Prolonged Exposure to NMDAR Antagonist Suppresses Inhibitory Synaptic Transmission in Prefrontal Cortex

Abstract

Postmortem studies have shown that schizophrenia produces a reduction in the 67-kilodalton isoform of glutamic acid decarboxylase (GAD67), a key enzyme for γ-aminobutyric acid (GABA) synthesis. N-methyl-d-aspartate receptor (NMDAR) antagonists have been extensively used to study schizophrenia because they can induce many aspects of the disease, including the decrease in GAD67. It is generally thought that this reduction in GAD implies a reduction in functional inhibition, but direct evidence had been lacking. We have therefore performed physiological studies in slices of prefrontal cortex taken from rats treated with the NMDAR antagonist ketamine. Both frequency and amplitude of miniature inhibitory postsynaptic currents were reduced. Consistent with a reduction of inhibition, we observed an increase in postsynaptic excitability. The increased excitability is likely to result from disinhibition because miniature excitatory postsynaptic current properties and intrinsic excitability were not changed. Ketamine did not affect inhibition or GAD levels in young rats, indicating a developmental regulation that may be related to the developmental increase in ketamine sensitivity that occurs in humans. Our results show that NMDAR antagonist produces biochemical changes in the GABA system that lead to a functional disinhibition. Such disinhibition would be expected to decrease gamma oscillations, which are reduced in schizophrenia.

INTRODUCTION

The 67-kilodalton isoform of glutamic acid decarboxylase (GAD67) is critical for the synthesis of the major inhibitory neurotransmitter, γ-aminobutyric acid (GABA) (Bu et al. 1992). Postmortem studies of schizophrenia patients have shown that GAD67 mRNA and protein are substantially reduced in GABAergic interneurons (Akbarian et al. 1995; Benes et al. 2007; Dracheva et al. 2004; Guidotti et al. 2000; Volk et al. 2000). The decrease does not occur in all classes of interneurons, but is observed mainly in parvalbumin (PV)-containing interneurons (Hashimoto et al. 2003). These neurons are characterized by a fast-spiking firing pattern (Kawaguchi and Kondo 2002) and control the excitability of pyramidal neurons (Fuchs et al. 2007; Zhu et al. 2004). Furthermore, GABAergic activity controls spike timing in pyramidal neurons, thereby contributing to the synchronized firing that underlies gamma-frequency oscillations (Fisahn et al. 1998), oscillations that have an important role in cognitive function. There are thus likely to be major consequences of the reduction in GAD67 in schizophrenia if the reduction produces deficits in GABAergic transmission (Lewis and Gonzalez-Burgos 2006).

It has not been possible to study directly the functional status of inhibiton in schizophrenia patients and it is therefore of interest to examine inhibition in animal models of the disease. N-Methyl-d-aspartate receptor (NMDAR) antagonists, such as ketamine, are extensively used as models of schizophrenia because they induce symptoms in healthy volunteers that are very similar to those of schizophrenia (Krystal et al. 1994; Newcomer et al. 1999). Furthermore, neurochemical changes observed in schizophrenia are reproduced by NMDAR antagonists. Importantly, several days of in vivo ketamine treatments produce a reduction in GAD67 in PV interneurons (Behrens et al. 2007; Kinney et al. 2006). To determine whether this treatment has functional effects on inhibition, we have characterized the strength of inhibition in slices of prefrontal cortex after 2 days' treatment of rats with the NMDA antagonist ketamine.

METHODS

Animal model and brain slices preparation

Male Long–Evans rats (3 mo old, 300–380 g; Charles River, Wilmington, MA) were housed under a 12-h light/dark cycle in a temperature- and humidity-controlled environment with free access to food and water. Animals were handled for 2 days prior to experiments to reduce stress during treatments. Animals were given an intraperitoneal injection of 30 mg/kg ketamine (ketamine hydrochloride, dissolved in physiological saline; Sigma, St. Louis, MO) daily on 2 consecutive days. The efficacy of this dose was established in a previous study (Keilhoff et al. 2004). Animals from the control group received saline injections. The animals were killed on the day following the second series of injections. The brains were rapidly removed and were cut into 300-μm-thick coronal slices with a vibrotome (Leica VT 1000S, Nussloch, Germany) in an oxygenated ice-cold solution containing (in mM): NaCl 124, KCl 2.5, NaCO₃ 26, NaH₂PO₄ 1.25, dextrose 10, CaCl₂ 2.5, and MgSO₄ 4. Slices containing prefrontal cortex were collected and incubated for ≥1 h before being transferred into the chamber for recording.

