small family of sequence-related candidate editases which are expressed in brain and other tissues at distinct levels and patterns. We have employed single-cell polymerase chain reaction of hippocampal CA1 and CA3 pyramidal neurons and cerebellar Purkinje and Bergmann glial cells in an attempt to evaluate the expression of these enzymes at a cellular level. We found ADAR2 expressed in all cells analyzed; approximately 50% of the cells co-expressed ADAR1 or ADAR3. The differential ADAR expression revealed by our study might underlie the distinct editing efficiencies and selectivities in different GluR subunit transcripts.

Subtype-specific regulation of recombinant NMDA receptor-channels by protein tyrosine kinases of the src family

1. Tyrosine kinases regulate NMDA receptor-channel activity in cultured neurons, and NMDA receptor subunits are tyrosine phosphorylated in the brain. 2. Heteromeric NMDA receptor-channels were transiently expressed in human embryonic kidney (HEK) 293 cells and glutamate (100 microM)-activated whole-cell currents (500 ms) were studied when tyrosine kinases of the src gene family were included in the pipette solution. 3. Glutamate-activated currents (evoked every 20 s for up to 20 min) were increased by src and fyn kinases without affecting the desensitization and deactivation kinetics in NR1-NR2A but the kinases had no effects in NR1-NR2B, NR1-NR2C and NR1-NR2D receptor-channels, suggesting that a phosphorylation site in NR2A is targeted. 4. In a mutant channel consisting of NR1 and a C-terminal deletion mutant of NR2A (NR2A delta C), src and fyn kinases lost their potentiating effects indicating that the phosphorylation of tyrosine(s) in the C-terminal domain of NR2A affects the current flux through native NMDA receptor-channels.

Kindling increases N-methyl-D-aspartate potency at single N-methyl-D-aspartate channels in dentate gyrus granule cells

Dose-response studies of N-methyl-D-aspartate channel openings were carried out using cell-attached patches in dentate gyrus granule cells acutely isolated from control and kindled rats. The tips of the patch electrodes were first filled with regular extracellular solution, followed by backfilling through the shank with the agonist containing solution. As the two solutions joined, the agonist (N-methyl-D-aspartate, 25 microM) steadily diffused to the cell membrane, and the concentration gradually built up resulting in the progressive increase in the opening probability of N-methyl-D-aspartate channels. The reliability of this cell-attached diffusional drug delivery method was tested by determining the concentration dependence of competitive antagonism of N-methyl-D-aspartate induced channel activity by D(-)-2-amino-5-phosphonopentanoic acid. The Ki for D(-)-2-amino-5-phosphonopentanoic acid in the presence of 25 microM N-methyl-D-aspartate was found to be 6.8 microM. Twenty-four hours following the last seizure, N-methyl-D-aspartate channels on kindled neurons were consistently activated by lower N-methyl-D-aspartate concentrations than channels on control granule cells, indicating a higher potency of agonist at epileptic N-methyl-D-aspartate channels. The higher potency of the agonist is most likely a reflection of the long-term alterations in the modulation of N-methyl-D-aspartate receptor function in epileptic neurons.

High-level expression of functional glutamate receptor channels in insect cells

We have expressed glutamate-gated ion channels in Spodoptera frugiperda Sf21 insect cells using a recombinant baculovirus system. Cells infected with recombinant baculoviruses encoding the alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA)-selective glutamate receptor channel subunits GluR-B and GluR-D displayed specific high-affinity [3H]AMPA binding (apparent dissociation constant Kd of 15 nM for GluR-B and 40 nM for GluR-D) with pharmacological profiles typical of AMPA receptors. The binding reached maximal levels (Bmax of 15-30 pmol per mg of membrane protein) by 3-4 days postinfection. AMPA, glutamate and kainate triggered inward currents in GluR expressing cells, indicating assembly of functional homomeric channels. Formation of heteromeric GluR-B/D channels in doubly-infected cells was evident from the diagnostic current-voltage relations of AMPA-activated whole-cell currents. For the solubilization of the receptor, nonionic detergents Triton X-100, n-octyl-D-glucoside and n-dodecylmaltoside proved most effective. Detergent-solubilized receptor preparations were stable, retained their characteristic ligand-binding properties and bound to immobilized wheat germ lectin, demonstrating the glycosylation of insect cell-expressed GluR subunits. The expression level of 300-400 micrograms of receptor protein per liter of suspension culture should facilitate production of glutamate receptors for biochemical and structural studies.

