Endogenous activation of nAChRs and NMDA receptors contributes to the excitability of CA1 stratum radiatum interneurons in rat hippocampal slices: effects of kynurenic acid

Abstract

CA1 stratum radiatum interneurons (SRIs) express α7 nicotinic receptors (nAChRs) and receive inputs from glutamatergic neurons/axons that express α3β4β2 nAChRs. To test the hypothesis that endogenously active α7 and/or α3β4β2 nAChRs control the excitability of CA1 SRIs in the rat hippocampus, we examined the effects of selective receptor antagonists on spontaneous fast current transients (CTs) recorded from these interneurons under cell-attached configuration. The frequency of CTs, which represent action potentials, increased in the absence of extracellular Mg(2+) and decreased in the presence of the α3β4β2 nAChR antagonist mecamylamine (3 μM) or the NMDA receptor antagonist APV (50 μM). However, it was unaffected by the α7 nAChR antagonist MLA (10 nM) or the AMPA receptor antagonist CNQX (10 μM). Thus, in addition to synaptically and tonically activated NMDA receptors, α3β4β2 nAChRs that are present on glutamatergic axons/neurons synapsing onto SRIs and are activated by basal levels of acetylcholine contribute to the maintenance of the excitability of these interneurons. Kynurenic acid (KYNA), an astrocyte-derived kynurenine metabolite whose levels are increased in the brains of patients with schizophrenia, also controls the excitability of SRIs. At high micromolar concentrations, KYNA, acting primarily as an NMDA receptor antagonist, decreased the CT frequency recorded from the interneurons. At 2 μM, KYNA reduced the CA1 SRI excitability via mechanisms independent of NMDA receptor block. KYNA-induced reduction of excitability of SRIs may contribute to sensory gating deficits that have been attributed to deficient hippocampal GABAergic transmission and high levels of KYNA in the brain of patients with schizophrenia.

Fig. 1

Pattern of firing of CA1 SRIs in rat hippocampal slices. Samples of current transients (CTs) recorded from an SRI using cell-attached configuration. All recordings were performed in the voltage-clamp mode at −60 mV in the continuous presence of the GABA_A receptor antagonist bicuculline (10 μM) and the muscarinic receptor antagonist atropine (0.5 μM). A. Segments (20-s each) of recordings obtained from an SRI under various conditions are shown. In the nominal absence of extracellular Mg²⁺, the firing rate of the SRI increased. This pattern illustrates the presence of non-overlapping CTs even during high-frequency firing. B. Segments of recordings obtained from another SRI under two different ACSF conditions. This pattern illustrates the presence of both isolated events and bursts of CTs as revealed in the expanded traces. Single CT shown at the bottom reveals a fast inward component and a slow outward component that resembles the components of an action potential in extracellular recordings [22,23]. C. Graph shows CT frequency recorded from several neurons in Mg²⁺-containing ACSF during 10 min and after exposure to Mg²⁺-free ACSF. In Mg²⁺-free ACSF, 5 min recording segments (i.e., 5–10, 15–20, and 25–30 min) were used for the analysis. Graph and error bars represent mean and S.E.M. of results from 92 neurons in Mg²⁺-containing ACSF and 58 neurons in Mg²⁺-free ACSF. b p < 0.01; d p < 0.0001 compared to Mg²⁺-ACSF by one-way ANOVA followed by Bonferroni comparison.

Fig. 2

Effect of MLA on CT frequency recorded from CA1 SRIs. A. Graph depicts the average frequency of CTs observed in each SRI under control condition (Mg²⁺-containing ACSF, 10 min), during 5–15 min bath exposure to 10 nM MLA, and during 10–40 min wash with MLA-free ACSF. Only cell #1 and cell #3 revealed a reversible decrease of CT frequency by MLA bath application. B. Graph depicts the frequency of CTs recorded in the continuous presence of 10 nM MLA from several SRIs in MLA (10 nM)-incubated slices and normalized to the mean frequency of CTs recorded from neurons in control slices (i.e., slices incubated in MLA-free ACSF). Graph and error bars represent mean and S.E.M., respectively, of results obtained from 13 neurons. C. Time course of increase in the frequency of CTs recorded from the same neurons as in B during their superfusion with Mg²⁺-free ACSF. Data points and error bars represent mean and S.E.M., respectively, of results obtained from 13 neurons in the MLA group and 58 neurons in the control group.

