It has been postulated that endogenous kynurenic acid (KYNA) modulates α7* nicotinic acetylcholine receptor (nAChR) and NMDA receptor activities in the brain.a To test this hypothesis, α7* nAChR and NMDA receptor functions were studied in mice with a targeted null mutation in the gene encoding kynurenine aminotransferase II (mKat-2-/- mice), an enzyme responsible for brain KYNA synthesis. At 21 postnatal days, mKat-2-/- mice had lower hippocampal KYNA levels and higher spontaneous locomotor activity than wild-type (WT) mice. At this age, α7* nAChR activity induced by exogenous application of agonists to CA1 stratum radiatum interneurons was ∼65% higher in mKat-2-/- than WT mice. Binding studies indicated that the enhanced receptor activity may not have resulted from an increase in α7* nAChR number. In 21-d-old mKat-2-/- mice, endogenous α7* nAChR activity in the hippocampus was also increased, leading to an enhancement of GABAergic activity impinging onto CA1 pyramidal neurons that could be reduced significantly by acute exposure to KYNA (100 nM). The activities of GABAA and NMDA receptors in the interneurons and of α3β4* nAChRs regulating glutamate release onto these neurons were comparable between mKat-2-/- and WT mice. By 60 d of age, KYNA levels and GABAergic transmission in the hippocampus and locomotor activity were similar between mKat-2-/- and WT mice. Our findings that α7* nAChRs are major targets for KYNA in the brain may provide insights into the pathophysiology of schizophrenia and Alzheimer's disease, disorders in which brain KYNA levels are increased and α7* nAChR functions are impaired.
Neuronal nicotinic acetylcholine receptors (nAChRs) are acetylcholine-gated channels that are essential for development and plasticity of the mammalian CNS (McGehee, 2002). Of the several neuronal nAChR subtypes identified to date, α7*, α4β2*, and α3β4* nAChRs are present in the hippocampus (Albuquerque et al., 1997).a For several reasons, α7* nAChRs are unique among nAChRs: (1) they have high Ca2+ permeability, short open time, and fast rate of 1desensitization, (2) they can be blocked by nanomolar concentrations of methyllycaconitine (MLA) or α-bungarotoxin (α-BGT), and (3) they can be fully activated by the ACh metabolite choline (Pereira et al., 2002). In addition, α7* nAChRs and chimeric α7/5-HT3 receptors in HEK293 cells are highly sensitive to blockade by kynurenic acid (KYNA) (IC50 of ∼7 μM), a major brain metabolite of the kynurenine pathway of tryptophan degradation (Hilmas et al., 2001; Pereira et al., 2002).
The well established neuroprotective and anticonvulsant properties of KYNA have been traditionally attributed to its action as a competitive antagonist of glycine at NMDA receptors (Stone, 1993, 2001; Schwarcz and Pellicciari, 2002). In the nominal absence of glycine and in the presence of glycine concentrations similar to those found in the extracellular compartment of the brain (Kennedy et al., 2002), KYNA blocks NMDA receptors with IC50 values of ∼15 and 230 μM, respectively (Hilmas et al., 2001). Considering that, under physiological conditions, brain KYNA concentrations range from low nanomolar to low micromolar (Moroni et al., 1988; Turski et al., 1988), it is unclear whether sufficient KYNA is present to regulate NMDA receptor activity in vivo (Scharfman et al., 1999; Urenjak and Obrenovitch, 2000).
Brain KYNA is synthesized primarily in astrocytes as a result of the irreversible transamination of L-kynurenine. In the adult rat and the human brain, >70% of this synthesis is catalyzed by kynurenine aminotransferase II (KAT II) (L-α-aminoadipate aminotransferease; E.C. 126.96.36.199); the remainder is catalyzed by KAT I (Guidetti et al., 1997; Kiss et al., 2003). Therefore, the actions of endogenous KYNA in the brain could be elucidated by in vivo manipulations of KAT II activity. The lack of selective KAT II inhibitors, however, and the characterization of the mouse Kat-2 gene led to the development of a KAT II-deficient mouse by targeted deletion (Yu et al., 1999a,b).
The present study uses mKat-2-/- mice to test the hypothesis that α7* nAChRs mediate the actions of endogenous KYNA in the brain. Results presented herein reveal that, at 21 d of age, mKat-2-/- mice had lower hippocampal KYNA levels and higher locomotor activity than wild-type (WT) mice. At this age, α7* nAChR activity in CA1 stratum radiatum (SR) interneurons and GABAergic synaptic activity impinging onto CA1 pyramidal neurons were also higher in mKat-2-/- than in WT mice. In contrast, NMDA receptor activity in the hippocampus of 21-d-old mutant and WT mice did not differ significantly. Furthermore, hippocampal levels of KYNA in WT and mKat-2-/- mice became similar by 60 d of age, as did GABAergic synaptic activity impinging onto CA1 pyramidal neurons. Thus, in the hippocampus, KYNA is an endogenous modulator of the α7* nAChR activity.
Parts of this work have been published previously in abstract form (Alkondon et al., 2003b).
Materials and Methods
Genotyping mice nullizygous for mKat-2 (129SvEv)
The mouse KAT II sequence and genomic organization were determined from BAC clones made from a 129/SVJ genomic library (Yu et al., 1999a,b). Mice with targeted deletion of the mKat-2 gene were made at the National Institutes of Health under the Institutional Animal Care and Use Committee-approved animal protocol G-99-3 (Yu et al., 1999a). The animals were maintained in a 129SvEv background strain (Taconic, Germantown, NY). Two pairs of primers (see Fig. 1 A) were used for screening and genotyping the mice: P1 (5′-ACA TGC TCG GGT TTG GAG AT-3′) and P3 (5′-AAG CTT TGG AAC TCA GTG GG-3′), and P2 (5′-GTG GAT GTG GAA ATG TGT GTG GG-3′) and P4 (5′-GAG ACA GAC ACC TTG ATA CT-3′). Genotypes were verified by PCR analysis under the following conditions. The reaction contained 50 ng of mouse tail DNA, in MasterAmp buffer (Epicentere Technologies, Madison, WI), with 5 pmol of each primer and 0.2 U of TaqDNA polymerase (Roche, Indianapolis, IN). Thermocycling parameters were as follows: 2 min at 94°C, followed by 25 cycles of 40 sec at 94°C, 40 sec at 58°C, and 1 min at 72°C. Once mKat-2-/- mice were successfully generated, homozygous breeding pairs were established. Separate 129SvEvTac mice were bred to form a homozygous WT line. To minimize the risk of genetic drift in either homozygous line, mKat-2-/- mice were paired with WT 129SvEvTac mice after every fourth generation. mKat-2-/- and WT mice from these pairings served as the breeders for subsequent generations. Mice were maintained in an Association for Assessment and Accreditation of Laboratory Animal Care-approved animal facility on a 12 hr light/dark cycle, with food and water available ad libitum. Animal care and handling were done strictly in accordance with the guidelines set forth by the Animal Care and Use Committee of the University of Maryland, Baltimore and National Institutes of Health.