NMDA receptor channels: subunit-specific potentiation by reducing agents

Sulfhydryl redox agents affect NMDA receptor activity. We investigated a putative redox site in four recombinant NMDA receptors. In 293 cells expressing NR1-NR2A channels dithiothreitol (DTT) rapidly potentiated L-glutamate-activated whole-cell currents and decreased the time course of desensitization and deactivation. Part of the current potentiation (reversible component) and all kinetic changes reversed upon washout. The remaining potentiation (persistent component) was abolished by an oxidizing agent. The N-terminal 370 residues of NR2A mediate the reversible component in chimeric NR2 subunits. In cells expressing the NR1-NR2B, -NR2C, and -NR2D channels DTT elicited only a slowly developing, persistent potentiation and increased the deactivation time course. In these, but not in NR1-NR2A, the DTT effect was rendered insensitive to reoxidation by alkylation. Reduced glutathione mimicked the DTT effects only in the NR1-NR2A receptor. Hence, molecularly distinct NMDA receptors differ profoundly in their responses to sulfhydryl redox agents.

Properties of NMDA receptor channels in neurons acutely isolated from epileptic (kindled) rats

The hyperexcitability accompanying chronic epileptiform activity may result from long-term alterations of ligand- and voltage-gated channels. Previous studies have indicated that NMDA responses and other electrophysiological characteristics of dentate gyrus granule cells are profoundly altered following chronic epilepsy (kindling). We have now investigated channels activated by NMDA using whole-cell patch-clamp and cell-attached single-channel recordings in granule cells acutely isolated from control and epileptic (kindled) rats. In control neurons, the amplitude of whole-cell NMDA currents was not sensitive to the presence of an intracellular ATP regeneration system, whereas NMDA currents in kindled cells showed a great variability, with larger amplitudes consistently recorded in the presence of intracellular high-energy phosphates. The ratio of peak to steady-state NMDA current (desensitization) was comparable (approximately 51%) in control and kindled neurons. Single-channel conductance determined from fluctuation analysis of whole-cell NMDA currents ranged between 21 and 35 pS in control and between 17 and 37 pS in kindled cells. Whole-cell NMDA channel noise power spectra yielded a single normal distribution of long channel lifetimes (mean, 4.3 msec) in control neurons, and the sum of two normal distributions (means, 4.6 and 7.1 msec) in kindled cells. The voltage-dependent Mg2+ block of NMDA channels was altered following kindling. From curves fitted to voltage-ramp-evoked currents in the presence of NMDA, the calculated affinity for Mg2+ of kindled channels at 0 mV was lower (12 mM) than that of controls (1.7 mM). Cell-attached recordings in the absence of Mg2+ have substantiated the lack of effect of kindling on single-channel conductance (approximately 50 pS), and have demonstrated large increases in mean open times (from 1.26 msec in control to 2.05 msec in kindled), burst lengths (from 1.91 msec to 4.18 msec), and cluster lengths (from 9.11 msec to 20.86 msec) of NMDA channels in kindled neurons. In summary, kindling, an NMDA receptor-dependent form of activity-dependent neuronal plasticity induced in vivo, results in lasting modifications in the function of single NMDA receptor channels that can be studied in acutely dissociated neurons. Kindling-induced epilepsy predominantly affects the mean open time, burst, and cluster duration of NMDA channels, their sensitivity to intracellular high-energy phosphates, and their block by Mg2+, but not the desensitization or single-channel conductance. Such alterations may reflect a change in the molecular structure of NMDA channels and may underlie the maintenance of the epileptic state.