Electrophysiology

Patch electrodes were prepared from borosilicate glass using a horizontal electrode puller (P-87, Flaming/Brown; Sutter Instrument, Novato, CA) to produce tip openings of 1–2 μm (3–5 MΩ). For miniature inhibitory postsynaptic current (mIPSC) recordings, electrodes were filled with an internal solution containing (in mM) CsCl 135, NaCl 6, MgCl 1, HEPES 10, EGTA 2, tetraethylammonium (TEA) chloride 5, QX-314 2, Mg-ATP 2, and CaCl₂ 0.2. For miniature excitatory postsynaptic current (mEPSC) recordings, the internal solution contained (in mM) CsCl 43, CsMeSO₄ 92, TEA 5, EGTA 2, MgCl₂ 1, HEPES 10, and adenosine 5′-triphosphate (ATP) 4; pH was adjusted to 7.2–7.4 with CsOH. For current-clamp recording, electrodes were filled with an internal solution containing (in mM): KMeSO₄ 120, KCl 12, MgCl₂ 1, EGTA 2, CaCl₂ 0.2, HEPES 10, Mg-ATP 2, and Na-GTP 0.4; pH was adjusted to 7.2–7.4 with KOH and the final osmolarity was about 290 mOsm. Brain slices were held down in the recording chamber with a staple pin and were immersed in oxygenated artificial cerebrospinal fluid at a flow rate of 2–3 ml/min. Pyramidal neurons in layer 5 of prefrontal cortex were visualized using dark-field illumination and a charge-coupled detector camera. Pyramidal cells were identified by their morphology, a pyramid-like cell body, and a thick apical dendrite. The series resistance of the pipette was about 10 MΩ. Both voltage-clamp and current-clamp recordings were performed with an Axopatch 200B amplifier (Molecular Devices, Foster City, CA). Series resistance was monitored throughout the experiments. The recording was terminated if series resistance varied >15% of the baseline. Signals were filtered at 2 kHz and digitized at a sampling rate of 5 kHz using a data-acquisition program (Igor Pro 5.0, WaveMetrics, Portland, OR). The recording was performed at room temperature (∼22°C). Miniature IPSCs were analyzed with the MiniAnalysis 6.0 program (Synaptosoft, Decatur, GA). All events visually judged as mIPSCs were detected by the MiniAnalysis program. Settings of the MiniAnalysis program were: threshold amplitude, 8 pA; area threshold, 50 pA/ ms; period to search a peak, 20 ms; and period to average a baseline, 10 ms.

Immunohistochemistry

Slices used for electrophysiology were fixed in 4% paraformaldehyde for 24 h, washed in phosphate-buffered saline (PBS), and then equilibrated in 30% sucrose in PBS for 24 h at 4°C. Slices were then mounted in OCT (Optimal Cutting Temperature embedding medium), frozen, and cut coronally at 40-μm thickness in a cryostat. Outermost slices were discarded and the rest were processed for floating-section double immunohistochemistry for the detection of PV and GAD67. When detecting GAD67, antigen retrieval was performed by incubation of the slices in 1% sodium borohydride for 15 min as described (Stanley and Shetty 2004), followed by washing in PBS and incubation in 10% normal goat serum in PBS for 16 h at 4°C. Primary antibodies (PV: 1:3,000 rabbit polyclonal; GAD67: 1:2,000 mouse monoclonal) were diluted in 2% normal goat serum in PBS and applied to the slices for 18 h at 4°C. Following several washes in PBS the slices were incubated in a 1:1,000 dilution of AlexaFluor conjugated goat anti-rabbit (568) and goat anti-mouse (488) antibodies for 1 h at room temperature. Slices were washed in PBS and mounted sequentially on glass slides using Vectashield, covered with a coverslip, and allowed to dry for ≥24 h before confocal imaging.

Confocal microscopy and image analysis

Mounted slices were evaluated for fluorescence under settings for 568 and 488 emissions on a LSM510 Meta multiphoton laser confocal microscope using a ×40 water-immersion objective. Each slice was then imaged across the prefrontal cortex (three regions per slices). For each slice a z-stack of eight images was obtained. Cell numbers were measured in a volume of 230.3 × 230.3 × 3.5 μm³. All PV neurons in the images were analyzed for their parvalbumin and GAD67 fluorescence intensity using Metamorph as described (Behrens et al. 2007). The number of PV cells per slice was counted using Abercrombie's correction (Abercrombie 1946).

