• mRNA;
  • GABA;
  • Hippocampus;
  • Interneuron;
  • Brain slice;
  • Epilepsy


  1. Top of page
  2. Abstract

Summary:  Purpose: The balance between synaptic excitation and inhibition within the hippocampus is critical for maintaining normal hippocampal function. Even mild reduction in inhibition or enhancement of excitation can produce seizures. Synaptic excitation is produced by pyramidal cells and granule cells, whereas inhibition is produced by a smaller number of interneurons. To understand how two subpopulations of these excitatory and inhibitory neurons are regulated at the molecular level, we analyzed specific mRNA expression profiles for receptors that are significantly involved in synaptic transmission and in the synthesis and storage of the principal inhibitory neurotransmitter, γ-aminobutyric acid (GABA). Our hypothesis was that differences in gene expression between inhibitory and excitatory neurons in the rat hippocampus might point to specific new targets for seizure pharmacotherapy.

Methods: We combined the techniques of (a) whole-cell patch clamping in rat hippocampal slices, (b) biocytin staining for cell identification, (c) single-cell mRNA amplification, and (d) small-scale cDNA microarray analysis to allow us to obtain expression profiles for candidate genes from identified CA1 pyramidal neurons and interneurons. Electrophysiologic and morphologic data and expression profiles were obtained from 12 stratum pyramidale and seven stratum radiatum cells.

Results: Presumed inhibitory neurons expressed significantly more GAD65, GAD67, vGAT, GABAA-receptor α3, and N-methyl-d-aspartate (NMDA)-receptor IIB mRNA, and presumed excitatory neurons expressed more GABAA-receptor α1, and NMDA-receptor I mRNA.

Conclusions: Differential expression of candidate neurotransmitter-receptor subunits distinguished CA1 pyramidal neurons from interneurons. These differences may indicate potential new targets for altering the balance of inhibition and excitation in the treatment of epilepsy.

Hippocampal excitatory and inhibitory neurons differ from one another in a variety of features including morphology, action potential–firing characteristics, neurotransmitter choice, peptide composition, calcium-binding protein composition, neurotransmitter-receptor subunit expression, and neurotransmitter release mechanisms (1). There are likely to be many other fundamental differences between these two classes of neurons. In general, these differences have been detected when new techniques or reagents have become available and are then applied to neurons of interest, for example, immunohistochemistry with a new antibody, or molecular techniques to identify newly cloned receptors. It would be useful if a simultaneous identification of many of these phenotypic characteristics of inhibitory interneurons could be determined in a cell-type–specific manner. This would allow identification of the expression of genes of interest and unknown genes that subserve important functions in these neurons and could serve to establish a coordinated molecular phenotype of the neurons in question.

One specific area in which differences between excitatory and inhibitory neurons could have very significant functional consequences relates to the receptors they express for both glutamate and γ-aminobutyric acid (GABA), the two most important neurotransmitters involved in information processing within the hippocampus. These two transmitter systems also have been demonstrated to be critical for overall control of hippocampal excitability and seizures. The subunit composition of both glutamate and GABAA-receptor subtypes is likely to confer specific pharmacologic and physiologic properties to neuronal subpopulations, which would be important in antiepileptic drug (AED) therapy. Most of the work done analyzing the cellular distribution of glutamate and GABAA receptors has described protein expression by using immunohistochemistry (2,3). This technique can resolve some subtypes within each receptor grouping, but is often hard to evaluate at the single-cell level. In addition, immunostaining can only be used with one or, at most, two receptors at any time and is not suitable for quantitative estimates or for evaluating relative amounts of multiple subunits within single cells. Furthermore, correlation of the physiology of a neuron with expression of multiple ion-channel subtypes or neurotransmitter subtypes is difficult. Reverse transcription polymerase chain reaction at the single-cell level is useful for determining differences in gene expression between individual cell types but is limited in the number of mRNAs that can be investigated by sequence primers required.