The rat delta-1 and delta-2 subunits extend the excitatory amino acid receptor family

We have characterized a second member (delta-2) of a new class of subunits for the ligand-gated excitatory amino acid receptor superfamily. The sequence of delta-2 exhibits an average identity of 25% and 18.5% to the non-NMDA and NMDA receptor subunits, respectively. The rat delta-2 gene is expressed predominantly in Purkinje cells of the cerebellum whereas only low levels of delta-1 transcripts are found in the adult brain. However, delta-1 gene expression undergoes a pronounced developmental peak, with particularly high mRNA levels in the caudate putamen of late embryonic/early postnatal stages.

Endogenous intracellular calcium buffering and the activation/inactivation of HVA calcium currents in rat dentate gyrus granule cells

Granule cells acutely dissociated from the dentate gyrus of adult rat brains displayed a single class of high-threshold, voltage-activated (HVA) Ca2+ channels. The kinetics of whole-cell Ca2+ currents recorded with pipette solutions containing an intracellular ATP regenerating system but devoid of exogenous Ca2+ buffers, were fit best by Hodgkin-Huxley kinetics (m2h), and were indistinguishable from those recorded with the nystatin perforated patch method. In the absence of exogenous Ca2+ buffers, inactivation of HVA Ca2+ channels was a predominantly Ca(2+)-dependent process. The contribution of endogenous Ca2+ buffers to the kinetics of inactivation was investigated by comparing currents recorded from control cells to currents recorded from neurons that have lost a specific Ca(2+)-binding protein, Calbindin-D28K (CaBP), after kindling-induced epilepsy. Kindled neurons devoid of CaBP showed faster rates of both activation and inactivation. Adding an exogenous Ca2+ chelator, 1,2-bis-(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA), to the intracellular solution largely eliminated inactivation in both control and kindled neurons. The results are consistent with the hypothesis that endogenous intraneuronal CaBP contributes significantly to submembrane Ca2+ sequestration at a concentration range and time domain that regulate Ca2+ channel inactivation.

Calbindin-D28K (CaBP) levels and calcium currents in acutely dissociated epileptic neurons

Nerve cells that lack the cytoplasmic Ca2+ binding protein Calbindin-D28K (CaBP) appear to be selectively vulnerable to Ca(2+)-related injury consistent with a postulated intraneuronal Ca(2+)-buffering role of CaBP. We have confirmed the selective loss of CaBP from the dentate gyrus during kindling-induced epilepsy in acutely dissociated granule cells (GCs) from kindled rats. Immunohistochemically stained kindled neurons showed a significant loss of CaBP when compared to controls (p less than 0.001; ANOVA). The Ca(2+)-buffering role of CaBP was assessed in acutely dissociated control and kindled GCs by examining a physiological process highly sensitive to intracellular Ca(2+)-buffering: the Ca(2+)-dependent inactivation of high-voltage activated (HVA or L-type) Ca2+ currents in the absence (or presence) of exogenous Ca(2+)-chelators. Whole-cell patch clamp recordings in kindled GCs demonstrated a markedly enhanced Ca(2+)-dependent inactivation of Ca(2+)-currents. After brief conditioning Ca2+ currents, in the absence of an exogenous intraneuronal Ca(2+)-chelator, subsequent test Ca2+ currents were inactivated by 58.3% in kindled GCs, a significant increase from the 37.4% inactivation observed in control GCs (p less than 0.005; ANOVA). The differential Ca2+ current decay and Ca(2+)-dependent inactivation were prevented in both control and kindled GCs upon loading the neurons with the exogenous Ca(2+)-chelator BAPTA. These experiments demonstrate a high correlation between the loss of CaBP and changes in Ca2+ current inactivation and are consistent with the hypothesis that CaBP contributes to the physiological Ca(2+)-buffering in mammalian neurons.

Effects of the tetronic acid derivatives AO33 (losigamone) and AO78 on epileptiform activity and on stimulus-induced calcium concentration changes in rat hippocampal slices

The effects of members of a new class of anticonvulsants, the tetronic acid derivatives, were studied in 3 in vitro models of epileptogenesis in rat hippocampal slices; the picrotoxin, the low magnesium and the low calcium model. The effects of AO33 (losigamone) and AO78 on stimulus-induced decreases in extracellular calcium concentration were also investigated. In all 3 models of epileptogenesis, both drugs blocked spontaneous and reduced stimulus-induced epileptiform discharges dose dependently and reversibly. Stimulus-induced changes in [Ca2+]0 were markedly diminished by these agents. The fact that the tetronic acid derivatives block the low Ca seizure-like events which develop independently from chemical synaptic transmission suggests that these agents have non-synaptic or direct membrane actions with subsequently reduced cellular excitability.