Total RNA from kidneys of 2-month-old WT and mKat-2-/- mice were extracted using the Trizol reagent according to the protocol of the manufacturer (Invitrogen, Gaithersburg, MD). Total RNA of each (10 μg) was subjected to electrophoresis on a 0.8% agarose gel and transferred to nitrocellulose membrane. The membrane was probed with [α-32P]-labeled full-length mKat-2 cDNA in preHYB/HYB buffer (Quality Biological, Gaithersburg, MD), washed at high-stringency conditions with 0.2× SSC, and subjected to autoradiography at -70°C.
Immunotitration of KAT II activity in mouse hippocampal extracts
Male mice were anesthetized in a CO2 atmosphere and killed by decapitation. Their hippocampi were rapidly removed and placed on ice. The tissues were suspended (1:20, w/v) in distilled water and sonicated. The homogenate was then centrifuged (10,000 × g), and the supernatants were dialyzed overnight at 4°C against 4 l of 5 mm Tris acetate buffer, pH 8.0, containing pyridoxal-5-phosphate (50 μm) and 2-mercaptoethanol (10 mm). KAT II was immunoprecipitated using a polyclonal rabbit anti-human KAT II antibody (Okuno et al., 1993; Yu et al., 1999b). To this end, either 1 μl of antibody or 1 μl of water was mixed with 100 μl of the dialyzed tissue supernatant and incubated at 25°C for 1 hr. One milligram of insoluble protein A, dissolved in 10 μl of water, was then added, and the mixture was vigorously stirred and incubated at 25°C for 1 hr. After centrifugation (12,000 × g, 10 min), KAT activity was measured in a total volume of 200 μl, containing 80 μl of supernatant fluid, 150 mm Tris-acetate buffer, pH 7.4, 2 μm [3H]kynurenine (2.4 nCi), 1 mm pyruvate, and 80 μm pyridoxal-5-phosphate. Samples were incubated for 24 hr at 37°C, and the reaction was terminated by adding 50% (w/v) trichloroacetic acid. The denatured protein was removed by centrifugation, and 1 ml of the resulting supernatant was added to a Dowex 50W H+ cation exchange column. After successive washes with 0.1 m HCl and distilled water, [3H]KYNA was eluted from the column with 2× 1 ml of distilled water and quantified by liquid scintillation spectrometry (Guidetti et al., 1997).
Measurement of kynurenine, quinolinic acid, and KYNA levels in the brain
Animals were anesthetized in a CO2 atmosphere and killed by decapitation. Their brains were rapidly removed from the skull and placed on ice. The hippocampi were dissected out, placed on dry ice, and stored at -80°C. On the day of the assay, the tissue was thawed and homogenized (1:10, w/v) in ultrapure water. Aliquots (100-200 μl) of the homogenates were diluted to 450 μl. Then, internal standard (deuterated KYNA, 50 μl) was added, and the solution was acidified with HCl (100 μl, 5N). To remove fatty compounds, chloroform (500 μl) was added to the acidified aliquots. After centrifugation (10 min, 10,000 × g), a 500 μl aliquot of the supernatant was applied to a Dowex 50W H+ column. The column was washed successively with HCl (200 μl, 1N), distilled water (500 μl), and methanol (500 μl, 50% in water), and KYNA was eluted with methanol (2 ml, 50% in water). Tetrabutylammonium hydrogen sulfate (50 μl) was added to 1 ml of the eluate, and the solution was evaporated to dryness. The samples were then reacted with 25 μl of acetone containing 7.5% of diisopropylethylamine and 3% pentafluorobenzyl bromide (PFB-Br) at 60-65°C for 15 min. Subsequently, 50 μl of decane and 750 μl of water were added. The samples were thoroughly mixed, and the decane phase was removed. One microliter of the decane phase was injected into a gas chromatograph. The gas chromatography (GC) with electron capture negative ionization mass spectrometry (MS) system (ThermoFinnigan, San Jose, CA) was a Trace GC coupled with a Trace MS quadrupole mass spectrometer. Chromatographic separation was achieved using a 30 m Rtx-200MS capillary column (0.25 mm inner diameter, 0.5 μm film thickness; Restek, Bellafonte, PA) with helium as a carrier gas. A split-splitless injection port (1 μl injection volume) was used. The temperature program was as follows: 155°C for 1.25 min, 40°C/min to 240°C, 10°C/min to 325°C, 1 min at 325°C, 18°C/min to 155°C, and 1.5 min at 155°C, injection port at 250°C. The GC/MS uses electron capture negative ion chemical ionization mass spectroscopy with methane as the reagent gas. The ion source temperature was 220°C. Selected ion monitoring was performed by recording the signal of the characteristic di-PFB derivative [mass/charge ratio (m/z) of 368] (Naritsin et al., 1995).
Levels of quinolinic acid and kynurenine were also quantified by GC/MS. To this end, the tissue homogenate used for determination of KYNA was further diluted (1:2.5, v/v) in water. Fifty microliters of internal standard (100 nm 3,5-pyridinedicarboxylic acid and 200 nm homophenylalanine) were added to 50 μl of homogenate, and proteins were precipitated with HCl (25 μl, 5N). Fatty compounds were removed by extraction with chloroform (100 μl). After centrifugation (10 min, 10,000 × g), 100 μl of the acidic supernatant were added to a glass tube containing 62.5 μl of 50 mm tetrabutylammonium hydrogen sulfate and lyophilized overnight. The samples were then reacted with 25 μl of methylene chloride containing 7.5% of diisopropylethylamine and 3% PFB-Br at 60-65°C for 15 min. Subsequently, 50 μl of decane and 750 μl of water were added. The samples were thoroughly mixed, and the decane phase was removed. One microliter of the decane phase was injected into the gas chromatograph. Chromatographic separation and GC/MS analysis was achieved using a 30 m Rtx-5MS capillary column (0.25 mm inner diameter, 0.25 μm film thickness; Restek) and the same GC/MS instrument, injection port, and ion source temperature as described above for determination of KYNA levels, again using electron capture negative ion chemical ionization. The temperature program was as follows: 155°C for 1.25 min, 40°C/min to 275°C, 10°C/min to 320°C, 1 min at 320°C, injection port at 228°C. For quinolinic acid and kynurenine, the characteristic (M-PFB)- ions are di-PFB derivatives, with m/z of 346 and 387, respectively (Naritsin et al., 1995).