Our strategy has been to combine the techniques of whole-cell patch clamping in rat hippocampal slices with a powerful new technique in molecular biology, single-cell antisense RNA (aRNA) amplification, which generates the rapid simultaneous analysis of many mRNAs to establish “expression profiles” of physiologically and anatomically characterized neurons (4). Our hypothesis is that inhibitory and excitatory neurons have differences in the transcription of glutamate or GABAA-receptor genes that would regulate their functional phenotype and determine how they respond to synaptic inputs. Our data implicate differences in gene expression that may regulate inhibition and provide new selective targets for epilepsy treatment. This could lead to new approaches in epilepsy therapy that can be targeted either to specific receptor subtypes at the receptor level with conventional pharmacotherapy or at the level of gene expression with gene therapy or antisense treatment.


  1. Top of page
  2. Abstract

Slice preparation and maintenance

Sprague–Dawley rats aged 3–4 weeks were deeply anesthetized with pentobarbital (PTB), decapitated, and their brains were removed; 400-μm-thick coronal sections of the hippocampus were prepared on a vibratome and bathed in cooled (4°C) artificial cerebrospinal fluid (aCSF), composed of (in mM): 145 NaCl, 3.0 KCl, 10 HEPES, 2.0 CaCl2, 1.0 MgCl2, and 8.0 glucose. Slices were transferred to a submersion-type recording chamber and perfused with room-temperature (∼25°C), oxygenated (95% O2/5% CO2) aCSF at a rate of 4 ml/min. Slices were allowed to equilibrate for a minimum of 1 h before recording.

Electrophysiological recording methods

Whole-cell patch recordings were made from visually identified CA1 pyramidal neurons in the stratum pyramidale and nonpyramidal cells in the stratum radiatum, with the use of infrared differential interference contrast (IR/DIC) videomicroscopy. Recordings were performed with glass (1.0-mm outer diameter) micropipettes pulled in two stages to tip diameters of 0.5–1.0 μm and resistance of 5–9 MΩ. The intracellular solution was composed of (in mM): 130 K-gluconate, 10 KCl, 10 N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES), 1.0 ethylene glycol-bis (B-aminoethyl ether)-tetraacetic acid (EGTA), 1.0 CaCl2, and 2.5 Mg/adenosine triphosphate (ATP). The solution was buffered to pH 7.25 with KOH. In all recordings, 0.5% biocytin (Sigma, St. Louis, MO, U.S.A.) was dissolved in the intracellular patch solution. After filling the soma and dendrites with biocytin, the soma contents were aspirated into the patch electrode and harvested into an Eppendorf tube for mRNA amplification and the synthesis of radiolabeled antisense RNA probes.


Cells were filled with biocytin by passive diffusion during whole-cell recording for ≥30 min, which was sufficient time for cell staining to be successful, even with aspiration of the somal contents. Slices were fixed in 4% paraformaldehyde and processed with the avidin-biotin-peroxidase method (Vector Labs ABC Kit, Burlingame, CA, U.S.A.) and reacted with diaminobenzidine dihydrochloride (DAB; Cappel, West Chester, PA, U.S.A.).

aRNA amplification

After aspirating an isolated neuron's somal contents, the electrode contents were collected with positive pressure into a reaction tube, and mRNA from single cells was amplified according to previously published methods (4,5). In brief, an 11-nucleotide oligo-dT primer coupled to a T7 RNA polymerase promoter sequence (oligo-dT-T7) hybridized to the poly-A tail of mRNA is used to convert mRNA into cDNAs in a reaction mixture containing avian myeloblastosis virus reverse transcriptase (AMVRT; 20 U/ml; Seikagaku America, Rockville, MD, U.S.A.), dNTP (containing dATP, dCTP, dGTP, and dTTP), dithiothreitol (DTT), 10× electrode buffer (containing 100 mM MgCl2, 1.2 M KCl, and 100 mM HEPES adjusted to pH 8.3), and RNAsin. This mixture was incubated for 60 min at 37°C for 5 min followed by blunt ending in a reaction mixture containing 200 mM pH 7.5 Tris, 100 mM MgCl2, 50 mM NaCl, 100 mM DTT, 2.5 mM dNTPs, and T4 DNA polymerase at 37°C for 15 min.