Anticonvulsant effects of tetronic acid derivatives on picrotoxin induced epileptiform activity in rat hippocampal slices

We have investigated the effects of a new class of anticonvulsants, the tetronic acid derivatives AO33 (generic name: losigame) and AO78, on field potentials, extracellular calcium concentration changes and intracellular potentials in rat hippocampal slices treated with the non-competitive GABAA antagonist picrotoxin (PTX). The tetronic acid derivatives reduced and eventually blocked spontaneous epileptiform events, induced by 10 to 30 microM PTX. Stimulus induced burst discharges were shortened in duration, but not blocked. Extracellular calcium concentration changes and associated slow negative field potentials were diminished in a dose dependent manner. Intracellular recordings revealed no effect of AO33 on resting membrane potential, little effect on input resistance, a small increase in the threshold of action potentials and an attenuation of stimulus induced paroxysmal depolarisation shifts (PDSs). Spontaneous PDSs initially decreased in duration until they were no longer observable.

Effects of NMDA antagonists on picrotoxin-, low Mg2+- and low Ca2+-induced epileptogenesis and on evoked changes in extracellular Na+ and Ca2+ concentrations in rat hippocampal slices

The anticonvulsant properties of ketamine and 2-APV were compared on 3 types of convulsant activity in hippocampal area CA1: the 'picrotoxin-epilepsy,' the 'low magnesium epilepsy' and the 'low calcium epilepsy.' In particular the spontaneous activity, the synaptically evoked responses and the changes in [Ca2+]0 were examined, since in many cases of epilepsy, Ca2+ uptake into cells is enhanced. In normal medium, ketamine and 2-APV have nearly no effect on stimulus evoked decreases in [Ca2+]0, although they clearly depress NMDA-induced ionic changes. However, ketamine and 2-APV prevent to some extent the augmentation of stimulus-induced changes in [Ca2+]0, observed after treating slices with picrotoxin or Mg2+-free medium. This extra Ca2+ uptake is probably mediated by NMDA operated channels. Our findings also show that ketamine, like 2-APV, has a stronger anticonvulsant effect on the low Mg-than on the picrotoxin-induced epileptiform activity. Responses to iontophoretically applied NMDA are facilitated in the 'low calcium epilepsy' and can be selectively blocked by ketamine. Spontaneous epileptiform activity occurring in low calcium can be blocked by ketamine only when some synaptic transmission is still present.

Differences in magnesium and calcium effects on N-methyl-D-aspartate- and quisqualate-induced decreases in extracellular sodium concentration in rat hippocampal slices

Decreases in extracellular sodium concentration [( Na+]o) and associated slow negative field potentials (fp's) were monitored with double barreled sodium sensitive/reference microelectrodes in area CA1 of rat hippocampal slices during iontophoretic application of the glutamate receptor agonists N-methyl-D-aspartate (NMDA) and quisqualate (quis). The effects of lowering [Ca2+]o on these signals were compared to those of lowering [Mg2+]o. Both NMDA- and quis-induced decreases in [Na+]o of up to 60 mM and in the fp's of up to 8 mV. Decreasing [Mg2+]o enhanced NMDA-induced signals, whereas quis-induced signals were unaffected. Lowering [Ca2+]o also enhanced NMDA signals, although somewhat less than lowering [Mg2+]o. This effect was still present, even when voltage dependent Na+ currents were blocked by 10(-7) tetrodotoxin. Interestingly, quis-induced signals could be enhanced in a low Ca2+ medium as well, but only when high quis concentrations were used. The results suggest that, during the sorts of large decreases of [Ca2+]o observed during seizure activity, activation of NMDA receptors is facilitated.