Assessment of locomotor activity
Male mutant and WT mice (n = 10 animals per group) were assessed for changes in locomotor activity (Reddy et al., 1999). Mice were placed in an open-field arena (35 × 42 cm) for a period of 5 min. The floor of the arena was divided into 7 cm2 sectors, and locomotor activity was quantified as the number of sector crossings during the 5 min session. Data were obtained from a videotape by an investigator who was unaware of the genotype of the test animals.
Preparation of hippocampal slices. Slices of 250 μm thickness were obtained from the hippocampus of 21- or 60-d-old male mice by procedures similar to those described previously for rats (Alkondon and Albuquerque, 2001). Slices were stored at room temperature in artificial CSF (ACSF), which was aerated with 95% O2 and 5% CO2 and composed of the following (in mm): 125 NaCl, 25 NaHCO3, 2.5 KCl, 1.25 NaH2PO4,2 CaCl2, 1 MgCl2, and 25 glucose. In most experiments, biocytin labeling was used to identify morphologically the neurons from which recordings were obtained (Alkondon and Albuquerque, 2001).
Recordings. Nicotinic whole-cell currents, EPSCs, and IPSCs were recorded from the soma of CA1 pyramidal neurons or SR interneurons according to the standard whole-cell patch-clamp technique (Hamill et al., 1981), using an LM-EPC7 amplifier (List Electronic, Darmstadt, Germany). Pyramidal neurons and SR interneurons in the CA1 field of the slices were visualized by means of infrared-assisted videomicroscopy. In some experiments, recordings were also obtained from neurons in the CA1 pyramidal layer visualized under light microscopy. Agonists were applied to the slices via a U-tube, and antagonists were applied via bath perfusion (Alkondon and Albuquerque, 2001). The U-tube had a pore diameter of 250 μm and was positioned ∼200 μm above the surface of the slice. Signals were filtered at 3 kHz (eight-pole, low-pass Bessel filter) and either recorded on a videotape recorder for later analysis or digitally sampled on-line at 1 or 10 kHz by a microcomputer using the Digidata 1200 or 1322A and the pClamp6 or pClamp9 software (Axon Instruments, Foster City, CA). Slices were superfused with ACSF at 2 ml/min. Nominally Mg2+-free ACSF was used in experiments in which NMDA-evoked whole-cell currents were studied. Atropine (1 μm) was included in the ACSF in all experiments to inhibit muscarinic receptors. In some experiments, bicuculline (10 μm) was added to the ACSF to block GABAA receptors. Patch pipettes were pulled from borosilicate glass capillary (1.2 mm outer diameter) and, when filled with internal solution, had resistances between 3 and 6 MΩ. The series resistance ranged from 8 to 20 MΩ. At -68 mV, the leak current was generally between 25 and 100 pA. Data were not included in the analysis when the leak current exceeded 150 pA. For most voltage-clamp recordings, the internal solution contained 0.5% biocytin in addition to the following (in mm): 10 EGTA, 10 HEPES, 130 Cs-methane sulfonate, 10 CsCl, 2 MgCl2, and 5 lidocaine N-ethyl bromide (QX-314). The pH was adjusted to 7.3 with CsOH, and the osmolarity was 340 mOsm. In experiments in which GABA-evoked currents were studied, the internal solution contained the following (in mm): 120 CsCl, 10 HEPES, 10 EGTA, 2 MgCl2, and 5 QX-314, pH adjusted to 7.3 with CsOH. Membrane potentials were corrected for liquid junction potentials. In some neurons, recordings of fast current transients (the counterparts of action potentials under voltage clamp) were initially obtained under cell-attached configuration before entering into the whole-cell configuration. Thus, the ability of nicotinic agonists to trigger action potential firing in the interneurons could be assessed without altering the intracellular environment. All experiments were performed at room temperature (20-22°C).
Data analysis. The frequency, peak amplitude, 10-90% rise time, and decay-time constants (τdecay) of EPSCs mediated by NMDA and AMPA receptors were analyzed using the WinEDR V2.3 (Strathclyde Electrophysiology Software, Glasgow, Scotland). The net charge of the NMDA component of EPSCs as well as the amplitude and/or net charge of whole-cell currents evoked by different agonists were calculated using the pClamp9 software. The Mini Analysis software (Synaptosoft, Leonia, NY), which uses a threshold-based event-detection algorithm, was also used to analyze IPSCs recorded from CA1 pyramidal neurons. Frequencies and amplitudes of IPSCs were measured in 5 min recordings. Area and amplitude thresholds were set above the noise level and kept constant in each experiment. Events that did not show a typical synaptic waveform were rejected manually. Histograms of IPSC amplitude distributions were made from measurements of amplitude versus number of events. In the histograms, the amplitudes were grouped in 3 pA bins. Cumulative distributions of IPSC amplitudes and interevent intervals, which make no assumption regarding the transmitter release process, were analyzed and compared using the nonparametric Kolmogorov-Smirnov (K-S) test.
Membrane preparation. Hippocampi were homogenized in 2 ml of ice-cold HBSS containing antiproteases (2 mm PMSF, 1 μg/ml pepstatin A, 2 μg/ml aprotinin, and 0.3 μg/ml leupeptin), and the homogenates were centrifuged at 4000 × g at 4°C for 15 min. The tissue pellets were resuspended in antiprotease-containing HBSS (2 ml) and freeze thawed five times using liquid nitrogen. The suspensions were then centrifuged at 12,000 × g at 4°C for 15 min, and the membrane pellets were resuspended in the antiprotease-containing HBSS (1 ml) and stored at -80°C until needed. Protein levels were determined on SDS/NaOH (0.1%/0.1 m) solubilized membranes using the Micro BCA protein assay (Pierce, Rockford, IL); bovine serum albumin was the standard.
Binding assay. Binding assays were performed on 100 μg of membrane suspension in 0.25 ml of HBSS containing 0.1% BSA with [125I]α-BGT (20 nm) or [3H]epibatidine (10 nm). Nonspecific binding was determined by addition of cold α-BGT (2 μm) or cold epibatidine (100 μm) before the addition of the corresponding radioactive ligand. The samples were incubated on a rocker platform at 4°C for 90 min and subsequently transferred to glass microfiber filters that had been soaked for at least 3 hr in HBSS containing 0.5% polyethyleneimine. The filters were then washed with cold HBSS (20 ml) and placed in tubes for counting. Each sample was assayed in triplicate.
Quantification of dendritic arborizations
In most experiments, biocytin labeling was used to identify morphologically the neurons from which recordings were obtained (Alkondon and Albuquerque, 2001). The Neurolucida software was used to measure the length of dendrites of neurons from which electrophysiological recordings were obtained.