The repeated synthesis of antisense RNA (aRNA) from this double-stranded cDNA template results in a 1,000-fold amplification of mRNA recovered from individual neurons. aRNA amplification is performed with a reaction mixture made from 10× RNA amplification buffer, DTT, NTP (containing ATP, GTP, and UTP, 100 mM CTP), [α-[32P]CTP (800 Ci/mmol)], RNAsin, and T7 RNA polymerase (1,000 U/ml; Epicenter Technologies, Madison, WI, U.S.A.) and incubated at 37°C overnight. Samples from the reaction mixture are withdrawn for scintillation counting and denaturing RNA gels to assess the extent of aRNA synthesis and population size, respectively.

mRNA expression profiling

The resultant radiolabeled amplified aRNAs were used as probes of small-scale cDNA arrays containing candidate genes of interest. For each 32P-labeled aRNA from a cell, duplicated slot blots were used for each hybridization reaction. A mRNA expression profile could then be obtained by scanning the probed blot with a laser densitometric scanner (PhosphoImager model 425E, Molecular Dynamics) to quantitate the amount of radioactivity for each cDNA band. The intensity of each cDNA band signal represents the abundance of its corresponding mRNA in the cell. All 48 clones used in this study were sequenced to ensure their identity. cDNA clones were obtained as gifts from other laboratories included NT-3, NT-4, TRK-B, TRK-C, NGFR, and GFAP from James Eberwine (University of Pennsylvania), β−actin, NMDAR-1, 2A, 2B, 2C, 2D and GABAA receptors α1, α2, α3, α4, α5, α6, β1, β2, β3, γ2, and γ3 from Amy Brooks-Kayal (University of Pennsylvania), Kir 2.3, 4.1, 5.1 from J.P. Adelman (Oregon Health Sciences), Kir 6.2 from Makoto Takano (Kyoto University), Kir 1.1a from Jason Xu (Vanderbilt University Medical Center), NOS from David Bredt (Solomon Snyder-Brian Brooks; Johns Hopkins University), GAD65 and GAD67 from Niranjala Tillakartne-Jim (University of California, Los Angeles), Naβ1 and Na8.4(NaIIa) from Beth Sharp (University of Washington), vGAT from Robert Edwards (University of California, San Francisco), kv1.6 (kv2) from Leonard Kaczmarck (Yale University School of Medicine), and Kir3.1 from Norman Davidson (Caltech).

To determine the relative expression profile of any given mRNA from a single neuron, two normalization procedures were used. First, the hybridization intensity (amount of labeling) from a single gene was compared with the total hybridization signal on the array, providing an estimate of the expression of a given gene to the sum of the amplified expression of all the 48 genes being examined in that array. We did not find that the expression of any single gene was a sufficiently constant from cell to cell so that it could be used as an internal control. Second, when multiple members of a family of a common receptor were present (for example, 6 α subunits of the GABAA receptor), we analyzed the data as fraction of the mRNA hybridization of a given subtype versus total mRNA of that family (e.g., α1 GABAA receptor subunit as percentage of total GABAA receptor α subunit signal.) This latter approach allowed us to examine the relative amount of any given subunit expression within a specific receptor type. As a control for potential contamination from glia, any sample that exhibited significant amplification of glial fibrillary acidic protein (GFAP) was discarded from the analysis.

Statistical significance was determined for electrophysiologic data with the use of a two-tailed Student's t test for paired samples and for blot densities with the use of a univariate analysis of variance (ANOVA). Numeric values are expressed as means ± standard error of the mean (SEM). Hybridization intensity for each slot was measured with the PhosphoImager and expressed in relative terms to the total expression measured for 48 different cDNAs. Expressions of the 48 genes of interest were then compared with the ANOVA.