Statistics

The data are presented as means ± SE. The two-tailed unpaired t-test was used for two-group comparisons, except for the comparison of mIPSC amplitudes in which the Kolmogorov–Smirnov test was used. ANOVA followed by Tukey's test was used for multigroup comparisons for fluorescence intensity analysis. The difference was considered significant when P < 0.05.

RESULTS

Rats were injected with ketamine for 2 days. Brain slices were then prepared and used for electrophysiological analysis (see following text). The same slices were examined for effects on GAD67 and PV. We used PV/GAD67 double-immunocytochemical staining, as recently described in mouse studies (Behrens et al. 2007). As shown in Fig. 1, A and B, this treatment produced a 31% reduction of GAD67 fluorescence in the cell bodies of PV interneurons (P < 0.001, n = 12). Ketamine treatment also reduced the PV fluorescence (58% of control, Fig. 1, A and B). There was no substantial reduction in the number of PV staining neurons [7.5 ± 2.1 in the control group (saline injected); 8.6 ± 1.8 in the ketamine group, P = 0.15]. These results demonstrate that PV-containing interneurons in rat contain less GAD67 after 2 days of ketamine treatment.

FIG. 1.

In vivo 2-day ketamine treatment decreased the 67-kDa isoform of glutamic acid decarboxylase (GAD67) and parvalbumin (PV) immunoreactivity in PV interneurons in the brain slices used for electrophysiological studies. A: pictures show PV (red) and GAD67 (green) double immunostaining in slices from saline-injected animals. B: ketamine decreased GAD67 from 71.1 ± 3.6 to 49.1 ± 3.3. Ketamine decreased PV from 121.6 ± 3.8 to 70.9 ± 17.0. Rats used were 3 mo old.

To determine whether inhibition is affected in such slices, we recorded mIPSCs. 6-Cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 μM), d-2-amino-phosphonopentanoic acid (d-AP5, 50 μM), and tetrodotoxin (TTX, 0.5 μM) were bath-applied to block fast excitatory synaptic transmission and spontaneous action potentials. The holding potential was −70 mV. High chloride concentration in the internal solution made mIPSCs inward. The recorded events could be fully blocked by bicuculline (20 μM), confirming that they were GABA_A-receptor–mediated currents (data not shown). In the saline-injected rats, the mean amplitude of mIPSCs was 37.99 ± 1.44 pA (n = 30), whereas in the ketamine-injected rats, the averaged amplitude of mIPSCs was decreased by about 20% compared with the saline group (30.80 ± 1.10 pA, n = 27, P < 0.001, Fig. 2, A and B). The kinetics of mIPSCs were not significantly changed by ketamine treatment (Fig. 2D, inset).

FIG. 2.

Ketamine decreased both frequency and amplitude of miniature inhibitory postsynaptic currents (mIPSCs) in 3-mo-old rats, but had no effect in 1-mo-old rats. A: representative traces from the saline (top) group and the ketamine group (bottom) in 3-mo-old rats. B and C: group data show a decrease in amplitude and frequency of mIPSCs after ketamine treatment in 3-mo-old rats. D: histogram of mIPSC distributions in the control (saline) and ketamine groups. E: representative traces from the saline (top) group and the ketamine group (bottom) in 1-mo-old rats. F and G: group data show no change in the amplitude or frequency of mIPSCs after ketamine treatment in 1-mo-old rats.

The frequency of mIPSCs was also reduced by ketamine treatment. In the saline-injected rats, mIPSCs had a mean frequency of 2.65 ± 0.16 Hz, averaged over a 10-min recording (n = 30). In the ketamine-treated rats, the frequency was lower by about 37% (1.67 ± 0.15 Hz in the ketamine group, n = 27, P < 0.001, Fig. 2, A and C). We also analyzed the distribution of mIPSC amplitude. Ketamine treatment reduced the peak of the histogram from 20 pA in the control group to 14 pA in the ketamine group (Fig. 2D). However, not all classes of mIPSCs were equally decreased as would occur if there was a global shift of mIPSC amplitude. Rather, we found that ketamine had no effect on the frequency of large-amplitude mIPSCs (Fig. 2D).