Drugs and toxins used
ACh chloride, dl-2-amino-5-phosphonovaleric acid (APV), (-)bicuculline methiodide, choline chloride, tetrodotoxin (TTX), QX-314, atropine sulfate, and α-BGT were obtained from Sigma (St. Louis, MO). 6-Cyano-7-nitroquinoxaline-2,3-dione (CNQX) was purchased from Research Biochemicals (Natick, MA). MLA citrate was a gift from Professor M. H. Benn (Department of Chemistry, University of Calgary, Alberta, Canada). Dihydro-β-erythroidine.HBr and (±)mecamylamine.HCl were gifts from Merck, Sharp & Dohme (Rahway, NJ). Stock solutions of all drugs, except KYNA, were made in distilled water. Stock solution of KYNA was made in dimethylsulfoxide (DMSO). [3H]Kynurenine (specific activity, 9.3 Ci/mmol) was purchased from Amersham Biosciences (Arlington Heights, IL). [125I]α-BGT (specific activity, 143 Ci/mmol) and [3H]epibatidine (specific activity, 56.2 Ci/mmol) were purchased from PerkinElmer Life Sciences (Boston, MA). The rabbit anti-human KAT II antibody was kindly provided to us by Dr. Etsuo Okuno (Wakayama Medical College, Wakayama, Japan).
Genotypic and phenotypic characterization of mKat-2-/- mice
Targeted deletion of the mKat-2 gene (Fig. 1A) was verified by PCR and Northern analyses. The results indicated that mKat-2 expression was disrupted in the mKat-2-/- mice (Fig. 1B,C). In addition, immunotitration of enzymatic activity revealed that KAT II activity was absent in hippocampal extracts of mKat-2-/- mice (Fig. 1D). It should be noted that, in the adult rat brain, KAT II accounts for ∼70% of total KAT activity (Guidetti et al., 1997). In the hippocampus of adult mice, however, KAT II activity represents only 10% of total KAT activity (Fig. 1D). Mouse brain KAT II activity was found to be substantially more important during the first month of postnatal development, contributing ∼40% to total KAT activity (Fig. 1D). This ontogenetic pattern of enzyme activity translated into ∼55% reduction of hippocampal KYNA levels in 21-d-old mKat-2-/- mice compared with WT mice, whereas no such decrease was seen in 60-d-old mutant mice (Fig. 2C). Across ages, levels of the KYNA precursor kynurenine and of quinolinic acid, another major kynurenine metabolite, were very similar in hippocampal tissue of WT and mKat-2-/- mice (Fig. 2A,B).
Weight, food, and water intake were not significantly different between mKat-2-/- and WT mice at any given age (data not shown). In contrast, up to 28 d of age, mKat-2-/- mice showed more spontaneous locomotor activity than age- and strain-matched WT mice. Quantification of the locomotor activity in an open-field arena revealed that 21-d-old mKat-2-/- mice crossed the sectors in the arena twice as frequently as WT mice (Fig. 2 D). At 60 d of age, however, the frequency of sector crossings for mKat-2-/- and WT mice was not significantly different (Fig. 2 D).
Characterization of nicotinic responses recorded from CA1 SR interneurons in hippocampal slices of WT and mKat-2-/- mice
Different types of nicotinic responses could be recorded under whole-cell patch-clamp configuration from CA1 SR interneurons in hippocampal slices of WT and mKat-2-/- mice. In the presence of the Na+-channel blocker TTX (300 nM), 87% of the interneurons studied (n = 26 from 30 animals) responded to choline (10 mM) with whole-cell currents that decayed to the baseline during the agonist pulse. Considering that these currents were irreversibly blocked by α-BGT (100 nM) (Fig. 3A) and reversibly blocked by MLA (20 nM) (Fig. 3B), they should have resulted from activation of somatodendritic α7* nAChRs. In slices obtained from the hippocampi of 21-d-old mice, these currents had small amplitudes; the amplitudes of the largest responses recorded at -68 mV from interneurons in 21-d-old WT and mKat-2-/- mice were 16.9 and 26 pA, respectively. Nevertheless, α7* nAChR activation in the somatodendritic region of the CA1 SR interneurons was capable of modulating their excitability as demonstrated by the finding that, in Mg2+-free ACSF, U-tube application of choline (10 mM) to interneurons under cell-attached configuration elicited fast-current transients (Fig. 3C, top trace and inset) that were reversibly blocked by perfusion of the slices with TTX (100 nM) (Fig. 3C) and, therefore, represented action potentials. In the same neurons, ACh, at a subsaturating concentration for α7* nAChRs (0.1 mM), triggered action potentials that had lower frequency and longer delay than those evoked by the saturating concentration of choline (Fig. 3C). The concentration-response relationship for ACh and the finding that both ACh- and choline-triggered action potentials did not outlast the duration of the agonist pulse supported the notion that the action potentials resulted from activation of somatodendritic α7* nAChRs in the interneurons.
A small population of the CA1 SR interneurons studied at -68 mV in the presence of TTX (3 of 16 neurons from 9 WT animals) responded to ACh (up to 1 mM) with slowly desensitizing currents (Fig. 3D). At a concentration as low as 1 μM, ACh could evoke these currents (Fig. 3D). The apparent high potency of ACh suggested that the receptors subserving these responses were most likely α4β2* nAChRs.
In the absence of TTX, 91% of the interneurons studied in 21-d-old WT and mKat-2-/- mice (n = 32 from 35 animals) responded to ACh with EPSCs that were recorded under the whole-cell mode of the patch-clamp technique (Fig. 4). In interneurons that were voltage clamped at -68 mV and continuously perfused with ACSF containing 1 mM extracellular Mg2+ and 10 μM bicuculline, ACh triggered EPSCs that had an average τdecay of 2.8 ± 0.32 msec (n = 12 neurons) (Fig. 4A) and could be blocked by the AMPA receptor antagonist CNQX (10 μM) (Fig. 4A). The pharmacological and kinetic properties of these EPSCs indicated that they were glutamatergic events mediated by AMPA receptors (Arancio et al., 1994; Alkondon et al., 2003a). At +40 mV, however, ACh-elicited EPSCs had an average τdecay of 206 ± 26 msec (n = 13 neurons) (Fig. 4B); >95% of these currents could be blocked by the NMDA receptor antagonist APV (50 μM) (Fig. 4B). The pharmacological and kinetic properties of the majority of ACh-evoked EPSCs recorded at +40 mV were characteristic of glutamatergic events mediated by NMDA receptors (Alkondon et al., 2003a). In the nominal absence of extracellular Mg2+, spontaneously occurring and ACh-triggered EPSCs recorded from CA1 interneurons at any membrane potential were mediated primarily by NMDA receptors, because they could be blocked by APV (50 μM; data not shown). The ability of ACh to trigger NMDA EPSCs in the presence (Fig. 4C) or in the nominal absence of extracellular Mg2+ was inhibited by mecamylamine. Measurements of the net charge of EPSCs triggered by ACh in the presence and in the absence of 1 μM mecamylamine revealed that this antagonist reduced by 89 ± 4.4% the ACh response (n = 6 neurons). Previous studies have demonstrated that 0.3 and 1 μM mecamylamine reduced by ∼50 and 80%, respectively, the activity of native α3β4* nAChRs or ectopically expressed α3β4 nAChRs (Papke et al., 2001; Alkondon et al., 2003a). Other nAChR subtypes are far less sensitive to blockade by mecamylamine (Papke et al., 2001). Thus, the present results lead to the conclusion that glutamate release onto the CA1 SR interneurons in hippocampal slices from 129SvEv mice is regulated by presynaptically located α3β4* nAChRs and is capable of activating both postsynaptic NMDA and AMPA receptors. Similar results have been obtained from CA1 SR interneurons of rat hippocampal slices (Alkondon and Albuquerque, 2002; Alkondon et al., 2003a).