  1. Top of page
  2. Abstract

Identifying presumed excitatory and inhibitory cells

With the use of IR/DIC videomicroscopy, individual neurons were selected from the CA1 stratum pyramidale and stratum radiatum for whole-cell recording. Stratum pyramidale cells were packed tightly and had a pyramidal shape and a large apical dendrite. Stratum radiatum cells were smaller, round, and found isolated in their cell layer. After electrophysiologic recording for ≥15 min, during which the cell's physiology was characterized, the cytoplasmic contents of each cell were aspirated for aRNA amplification. Gentle aspiration followed by withdrawal of the electrode resulted in an intact soma and dendritic processes that could be identified by biocytin staining, which confirmed the identity of cells. Stratum pyramidale cells were spiny pyramidal cells, and stratum radiatum cells had a stellate shape with smooth dendrites and were varicosed (Fig. 1).


Figure 1. Anatomic computer reconstructions of biocytin-labeled cells are shown for a CA1 stratum pyramidale (upper) and stratum radiatum (lower) neuron with their characteristic voltage responses to current injection.

Download figure to PowerPoint

The firing and membrane properties of CA1 stratum pyramidale and stratum radiatum cells were typical. Stratum pyramidale cells fired in a regular spike-adapting firing pattern with wider action potentials, whereas stratum radiatum cells fired in a faster nonadapting pattern with thinner action potentials and a larger afterhyperpolarization (Fig. 1). The membrane properties for the stratum pyramidale cells were as follows: Vm, –67.6 ± 6.1; Rn, 133 ± 12.8; Tau, 48.5 ± 11.5; spike height, 98.9 ± 12.8. The membrane properties for the stratum radiatum interneurons were Vm, –59.2 ± 11.2; Rn, 228.6 ± 43.3; Tau, 26.6 ± 6.4; spike height, 78.27 ± 7.53.

Expression profiles for pyramidal versus stellate cells

Inhibitory interneurons from the stratum radiatum (n = 7) and excitatory pyramidal (n = 12) neurons from the stratum pyramidale were identified in slices by anatomic, morphologic biocytin staining, and electrophysiologic characterization. mRNA expression profiles were obtained for GABAA receptor subunits (α1-6, β1-3, γ2-3) (Fig. 2A), glutamate receptor subunits (GluR1,2, 4-7; NMDA-R1A, 2A-D) (Fig. 2B and C), the GABA-synthesizing enzymes, GAD65, GAD67, and the vesicular GABA transporter (vGAT) (Fig. 2D). Inhibitory cells express relatively more mRNA (p < 0.05) for GABAAα3 (8.6×) and NMDA-R2B (5×); the excitatory pyramidal neurons expressed relatively more NMDA-R1 (2.9×) and GABAAα1 (4.8×). Both types of neurons expressed ∼11–12% of their total mRNA as GABAAα receptor. However, within this family of receptor subtypes, pyramidal neurons had 21% of their GABAAα subunit mRNA as α1, compared with 4% in inhibitory neurons. By contrast, the inhibitory neurons had 17% of their GABAAα subunit as α3, compared with 2% of the pyramidal cell GABAAα subunit. Both cell types expressed similar fractions of α2, 4, 5 and 6.


Figure 2. aRNA expression profiles of CA1 pyramidal (stratum pyramidale) and stellate (stratum radiatum) neurons are shown. A:γ-Aminobutyric acid (GABA)A receptors. B: Glutamate receptors. C:N-methyl-d-aspartate (NMDA) receptors. D: GABA enzymes: GAD65, GAD67, VGAT. (* = p ≤0.05)

Download figure to PowerPoint

Both pyramidal and stellate neurons expressed NMDA-R2 family genes as similar proportions of the total amplified mRNAs on the microarrays (∼3%). However, within the NMDA-R2A-D receptor subclasses, pyramidal neurons expressed 28% of total NMDA-R2 expression as 2A and 10% as 2B, whereas stellate neurons expressed 11% as 2A and 49% as 2B. Thus inhibitory interneurons have higher relative expression of 2B as compared with excitatory neurons.