It was of interest to test whether the effect of ketamine on inhibition occurs in 1-mo-old rats because the ability of ketamine to induce psychosis in humans does not appear until adolescence (White et al. 1982). As shown in Fig. 2, E–G, ketamine had no effect on either frequency or amplitude of mIPSCs in 1-mo-old rats (36.89 ± 1.56 pA and 1.69 ± 0.14 Hz in the saline group, n = 21; 35.83 ± 1.24 pA and 1.65 ± 0.13 Hz in the ketamine group, n = 23; P = 0.85 for frequency comparison, P = 0.60 for amplitude comparison). Changes in inhibition may be due to altered levels of GAD67. Thus it was of interest to test whether the ketamine-induced changes in GAD67 and PV that occur in older animals were absent in younger animals. The results show no difference in GAD67 and PV levels between control and ketamine groups in 1-mo-old rats (GAD67: control, 75.6 ± 6.6, ketamine, 78.8 ± 5.1; PV: control, 167.4 ± 11, ketamine, 159.1 ± 6.1).

PV interneurons form inhibitory synapses surrounding the cell body and axon initial segment and thus can influence the excitability of pyramidal cells (Markram et al. 2004). A reduction in GAD67 would be expected to reduce GABA synthesis and thereby reduce the tonic and phasic concentrations of GABA (Farrant and Nusser 2005) produced by spontaneous release and by the release from spontaneously active fast-spiking interneurons (Tseng and O'Donnell 2004). This disinhibition should enhance pyramidal cell excitability. To test this prediction, we performed current-clamp recordings to measure the firing rate of pyramidal neurons. These experiments were performed with normal chloride concentration in the internal solution (see methods). We delivered a series of constant-current pulses through the recording pipette to induce action potentials. The firing rate was measured during a 1-s depolarizing current pulse. As shown in Fig. 3, we found a significant increase in firing rate in slices from ketamine-treated rats. For example, a 70-pA pulse of 1 s induced 3.1 ± 0.5 spikes in control neurons, whereas it induced 6.0 ± 0.9 spikes in neurons from the ketamine-treated group (P < 0.05, n = 7 in the control group, n = 9 in the ketamine group).

FIG. 3.

Excitability of pyramidal neurons was enhanced after in vivo ketamine treatment. Action potentials were induced with constant-current pulses (1-s duration), ranging from 10 to 100 pA with a step increase of 10 pA. A: representative traces from the saline group (black) and the ketamine group (gray) in response to a 70-pA current pulse. Calibration bars: 200 ms, 20 mV. B: group data show that firing rates during current injections were increased after ketamine treatment.

The increased neuronal excitability could be caused by enhanced excitatory synaptic drive, decreased inhibitory synaptic drive, or increased intrinsic excitability. To test whether ketamine altered properties of excitation, we measured the amplitude and frequency of miniature excitatory postsynaptic currents (mEPSCs). Membrane potential was held at −70 mV. Bicuculline (30 μM) and TTX (0.5 μM) were bath-applied to block fast inhibitory synaptic transmissions and spontaneous action potentials. As shown in Fig. 4, ketamine did not lead to a change in mEPSC frequency or amplitude (control group: 1.22 ± 0.09 Hz and 13.34 ± 0.41 pA, n = 10; ketamine group: 1.20 ± 0.08 Hz and 13.59 ± 0.54 pA, n = 11). Next, we tested whether intrinsic excitability of pyramidal neurons was changed by ketamine treatment. CNQX (20 μM), d-AP5 (50 μM), and bicuculline (30 μM) were bath-applied to block fast synaptic transmissions. Current pulses were delivered to induce spikes. The results are shown in Fig. 5. There was no difference in intrinsic membrane excitability between saline and ketamine groups. Resting membrane potential, input resistance, and spike threshold were also not significantly changed (Fig. 5). These results indicate that the major effect of ketamine treatment on pyramidal cell excitability is likely to be due to disinhibition.

FIG. 4.

Ketamine did not change miniature excitatory postsynaptic current (mEPSC) frequency or amplitude. A: representative traces from the saline (top) group and the ketamine group (bottom) in 3-mo-old rats. B and C: plot shows no change in the cumulative probability of mEPSC amplitude and inter-event intervals. D and E: group data show no change in mEPSC frequency and amplitude.

FIG. 5.

Ketamine did not change intrinsic neuronal excitability of pyramidal cells. A: representative traces from the saline (black) group and the ketamine group (gray) in 3-mo-old rats. B–D: group data show no change in resting membrane potential (RMP), input resistance (IR), and spike threshold. E: group data show no change in intrinsic membrane excitability.

DISCUSSION

Our results provide the first evidence that treatment with an NMDAR antagonist sets into motion biochemical changes in the GABAergic interneurons that lead to a reduction of inhibitory function. We show that this reduction of inhibition produces an enhanced excitability of pyramidal cells. This enhancement is likely to be due to disinhibition because we found no change in mEPSC properties or the intrinsic excitability of pyramidal cells.