Responses triggered in CA1 SR interneurons of mouse hippocampal slices by activation of somatodendritic α7* nAChRs, preterminal-presynaptic α3β4* nAChRs, and somatodendritic GABAA and NMDA receptors
The decay phase of α7* nAChR-mediated whole-cell currents triggered by application of choline (10 mM) to CA1 SR interneurons in hippocampal slices of 21-d-old WT and mKat-2-/- mice (Fig. 5A) did not differ significantly; the τdecay of currents recorded at -68 mV from neurons in slices of WT and mKat-2-/- mice were 1511 ± 222 msec (n = 8 neurons) and 1413 ± 234 msec (n = 16 neurons), respectively. However, the amplitude and net charge of the currents recorded from CA1 SR interneurons in slices of 21-d-old WT mice were significantly smaller than those of currents recorded from the interneurons in slices of 21-d-old mKat-2-/- mice (Fig. 5B). Choline-evoked currents in CA1 SR interneurons of WT mice ranged from 2 to 16.9 pA in amplitude and from 2 to 35 pC in net charge (Fig. 5B). In contrast, choline-evoked currents recorded from CA1 SR interneurons of mKat-2-/- mice ranged from 8 to 26 pA in amplitude and from 10 to 46 pC in net charge (Fig. 5B). The α7* nAChR activity in CA1 SR interneurons of 21-d-old mKat-2-/- mice may have increased as a direct and/or indirect result of the decreased hippocampal levels of KYNA. It could also have increased because of an enhancement in the number of surface α7 nAChRs.
To determine whether the number of surface α7* nAChRs changed with the age of the animals and/or as a consequence of the targeted mutation, a filter binding assay using [125I]α-BGT was performed on hippocampal membrane preparations from 21- and 60-d-old WT and mKat-2-/-mice. Across ages and animals, [125I]α-BGT binding to the membrane preparations was comparable (Table 1). These results indicated that the higher α7* nAChR activity observed in CA1 SR interneurons of 21-d-old mKat-2-/- mice may not have arisen from an increased number of surface α7* nAChRs. Similarly, levels of binding of the selective α4β2* nAChR ligand [3H]epibatidine to hippocampal membrane preparations of 21-d-old mKat-2-/- and WT mice were not significantly different (Table 1), suggesting that disruption of mKat-2 expression did not alter the number of surface α4β2* nAChRs in the hippocampus. It should be noted, however, that region- and/or cell-specific changes in nAChR expression could have been missed, because the binding assay was performed on membrane preparations from whole hippocampi.
Quantification of the activity of presynaptically located α3β4* nAChRs in the mouse hippocampus was made possible by the analysis of net charge carried by NMDA EPSCs triggered by ACh in CA1 SR interneurons using nominally Mg2+-free ACSF (Fig. 6A). At +40 and -68 mV, the net charge carried by ACh-triggered NMDA EPSCs in CA1 SR interneurons of 21-d-old WT mice was similar to that recorded from interneurons of age-matched mKat-2-/- mice (Fig. 6A,B). Likewise, the net charge carried by whole-cell currents evoked by application of GABA (20 μM) or NMDA (50 μM) plus glycine (10 μM) to CA1 SR interneurons of 21-d-old WT mice was comparable with that of currents recorded from interneurons of age-matched mKat-2-/- mice (Fig. 6C). We therefore concluded that the activities of NMDA and GABAA receptors in CA1 SR interneurons and of α3β4* nAChRs regulating glutamate release onto these interneurons were not affected by the targeted mutation.
Synaptic transmission in the CA1 field of the hippocampi of WT and mKat-2-/- mice
As mentioned above, in the presence of 1 mM extracellular Mg2+ and 10 μM bicuculline, EPSCs recorded from CA1 SR interneurons at -68 mV were primarily mediated by AMPA receptors. The kinetics of AMPA EPSCs recorded from CA1 SR interneurons of the hippocampi of 21-d-old WT and mKat-2-/- mice were very similar (Table 2). Likewise, the frequency and peak amplitude of AMPA EPSCs recorded from CA1 SR interneurons were not affected by the targeted mutation (Table 2).
Recordings obtained from CA1 pyramidal neurons at 0 mV, in the absence of bicuculline, contained primarily outward events (Fig. 7A,B) that were GABAergic in nature, because they could be blocked by the GABAA receptor antagonist picrotoxin (100 μM). In addition, these events had an average τdecay of 48 msec (Table 3), which is similar to that reported for GABAergic IPSCs recorded from rat hippocampal neurons (Alkondon et al., 1999; Alkondon and Albuquerque, 2001). The τdecay of IPSCs recorded from CA1 pyramidal neurons of 21-d-old WT mice were comparable with those of IPSCs recorded from CA1 pyramidal neurons of age-matched mKat-2-/- mice (Table 3). However, the frequency and amplitudes of IPSCs recorded from CA1 pyramidal neurons were significantly higher in mKat-2-/- than in WT mice (Fig. 7A,B; Table 3).
To determine how GABAergic transmission was affected by the targeted mutation, histograms and cumulative plots of IPSC amplitude were analyzed. To minimize contamination of IPSCs with baseline noise, the threshold for event detection was set to be 5 pA (approximately four times the root mean square noise level). The distributions of IPSC amplitudes recorded from CA1 pyramidal neurons of both 21-d-old WT and mKat-2-/- mice were similarly skewed toward larger events (Fig. 8A). Higher frequencies and amplitudes of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- compared with WT mice were demonstrated quantitatively in cumulative probability plots of interevent intervals and IPSC amplitudes (Fig. 8B). There was a significant shift toward larger amplitudes in the cumulative probability of IPSC amplitudes recorded from mKat-2-/- neurons compared with WT neurons (Fig. 8B). There was also a highly significant shift toward shorter interevent intervals in the cumulative probability of interevent intervals recorded from mKat-2-/- neurons compared with WT neurons (Fig. 8B).