With regard to enzymes in the GABA-synthesis and metabolic pathways, as expected, the inhibitory neurons demonstrated a significant increase in expression of mRNA for vGAT (8.6×), GAD65 (2.3×), and GAD67 (2.1×), although the excitatory pyramidal neurons did express the mRNAs for both GAD65 and GAD67 well above background levels, as had previously been reported for excitatory neurons in culture and in acute slices (5).

No significant difference was seen in mRNA expression for the GluR2 receptor. Because Ca(2+)-permeable AMPA receptors are predominately expressed in GABAergic interneurons (6–8), and others have described reduced expression of GluR2 in inhibitory hippocampal neurons in culture, this finding was somewhat surprising; however, GluR2 immunoreactivity has been shown in both GABAergic interneurons and pyramidal cells (9).


  1. Top of page
  2. Abstract

The analysis of mRNA expression in single cells provides a strategy to compare the transciptional profiles of individual phenotypically characterized neurons. This approach has been implemented in human and experimental epilepsy models and in live, as well as fixed, cell types (10–13). The use of an oligo-dT primer and T7 RNA polymerase permits amplification of a broad population of expressed genes across many gene families. The size range and complexity of the amplified mRNA provides a comprehensive view of differential gene expression in single cells. (4)

Our results indicate that specific gene expression is different between CA1 pyramidal neurons and stellate neurons for several distinct components of the GABA synthetic pathway and glutamate and GABA receptors, whereas for many others, the expression profile is similar. This has implications for the function of these neurons, as well as for the possible development of highly targeted pharmacotherapies for epilepsy and for other neurologic or psychiatric conditions in which selective regulation of either excitatory or inhibitory neurons in the hippocampus would be therapeutic.

Our findings for the differential distribution of mRNA expression for enzymes in the GABA metabolic pathway are consistent with previous reports. Although immunohistochemistry for GABA, and GAD protein, are often used to identify inhibitory neurons, several previous reports have indicated that mRNA for GAD67 and GAD65 can be found in relatively high levels in excitatory neurons (5,13). Our results confirm that in these neurons, but also indicate that there is a relative increase in expression for the GAD mRNAs in the inhibitory neurons. Conversely, the relative distribution of mRNA for the vesicular GABA transporter (VGAT) had not been analyzed in this way and, based on our results, appears to be a more specific molecular indicator of the inhibitory neuronal phenotype. These findings also are consistent with the immunoperoxidase localization of VGAT antigen to hippocampal interneuron perikarya (14).

Although mRNA expression differs between excitatory and inhibitory neurons, the level of functional receptor protein, and any associated regulatory component, will determine the functional significance of the findings reported here. Routine biochemical analysis (e.g., Western blots) cannot be performed at the single-cell level to determine how protein expression correlates with mRNA expression. Immunohistochemistry is a more qualitative technique, although it can be resolved at the single-cell level. However, it cannot be used for more than two simultaneous antigens and is still not very specific for receptor subunits. It is likely, thus, that physiologic or pharmacologic analyses will be required to determine the functional significance of the expression differences. Even under these circumstances, however, the fact that all of the cells show different relative amounts of multiple subunits of each receptor would suggest that different functional consequences might be relatively subtle and difficult to identify, unlike the case with expression systems in which single subunits can be expressed and pharmacologically tested. Despite these difficulties, however, when hippocampal dentate gyrus granule cells were studied after an epileptogenic lesion, an approximately twofold change in the relative distribution of mRNA for GABAA-receptor subunits produced significant changes in neuronal pharmacology (10).