Fig. 3

Effect of mecamylamine on CT frequency recorded from CA1 SRIs. A. Graph depicts the frequency of CTs recorded in the continuous presence of mecamylamine (3 μM) from several SRIs in mecamylamine (3 μM)-incubated slices and normalized to the mean frequency recorded from neurons in control slices incubated in ACSF free of mecamylamine. Graph and error bars represent mean and S.E.M., respectively, of results obtained from nine neurons. B. Time course of increase in the frequency of CTs recorded from the same neurons as in A during their superfusion with Mg²⁺-free ACSF. Data points and error bars represent mean and S.E.M., respectively of results obtained from nine neurons in the mecamylamine group and 58 neurons in the control group. a p < 0.05; b p < 0.01; c p < 0.001 compared to respective control by unpaired t-test.

Fig. 4

Effect of CNQX and APV on CT frequency in CA1 SRI of rat hippocampal slices. A. Graph depicts the frequency of CTs recorded in the continuous presence of CNQX (10 μM) from several SRIs in CNQX (10 μM)-incubated slices and normalized to the mean frequency of CTs recorded from neurons in control slices incubated in ACSF free of CNQX. B. Graph depicts the frequency of CTs recorded in the continuous presence of APV (50 μM) from several SRIs in APV (50 μM)-incubated slices and normalized to the mean frequency of CTs recorded from neurons in control slices incubated in ACSF free of APV. Graph and error bars in A and B represent mean and S.E.M., respectively, of results obtained from 13 neurons incubated with CNQX and 9 neurons incubated with APV. C. Time-dependent changes in the frequency of CTs recorded from the same neurons as in A and B during their superfusion with Mg²⁺-free ACSF are shown. Data points and error bars represent mean and S.E.M., respectively of results obtained from 13 neurons in the CNQX group, 9 neurons in the APV group, and 58 neurons in the control group. a p < 0.05; b p < 0.01; c p < 0.001; d p < 0.0001 compared to respective control by unpaired t-test.

Fig. 5

Effect of high concentration of KYNA on the frequency of CTs and EPSCs in CA1 SRI of rat hippocampal slices. Sample recordings illustrate that brief bath application of 200 μM KYNA abolishes bursts of CTs and EPSCs recorded from an SRI in Mg²⁺-free ACSF. A1. Representative traces of cell-attached recordings from an SRI under control condition, 3 min after bath applied KYNA (200 μM), and 10 min after wash with Mg²⁺-free ACSF. The regions marked by asterisks are shown in an expanded scale to the right to reveal individual events in a single burst of CTs. Note the fast onset and reversal of inhibition. A2. Representative traces of whole-cell recordings at −60 mV from the same SRI as in A1 under control (i.e., during wash in Mg²⁺-free ACSF), 3 min after bath applied KYNA (200 μM), 5 min wash in Mg²⁺-free ACSF, and 5 min after exposure to Mg²⁺-containing ACSF. The regions marked by asterisks are shown in an expanded scale to the right to reveal overlapping EPSCs in each burst. Note that the number of EPSC bursts matched the number of CT bursts and that a brief exposure (<3 min) to KYNA (200 μM) abolished both types of events completely. B. In another SRI, cell-attached recordings reveal the presence of CTs as both isolated events and bursts in Mg²⁺-free ACSF. Bath application of KYNA (200 μM) predominantly abolished CT bursts in this cell.