At 60 d of age, when levels of KYNA in the brain of mKat-2-/- and WT mice were comparable, the frequencies and amplitudes of IPSCs recorded from CA1 pyramidal neurons in hippocampal slices of either group of animals were very similar. The frequencies of IPSCs recorded from 13 neurons of 60-d-old WT and mKat-2-/- mice were 1.70 ± 0.26 and 1.91 ± 0.33 Hz (mean ± SEM), respectively. Thus, no significant differences were observed in the cumulative probability plots of interevent intervals or IPSC amplitudes recorded from 60-d-old WT and mKat-2-/- mice (Fig. 8C).
Effects of the NMDA receptor antagonist APV, the α7 nAChR antagonist α-BGT, and KYNA on the frequencies and amplitudes of spontaneous IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- and WT mice
To investigate the contribution of NMDA receptors to modulation of synaptic transmission between GABAergic neurons and CA1 pyramidal neurons in WT and mKat-2-/- mice, IPSCs were recorded from CA1 pyramidal neurons in either the presence or absence of the NMDA receptor antagonist APV (100 μM). Cumulative probability plots of interevent intervals and amplitudesof IPSCs recorded in the absence and in the presence of APV from pyramidal neurons of WT mice were not significantly different (Fig. 9A). The higher frequencies and amplitudes of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- mice could not be accounted for by increased NMDA receptor activity resulting directly and/or indirectly from decreased levels of KYNA in the brain of these mice, because cumulative probability plots of interevent intervals and amplitudes of IPSCs recorded in the absence and in the presence of APV from the CA1 pyramidal neurons in hippocampal slices of mKat-2-/- mice were also superimposable (Fig. 9B).
No significant difference existed between cumulative probability plots of interevent intervals and amplitudes of IPSCs recorded from CA1 pyramidal neurons of WT mice in the presence or in the absence of the α7* nAChR antagonist α-BGT (100 nM) (Fig. 9C). In the absence and in the presence of α-BGT (100 nM), the average IPSC frequencies recorded from neurons of four WT mice were 1.9 ± 0.28 Hz (n = 12; mean ± SEM) and 1.7 ± 0.46 Hz (n = 7; mean ± SEM), respectively. In contrast, a 10 min exposure to α-BGT (100 nM) reduced the frequencies and amplitudes of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- mice, causing a significant shift to the right in the cumulative plot of interevent intervals and a significant shift to the left in the cumulative plots of IPSC amplitudes (Fig. 9D). In the absence of α-BGT, the average frequency and amplitudes of IPSCs recorded from CA1 pyramidal neurons of three mKat-2-/- mice were 3.2 ± 0.44 Hz (n = 12 neurons; mean ± SEM) and 12.1 ± 0.26 pA (n = 5608 events; mean ± SEM), respectively. After 10 min of perfusion of the hippocampal slices with ACSF containing α-BGT (100 nM), the average frequency and amplitudes of IPSCs decreased to 1.6 ± 0.32 Hz (n = 9 neurons; mean ± SEM) and 9.7 ± 0.12 pA (n = 3013 events; mean ± SEM), respectively. Results obtained in the presence of α-BGT were significantly different from those obtained in the absence of the toxin (Fig. 9D), becoming similar to those obtained from CA1 pyramidal neurons of age-matched WT mice.
The significant reduction of IPSC frequency and amplitude caused by α-BGT could be accounted for by a presynaptic reduction of GABA release attributable to blockade of α7* nAChRs in GABAergic neurons synapsing onto the neurons under study or a decrease in the activity of postsynaptic GABAA receptors. The presynaptic action of α-BGT is favored, because numerous studies have reported that this toxin selectively blocks nAChRs (for review, see Albuquerque et al., 1997). The proximal dendritic region of CA1 pyramidal neurons is highly innervated by axons of CA1 SR interneurons (Gulyás and Freund, 1996; Oliva et al., 2000; Buzaski et al., 2003). Thus, it can be concluded that GABA release onto CA1 pyramidal neurons in 21-d-old mKat-2-/- mice is higher than in age-matched WT mice because the α7* nAChR activity in CA1 SR interneurons of the mutant mice is higher than in the interneurons of the WT mice. Similar results were obtained when hippocampal slices of 21-d-old WT or mKat-2-/- mice were exposed in vitro to 100 nM KYNA (Fig. 9 E, F).
Because of the slow equilibration of KYNA with α7* nAChRs (Hilmas et al., 2001) and the physical barrier imposed by the slice preparation, hippocampal slices were allowed to equilibrate in situ for 1 hr with 100 nM KYNA before IPSCs were recorded from the pyramidal neurons. The cumulative distributions of interevent intervals and amplitudes of IPSCs recorded from CA1 pyramidal neurons of WT mice were unaffected by in situ exposure to KYNA (Fig. 9E). In hippocampal slices that were perfused for 1 hr with ACSF containing DMSO in the same dilution as that present in the KYNA-containing ACSF (i.e., 1:100,000), the average frequency of IPSCs recorded from 12 CA1 pyramidal neurons of three WT mice was 1.8 ± 0.25 Hz (mean ± SEM). Likewise, in hippocampal slices that were perfused for 1 hr with KYNA (100 nM)-containing ACSF, the average frequency of IPSCs recorded from 10 CA1 pyramidal neurons of two other WT mice was 1.8 ± 0.41 Hz (mean ± SEM). In contrast, the cumulative distribution of interevent intervals of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- mice was significantly altered by the in situ exposure to KYNA (Fig. 9F) When hippocampal slices from mKat-2-/- mice were perfused with ACSF containing DMSO (1:100,000) or KYNA (100 nM), the average frequencies of IPSCs recorded from CA1 pyramidal neurons were 3.9 ± 0.47 Hz (mean ± SEM; n = 8 neurons from two mKat-2-/- mice) and 2.9 ± 0.29 Hz (mean ± SEM; n = 8 neurons from two mKat-2-/- mice), respectively. Thus, at 100 nM, KYNA reduced by ∼25% the frequency of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- mice. Changes in the amplitude of IPSCs by in situ exposure of the mKat-2-/- hippocampal slices to KYNA (100 nM) were very small (Fig. 9F). The average amplitudes of IPSCs recorded in the absence and in the presence of 100 nM KYNA were 12.6 ± 0.21 pA (n = 12,260 events) and 11.7 ± 0.22 pA (n = 8536 events), respectively.