Some predictions can be made based on the subunit differences noted here. Routine AMPA receptor–mediated excitatory synaptic function should be similar between the two classes of neurons, although our data say nothing about presynaptic mechanisms that may influence short-term synaptic plasticity at these synapses (15,16). With regard to the NMDA receptors, it is known that the NMDA-R IIB has a 10-fold higher affinity for glycine (at the strychnine-resistant glycine regulatory site) than the NMDA-R IIA (17,18). When examining hippocampal or neocortical NMDA-receptor subtypes at the tissue level, it has been shown that during maturation, NMDA-R IIA becomes the dominant subtype, although lesser amounts of NMDA-R IIB persist (19). If inhibitory interneurons have a relatively higher proportion of NMDA-R IIB, than excitatory neurons, even in the mature brain, the glycine regulatory sites on NMDA receptors of inhibitory neurons would be more likely to be saturated by ambient extracellular concentrations of glycine. Excitatory neurons, in which NMDA-R IIA predominates, would be more likely to be unsaturated at the glycine site and therefore would be subject to synaptic modulation by local control of extracellular glycine (20) or by the addition of drugs that act on this regulatory site. Thus one potential target for reducing the hyperexcitability associated with epileptic events would be the NMDA-R IIA receptor. A subtype-specific antagonist might selectively dampen excitability in excitatory pathways but leave excitation of inhibitory interneurons unaffected.

The differential distribution of different subunits of the GABAA receptor between excitatory and inhibitory neurons also may have significant functional consequences. The relatively high expression of the α3 subunit in inhibitory interneurons may confer a relatively higher sensitivity to the inhibitory effects of Zn at the GABA receptor on these cells. Zinc is thought to be released predominantly from excitatory terminals, especially mossy fiber terminals of dentate granule neurons. These terminals also innervate inhibitory neurons, and the Zn may serve to dampen inhibition at GABAA receptors in local regions around the activated terminals, which could paradoxically serve to enhance feed-forward inhibition associated with the activated pathway. The relatively increased α3-subunit expression also may decrease the relative sensitivity of inhibitory interneurons to the benzodiazepines-like drugs (BZDs), such as zolpidem. How the differential sensitivity to zolpidem's enhancement of GABA-mediated inhibitory postsynaptic currents (IPSCs) would affect overall circuit behavior is difficult to predict, but the differential subunit expression between inhibitory and excitatory neurons could become the basis for selective effects of different BZDs on a variety of normal and pathologic functions, including epilepsy.

In summary, we demonstrate that mRNA expression for a subgroup of NMDA and GABAA-receptor subunits differs between CA1 pyramidal neurons and nearby inhibitory interneurons in the adult rat hippocampus. Most subunit mRNAs, however, are relatively similarly expressed. The differences found, if reflected in their coded proteins, could confer significant differences in their pharmacologic properties. The complexity of predicting the outcomes of some pharmacologic manipulations is multiplied, given that GABAergic excitation also exists in some cells under certain circumstances. The studies reported here have been confined to one region of the hippocampus, and it is highly likely that neurons in other areas will demonstrate different molecular phenotypes. Whether the differences between excitatory and inhibitory neurons remains consistent from region to region, both within the hippocampus and in other parts of the brain, remains to be determined. Parenthetically, Brooks-Kayal et al. (21) demonstrated a relatively high expression of the α1 subunit of the GABAA receptor in normal adult rat dentate granule cells, another excitatory neuron in the hippocampus.

It has been demonstrated that the subunit compositions of many receptors change after epileptogenic lesions, and one could hypothesize that other kinds of changes might occur after trauma or during the course of chronic neurodegenerative diseases. Ultimately, to understand these conditions and develop highly targeted therapies, molecular characterization of individual neuronal cell types in critical areas of the involved CNS is likely to be necessary. This will obviously be a monumental task in neurobiology but will be significantly simplified if some common themes emerge among identifiable neuronal subtypes.