Ketamine treatment produced a reduction in both amplitude and frequency of mIPSCs. A reduction in frequency of miniature events could occur because of a failure to release or because amplitude reductions brought events below the detection limit. We thus consider it likely that the main effect of ketamine is to reduce GABA release and thereby to reduce mIPSC amplitude. This process does not appear to occur in all interneurons because the largest mIPSCs were not affected.

Disinhibition has been proposed as a core principle of NMDAR hypofunction in schizophrenia (Coyle 1996; Lisman et al. 2008; Olney et al. 1999). It now seems clear that when NMDAR hypofunction is induced by NMDAR antagonists at least two mechanisms are involved, one that is acute and one that depends on chronic treatment. Under control conditions, NMDARs contribute strongly to EPSPs in some types of interneurons (Lei and McBain 2002; Watanabe et al. 2005). Thus acute block of these receptors will dampen the activity of these interneurons (Homayoun and Moghaddam 2007) and thereby disinhibit pyramidal neurons (Grunze et al. 1996). Consistent with this idea, firing rates of pyramidal neurons are potentiated in prefrontal cortex on acute blockade of NMDARs using (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine maleate (MK-801; Jackson et al. 2004). Similar effects are seen in hippocampal pyramidal cells (MC Quirk, DL Sosulski, CE Feierstein, N Uchida, and ZF Mainen, unpublished observations). Our results indicate that chronic NMDAR antagonism reduces inhibition by a second process that reduces GAD67 and the efficacy of quantal inhibitory output. For a theory as to why this second process occurs see Lisman et al. (2008).

Recent work in mice has shed important light on the transduction events involved in the regulation of GAD67 by ketamine. The downregulation of GAD67 produced by ketamaine is associated with a persistent activation of the superoxide-producing enzyme reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and consequent accumulation of superoxide. Inhibition of the superoxide-synthesizing enzyme prevents the loss of GAD67 immunoreactivity in PV interneurons (Behrens et al. 2007).

The disinhibition produced by ketamine treatment may provide a way of explaining some of the symptoms of schizophrenia (Lisman et al. 2008). An important functional consequence of disinhibition in the hippocampus is to drive increased dopamine release (Legault et al. 2000; Lodge and Grace 2007), which may lead to psychosis. Disinhibition may also affect widespread cortical circuitry and reduce gamma oscillations. Recent work shows that interfering with the action of PV interneurons reduces gamma oscillations and produces cognitive deficits (Fuchs et al. 2007). Disinhibition may thus be an explanation for the reduction in the power of gamma oscillations that occurs in schizophrenia (Cho et al. 2006).

It appears that the GABAergic system is sensitive to modification and can be modified by a variety of factors. GAD67 is reduced in another model of schizophrenia produced by chronic injection of picrotoxin into the basolateral amygdala and there is a reduction of functional inhibition similar to what we have observed (Gisabella et al. 2005). GAD67 can also be reduced by sensory deprivation (Cotrufo et al. 2003; Hendry and Jones 1988). Recent work shows that GAD67 in barrel cortex is negatively modulated after whisker trimming and that this reduction is accompanied by a depression of inhibitory synaptic transmission (both frequency and amplitude of mIPSC are significantly decreased) (Jiao et al. 2006).

A noteworthy finding of the present study is that ketamine treatment failed to depress inhibitory transmission and reduce GAD67 levels in 1-mo-old rats. This may explain why an NMDA antagonist fails to affect heat shock gene activation in 1-mo-old rats, but does so strongly in 3-mo-old rats (Nakki et al. 1996). These findings are likely to be related to developmental changes in human sensitivity to ketamine; whereas ketamine induces psychosis in adults, it does not do so in children (White et al. 1982). The mechanism of these developmental changes is unclear. Not much is known about the developmental expression of NMDAR in interneurons, but one study found levels in 1-mo-old rats lower than those in adult rats (Lau et al. 2003). Another study found that PV interneurons in culture have a different NMDAR subunit composition compared with that of similar-age pyramidal neurons (Kinney et al. 2006). Besides the potential change of interneuron NMDAR composition during development, other receptors could also be involved in the age-dependent sensitivity to ketamine. Activation of dopamine D2 receptor excites interneurons only in adult rats, but not in preadolescent rats (Tseng and O'Donnell 2007); since there are strong dopamine–NMDA interactions (Tseng and O'Donnell 2004), the age dependence of the ketamine effect could be related to the age dependence of dopaminergic properties. Further work will be required to clarify this important issue.

GRANTS

This work was supported by National Institute of Mental Health Conte Center Grant 5P50 MH-060450-08.