KYNA-induced reduction in the frequency of IPSCs recorded from CA1 pyramidal neurons of mKat-2-/- mice was the result of its interaction with α7* nAChRs because, after a 10 min perfusion of the hippocampal slices with ACSF-containing α-BGT (100 nM), an additional 1 hr incubation of the slices with ACSF containing KYNA (100 nM) plus α-BGT (100 nM) caused no additional changes in the frequency of IPSCs. In hippocampal slices from mKat-2-/- mice, the frequency of IPSCs recorded in the presence of KYNA plus α-BGT from 10 CA1 pyramidal neurons was 101.3 ± 5.01% (mean ± SEM) of that recorded in the presence of α-BGT alone from six CA1 pyramidal neurons.
Correlation between dendritic length and α7* nAChR activity evoked by exogenous application of choline to CA1 SR interneurons in 21-d-old WT and mKat-2-/- mice
Previous studies have shown that α7* nAChR density increases along with dendritic length in hippocampal neurons in culture (Albuquerque et al., 1997; Zarei et al., 1999) and decreases along the dendrites of hippocampal CA1 interneurons in slices (Khiroug et al., 2003). To determine whether increased α7* nAChR activity in CA1 SR interneurons of mKat-2-/- mice was accompanied by changes in dendritic length in these neurons, two approaches were taken. First, the Neurolucida software was used to measure the dendritic length of CA1 SR interneurons from which recordings were obtained. Second, correlational analyses were performed to determine whether dendritic length was correlated with the magnitude of α7* nAChR activity induced by U-tube application of choline to the interneurons.
After the electrophysiological recordings were obtained, reconstruction of the image of biocytin-filled interneurons revealed that the morphological features of the neurons studied in hippocampal slices of WT mice were very similar to those of the neurons studied in slices of mKat-2-/- mice (Fig. 10A). All interneurons studied had their cell bodies located in the CA1 SR. Although the axons of all successfully labeled interneurons branched mostly within the CA1 SR, some projected into the stratum oriens (n = 12 of 24 neurons in WT mice and 6 of 20 neurons in mKat-2-/- mice) and others projected into the stratum lacunosum-moleculare (n = 12 of 24 neurons in WT and 14 of 20 neurons in mKat-2-/- mice). In contrast, the dendrites of the CA1 SR interneurons studied in hippocampal slices of WT and mKat-2-/- mice were confined primarily to the CA1 SR (Fig. 10A). No significant differences were observed between the length of dendrites of CA1 SR interneurons of 21-d-old WT and mKat-2-/- mice (Fig. 10A,B), and the amplitudes of choline-evoked whole-cell currents recorded from these interneurons showed no correlation with dendritic length (Fig. 10C). Therefore, it can be concluded that the increased α7* nAChR activity detected in CA1 SR interneurons of mKat-2-/- mice neither depend on nor caused significant changes in the dendritic length of these neurons.
In this study, the use of mKat-2-/- mice revealed that physiological levels of the tryptophan metabolite KYNA, although insufficient to modulate NMDA receptor activity, are sufficient to lower α7* nAChR activity in CA1 SR interneurons and, consequently, to attenuate GABAergic synaptic activity impinging onto CA1 pyramidal neurons in the hippocampus.
The present findings challenge the assumption that the biological effects of KYNA result exclusively from its action as a competitive antagonist of glycine at NMDA receptors (Stone, 1993; Danysz and Parsons, 1998; Schwarcz and Pellicciari, 2002). This mechanism of action for KYNA, which was originally proposed in the late 1980s (Mayer et al., 1988) (for review, see Danysz and Parsons, 1998), has been indirectly supported by evidence that the glycine site of NMDA receptors is not saturated in vivo (Salt, 1989; Thomson et al., 1989; Bergeron et al., 1998; Kinney et al., 2003) and that, in the nominal absence of glycine, KYNA blocks NMDA receptors with an IC50 of ∼15 μM (Henderson et al., 1990; Parsons et al., 1997; Hilmas et al., 2001). However, the apparent potency for KYNA to block NMDA receptors decreases sharply with increasing glycine concentrations. In the presence of glycine concentrations similar to those found extracellularly in the CNS (i.e., 5-10 μM) (Kennedy et al., 2002), KYNA blocks NMDA receptors with an IC50 of ∼250 μM. This concentration is unlikely to be reached in the brain in vivo (Scharfman et al., 1999; Frankiewicz et al., 2000; Urenjak and Obrenovitch, 2000; Hilmas et al., 2001).
α7* nAChRs are major targets for the actions of KYNA in the brain
As observed in the rat hippocampus (Alkondon and Albuquerque, 2001; Adams et al., 2002), in the mouse hippocampus, the majority of CA1 SR interneurons (>80% of the sampled interneurons) express somatodendritic α7* nAChRs, and α7* nAChR expression reaches adult levels by the third postnatal week. However, in CA1 SR interneurons of 21-d-old WT mice, the amplitudes of α7* nAChR-mediated whole-cell currents are considerably smaller than those recorded from CA1 SR interneurons in hippocampal slices of age-matched rats (Alkondon et al., 1999). This could be explained by differences in the number of surface α7* nAChRs in hippocampal interneurons between the two species. In fact, such differences have been reported even among different mouse strains (Marks et al., 1996; Gahring et al., 2004).
Analysis of amplitude and net charge of choline-evoked currents indicated that α7* nAChR activity in CA1 SR interneurons was ∼65% higher in 21-d-old mKat-2-/- mice than in age-matched WT mice. Binding studies performed in membrane hippocampal preparations suggested that this increased α7* nAChR activity may not have resulted from changes in numbers of surface α7* nAChRs. Although an indirect effect cannot be completely ruled out, the increased α7* nAChR activity in the interneurons of mKat-2-/- mice can be accounted for solely by a reduction in nAChR blockade by endogenous KYNA (Hilmas et al., 2001).
The effect of the targeted mutation appeared to be quite specific for α7* nAChRs. Decreased hippocampal KYNA levels in 21-d-old mKat-2-/- mice did not lead to significant changes in NMDA receptor activity in the CA1 SR interneurons, as indicated by measurements of whole-cell currents evoked by exogenously applied NMDA. In addition, ongoing endogenous NMDA receptor activity had no detectable contribution to regulation of GABAergic synaptic activity impinging onto CA1 pyramidal neurons of both WT and mutant mice.