  1. Top of page
  2. Abstract
  • 1
    Buzsaki G. Hippocampal GABAergic interneurons: a physiological perspective. Neurochem Res 2001;26: 899905.
  • 2
    Pickard L, Noel J, Henley JM, et al. Developmental changes in synaptic AMPA and NMDA receptor distribution and AMPA receptor subunit composition in living hippocampal neurons. J Neurosci 2000;20: 792231.
  • 3
    Macdonald RL, Olsen RW. GABAA receptor channels. Annu Rev Neurosci 1994;17: 569602.
  • 4
    Eberwine J, Yeh H, Miyashiro K, et al. Analysis of gene expression in single live neurons. Proc Natl Acad Sci U S A 1992;89: 30104.
  • 5
    Cao Y, Wilcox K, Martin C, et al. Presence of mRNA for glutamic acid decarboxylase in both excitatory and inhibitory neurons. Proc Natl Acad Sci 1996;93: 98449.
  • 6
    Racca C, Catania M, Monyer H, et al. Expression of AMPA-glutamate receptor B subunit in rat hippocampal GABAergic neurons. Eur J Neurosci 1996;8: 158090.
  • 7
    Leranth C, Szeidemann Z, Hsu M, et al. AMPA receptors in the rat and primate hippocampus: a possible absence of GluR2/3 subunits in most interneurons. Neuroscience 1996;70: 63152.
  • 8
    Tsuzuki K, Lambolez B, Rossier J, et al. Absolute quantification of AMPA receptor subunit mRNAs in single hippocampal neurons. J Neurochem 2001;77: 16509.
  • 9
    He Y, Hof PR, Janssen WG, et al. AMPA GluR2 subunit is differentially distributed on GABAergic neurons and pyramidal cells in the macaque monkey visual cortex. Brain Res 2001;921: 607.
  • 10
    Brooks-Kayal A, Shumate M, Jin H, et al. Selective changes in single cell GABA(A) receptor subunit expression and function in temporal lobe epilepsy. Nat Med 1999;4: 116672.
  • 11
    Crino PB, Duhaime AC, Baltuch G, et al. Differential expression of glutamate and GABA-A receptor subunit mRNA in cortical dysplasia. Neurology 2001;56: 90613.
  • 12
    White R, Hua Y, Scheithauer B, et al. Selective alterations in glutamate and GABA receptor subunit mRNA expression in dysplastic neurons and giant cells of cortical tubers. Ann Neurol 2001;49: 6778.
  • 13
    Sloviter R, Dichter M, Rachinsky T, et al. Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J Comp Neurol 1996;373: 593618.
  • 14
    Chaudhry FA, Reimer RJ, Bellocchio EE, et al. The vesicular GABA transporter, VGAT, localizes to synaptic vesicles in sets of glycinergic as well as GABAergic neurons. J Neurosci 1998;18: 973350.
  • 15
    Wilcox K, Dichter M. Paired pulse depression in cultured hippocampal neurons is due to a presynaptic mechanism independent of GABA(B) autoreceptor activation. J Neurosci 1994;14: 177588.
  • 16
    Cummings D, Wilcox K, Dichter M. Calcium-dependent paired-pulse facilitation of miniature EPSC frequency accompanies depression of EPSCs at hippocampal synapses in culture. Proc Natl Acad Sci U S A 1996;16: 531223.
  • 17
    Stern P, Behe P, Schoepfer R, et al. Single channel conductances of NMDA receptors expressed from cloned cDNAs: comparison with native receptors. Proc Natl Acad Sci U S A 1992;250: 2717
  • 18
    Williams K, Zappia AM, Pritchett DB, et al. Sensitivity of the N-methyl-d-aspartate receptor to polyamines is controlled by NR2 subunits. Mol Pharm 1994;45: 8039.
  • 19
    Sheng M, Cummings J, Roldan L, et al. Changing subunit composition of heteromeric NMDA receptors during development of rat cortex. Nature 1994;368: 1447.
  • 20
    Wilcox KS, Fitzsimonds RM, Johnson B, et al. Glycine regulation of synaptic NMDA receptors in hippocampal neurons. J Neurophysiol 1996;76: 341524.
  • 21
    Brooks-Kayal A, Jin H, Price M, et al. Developmental expression of GABAA receptor subunit mRNAs in hippocampal neurons in vivo and in vitro. J Neurochem 1998;70: 101728.