Fig. 6

Effect of kynurenine and low concentrations of KYNA on the frequency of CTs in CA1 SRIs. A1. Sample traces of cell-attached recordings in slices incubated with 2 μM KYNA in the presence and in the absence of added Mg²⁺ in the ACSF. A2. Graph depicts the frequency of CTs recorded in the continuous presence of 2 μM KYNA from several SRIs in KYNA (2 μM)-incubated slices and normalized to the mean frequency recorded from neurons in control slices incubated in KYNA-free ACSF. A3. Time-dependent changes in the frequency of CTs recorded from the same neurons as in A2 during their superfusion with Mg²⁺-free ACSF are shown. Data in A2 and A3 are presented as mean and S.E.M. of results obtained from 12 neurons in the KYNA group and 58 neurons in the control group. B1. Sample traces of cell-attached recordings in slices incubated with 200 μM kynurenine in Mg²⁺-free and Mg²⁺-containing ACSF. B2. Graph depicts the frequency of CTs recorded in the continuous presence of 200 μM kynurenine from several SRIs in kynurenine (200 μM)-incubated slices and normalized to the mean frequency recorded from neurons in control slices incubated in kynurenine-free ACSF. B3. Time-dependent changes in the frequency of CTs recorded from the same neurons as in B2 during their superfusion with Mg²⁺-free ACSF are shown. Data in B2 and B3 are presented as mean and S.E.M. of results obtained from 22 neurons in ACSF containing Mg²⁺ and kynurenine, 13 neurons in Mg²⁺-free ACSF containing kynurenine, and 58 control neurons. b p < 0.01; c p < 0.001; d p < 0.0001 compared to respective control by unpaired t-test.

Fig. 7

Effect of KYNA on the amplitude distribution of AMPA EPSCs in CA1 SRI of hippocampal slices. Histograms of pooled data from several neurons reveal the distribution pattern of AMPA EPSC amplitudes recorded under different experimental conditions. Note that there are fewer large-amplitude events in the KYNA and APV groups compared to control. Bottom graph illustrates the relative frequency of large-amplitude AMPA EPSCs recorded under different experimental conditions. Number of events with threshold amplitudes of 20, 25 and 30 pA was normalized to the total number of events pooled under each experimental condition. About 60% of the events in the control group had amplitudes smaller than 20 pA. The percent value of the large-amplitude component in relation to the total number of events is significantly decreased by both concentrations of KYNA and by APV. b p < 0.01; c p < 0.001; d p < 0.0001, compared to control by Fisher’s exact test. Results are from 1061 events pooled from 6 cells in control, 421 events pooled from 6 cells in 2 μM KYNA, 193 events pooled from 13 cells in 100 μM KYNA, and 403 events pooled from 5 cells in 50 μM APV.

Fig. 8

Schematic representation of the putative neurocircuitry/receptors and targets likely to mediate the effects of the various pharmacological agents on CA1 SRI excitability in resting hippocampal slices. Axon from pyramidal neuron (P) releases glutamate, which, in turn activates AMPA and NMDA receptors on the SRIs (I) and lead to firing of action potentials. Lack of significant effect of MLA suggest that either the number of cholinergic synapses targeting α7 nAChRs on the SRIs is too low and/or that ACh-induced activation of α7 nAChR is too weak to initiate CT in the neurons. The potent inhibitory effect of mecamylamine suggests that basal levels of ACh in the hippocampal slices are sufficient to activate α3β4β2 nAChRs in glutamatergic neurons/axons that synapse onto the SRIs. The potent inhibitory effect of APV suggests that α3β4β2 nAChR activation by endogenous ACh may act in conjunction with NMDA receptor activation on both the SRIs themselves and the glumatergic neurons that synapse onto these interneurons to control the excitability of the latter. High concentration of KYNA inhibits NMDA receptors on both pyramidal neuron and SRIs to suppress SRI excitation. Low concentration of KYNA (2 μM), however, acts on an unknown target to reduce SRI excitability.