In the mouse hippocampus, like in the rat hippocampus (Alkondon et al., 2003a), stimulation of α3β4* nAChRs in glutamatergic neurons enhanced glutamatergic input to CA1 SR interneurons. The activity of postsynaptic AMPA receptors on CA1 SR interneurons, the spontaneous release of glutamate onto the interneurons, and the activity of presynaptically located α3β4* nAChRs were not altered by the lower hippocampal levels of KYNA resulting from the targeted deletion of the mKat-2 gene. Endogenous levels of KYNA were also insufficient to modulate α4β2* nAChR expression in the hippocampus. Although prolonged in vitro exposure of hippocampal neurons to 10 μM KYNA causes upregulation of functional α4β2* nAChRs (Hilmas et al.,2001; Pereira et al., 2002), the number of [3H]epibatidine binding sites in hippocampal membranes of 21-d-old WT and mutant mice were comparable (Table 1). Therefore, endogenous levels of KYNA in the hippocampus selectively modulate α7* nAChR activity.
GABAergic synaptic transmission is selectively increased in the hippocampus of 21-d-old mKat-2-/- mice: relationship with higher α7* nAChR activity in CA1 interneurons, lower hippocampal levels of KYNA, and increased locomotor activity
In the mouse and rat hippocampus, GABA release is regulated by α7* nAChRs (Alkondon and Albuquerque, 2001; Dobelis et al., 2003). In the rat hippocampus, agonist-induced activation of α7* nAChRs in CA1 SR interneurons, which are known to synapse onto proximal dendrites of CA1 pyramidal neurons (Gulyás and Freund, 1996; Oliva et al., 2000; Buzaski et al., 2003), enhances GABAergic synaptic activity impinging onto the pyramidal neurons (Alkondon and Albuquerque, 2001). Thus, it can be hypothesized that a higher endogenous activity of α7* nAChRs in CA1 interneurons accounts for the increased GABAergic activity impinging onto the CA1 pyramidal neurons of the 21-d-old mKat-2-/- mice. This hypothesis is supported by the finding that the α7 nAChR antagonist α-BGT brought the IPSC frequency and amplitude recorded from CA1 pyramidal neurons of 21-d-old mKat-2-/- mice to values similar to those recorded from age-matched WT mice. Our hypothesis would also predict that adding a background level of KYNA to the hippocampal slices of mKat-2-/- mice should reduce the endogenous activity of α7* nAChRs in the interneurons and, consequently, decrease the GABAergic activity impinging onto the pyramidal neurons. In fact, in vitro exposure of hippocampal slices from mKat-2-/- mice to 100 nM KYNA significantly decreased the frequency of IPSCs recorded from CA1 pyramidal neurons, and this effect could not be observed after blockade of the α7* nAChRs by α-BGT.
In 21-d-old mKat-2-/- mice, the reduction of hippocampal KYNA levels was accompanied by a readily observable phenotype, i.e., enhanced spontaneous locomotor activity, which can be a result of the increased GABAergic activity in the hippocampus. There are reports that locomotor activity increases in DBA2 mice and rats after intrahippocampal microinjection of the GABAA receptor agonist muscimol, the GABA degradation inhibitor γ-vinyl-GABA, or GABA itself (van Abeelen and Boersma, 1984; Alvarez and Banzan, 1990). At 60 d of age, hippocampal KYNA levels and locomotor activity of mKat-2-/- mice converged to values comparable with those measured in WT mice. Thus, it is plausible to hypothesize that, as KYNA levels increase in the hippocampus of the 60-d-old mice, α7* nAChR activity in the CA1 SR interneurons decreases and consequently reduces the GABAergic transmission between these neurons and the pyramidal neurons.
Interactive signaling between glial cells and neurons is crucial for proper functioning of the hippocampus
Numerous studies have suggested that a network of interactive neurons and glial cells controls brain function (Boehning and Snyder, 2003). Astrocytes, which frequently engulf synapses with their processes and release a host of neuroactive compounds, are important players in the governance of these interactions, especially as they relate to excitatory neurotransmission (Bezzi et al., 2001). For example, astrocytes produce and release the amino acid D-serine, which increases NMDA receptor activity in neurons by acting as an agonist at the glycine site of these receptors (Snyder and Kim, 2000). Similarly, the enzyme glutamate carboxypeptidase II is expressed on the cytoplasmic membrane of astrocytes and influences neuronal glutamatergic activity by hydrolyzing the neuropeptide N-acetylaspartylglutamate, which is primarily stored in and released from CNS glutamatergic and cholinergic terminals (Wroblewska et al., 1993; Coyle, 1997).
The present demonstration that reduction of physiological levels of KYNA, which is preferentially produced by and released from astrocytes in the brain (Roberts et al., 1992; Guillemin et al., 2000, 2001; Kiss et al., 2003), is associated with a specific phenotype and regulates α7* nAChR activity in interneurons as well as synaptic transmission between these neurons and pyramidal cells in the hippocampus lends support to the concept that glial neuronal signaling is essential for proper brain functioning. In particular, KYNA released by astrocytes closely apposed to GABAergic axons synapsing onto pyramidal neurons would interact with α7* nAChRs in the GABAergic axons. Regulation by endogenous KYNA of the α7* nAChR activity in the GABAergic axons could then determine the level of GABAergic synaptic activity impinging onto the pyramidal neurons and, indirectly, regulate the excitability of these neurons. The interplay between glia-derived KYNA and the nicotinic cholinergic system could be essential in the pathogenesis of numerous catastrophic brain disorders, including schizophrenia and Alzheimer's disease, in which brain KYNA levels are elevated, and nicotinic and GABAergic functions are impaired (Tamminga, 1998; Baran et al., 1999; Court et al., 1999; Bartus, 2000; Schwarcz et al., 2001).
↵ a According to the current status of the nomenclature for nAChRs and their subunits (Lukas et al., 1999), the asterisk next to nAChR subunits throughout text is meant to indicate that the exact receptor subunit composition is not known.
This work was supported in part by United States Public Health Service Grants NS25296 and NS41671. We are indebted to Mabel Zelle for her expert technical assistance and Bhagavathy Alkondon for her technical support.
Correspondence should be addressed to Dr. Edson X. Albuquerque, Department of Pharmacology and Experimental Therapeutics, University of Maryland School of Medicine, 655 West Baltimore Street, Baltimore, MD 21201. E-mail:.
D. A. Tagle's present address: Neurogenetics, National Institute of Neurological Disorders and Stroke, National Institutes of Health, 6001 Executive Boulevard, Bethesda, MD 20892.
P. Yu's present address: Structure Biophysics Laboratory, National Cancer Institute, National Institutes of Health, Frederick, MD 21702.
Copyright © 2004 Society for Neuroscience 0270-6474/04/244635-14$15.00/0
↵* M.A., E.F.R.P., P.Y., and E.Z.A. contributed equally to this work.