Glutamatergic Modulation of GABAergic ... - Wiley Online Library

12 downloads 0 Views 136KB Size Report
group III metabotropic receptors or kainate receptors. ... Results: Agonists of group III metabotropic glutamate recep- ..... Ghauri M, Chapman AG, Meldrum BS.
Epilepsia, 43(Suppl. 5):174–178, 2002 Blackwell Publishing, Inc. © International League Against Epilepsy

Glutamatergic Modulation of GABAergic Signaling Among Hippocampal Interneurons: Novel Mechanisms Regulating Hippocampal Excitability Dimitri M. Kullmann and Alexey Semyanov Institute of Neurology, UCL Queen Square, London, England

Summary: Purpose: Because interneurons play a central role in regulating the excitability of the hippocampal formation, it is important to understand the mechanisms that modulate ␥-aminobutyric acid (GABA)ergic signaling among them. This study addresses the modulation of GABA release from interneuron terminals by presynaptic glutamate receptors. Methods: Whole-cell recordings were obtained from CA1 stratum radiatum interneurons in guinea pig hippocampal slices. Selective agonists and blockers of glutamate receptors were used to study modulation of GABAergic transmission by group III metabotropic receptors or kainate receptors. Antidromic action-potential initiation also was analyzed by stimulating the axons of interneurons. Results: Agonists of group III metabotropic glutamate receptors attenuated monosynaptic GABAergic signals in interneurons, but not in pyramidal neurons, in agreement with anatomic evidence on the distribution of these receptors. Submicromolar kainate enhanced the frequency and amplitude of spontaneous

GABAergic signals in interneurons. Kainate also depolarized the axons of hippocampal interneurons, and triggered spontaneous ectopic action potentials in axons. Synaptically released glutamate reproduced many of the effects of both agonists, implying that these receptors can sense the ambient glutamate concentration, and therefore indirectly respond to the excitatory traffic in the hippocampus. When the two classes of receptors were stimulated simultaneously, complex interactions were obtained. Conclusions: Group III metabotropic receptors and kainate receptors profoundly affect GABAergic signaling among interneurons of the hippocampus. Glutamatergic modulation of GABAergic signaling among interneurons represents a novel class of mechanisms that potentially plays a major role in determining the initiation, propagation, and termination of seizures. Key Words: Metabotropic—Kainate—Ectopic action potentials—Spillover—Axonal excitability.

The ␥-aminobutyric acid (GABA)ergic interneurons are conventionally thought to act as a brake on the propagation of excitatory traffic in the hippocampus. However, they also may have disinhibitory effects if their main target is other interneurons. Further complicating their role in regulating excitability, interconnected interneurons have been proposed to contribute to the synchronization of principal cells (1,2). These properties potentially contribute to the triggering of abnormal population discharges. To understand the role of interneurons in regulating excitability, it also is important to understand the mechanisms that regulate GABA release downstream of actionpotential initiation in interneurons. Several high-affinity glutamate receptors potentially play a major role in this phenomenon. These may detect very low concentrations

of glutamate, such as can be reached in the extracellular space after extrasynaptic “spillover” of neurotransmitter. If these receptors are present at presynaptic GABAergic terminals, they could potentially alter GABA release as a function of the overall level of glutamate release, and therefore play an important role in sampling the volumeaveraged excitatory traffic in the network. In light of evidence that synaptically released glutamate can have relatively remote effects in the hippocampus (3), we recently asked whether it modulates GABAergic signaling among interneurons. This work was prompted by the observation that glutamate receptors are present on GABAergic terminals or axons. Highresolution immunohistochemistry has demonstrated that some group III metabotropic glutamate receptors (mGluRs) are located on GABAergic terminals in the hippocampus. mGluR7 is located on presynaptic terminals within the synaptic cleft. Strikingly, whether the receptor is present at a synapse or not depends on the identity of the target cell: it often is present if the postsynaptic neuron is an interneuron, but not if it is a py-

Address correspondence and reprint requests to Dr. D.M. Kullmann at Institute of Neurology, UCL Queen Square, London, WC1N 3BG, U.K. E-mail: [email protected]

174

GLUTAMATERGIC MODULATION OF INHIBITION ramidal neuron (4). Coinciding with the immunohistochemical data, application of the selective group III mGluR agonist L-AP4 has been shown to depress excitatory postsynaptic currents (EPSCs) recorded in interneurons but not in pyramidal neurons (5). The closely related receptor mGluR4, also present in the hippocampus, has been demonstrated at the presynaptic side not only of glutamatergic, but also of GABAergic synapses, identified as such by the absence of a prominent postsynaptic density (6). Because this receptor subtype has a much higher affinity for glutamate and L-AP4 than mGluR7, it is an excellent candidate heteroreceptor, selectively modulating GABAergic transmission in response to ambient glutamate. By extrapolation from the target-cell dependence of mGluR7, it may actually selectively modulate GABAergic transmission among interneurons, rather than inhibition of pyramidal neurons. Far less is known of the distribution of kainate receptors at a subcellular level. Nevertheless, these receptors also have been implicated in presynaptic modulation of transmitter release. Although there are many reports that kainate and other agonists depress GABAergic signals evoked in pyramidal neurons (7,8), it remains unclear to what extent this can be explained by a direct effect on transmitter release (possibly via a metabotropic action of the receptor) (9,10). Kainate also very effectively brings interneurons to firing threshold (11,12), and it has been proposed that kainate-induced interneuron firing causes an accumulation of extracellular GABA, which, only as a secondary consequence, depresses GABAergic transmission to pyramidal neurons. We recently showed that both group III mGluRs and kainate receptors modulate GABAergic transmission among interneurons (13,14). Here we summarize these data, compare the effects of group III mGluRs and kainate receptors on GABAergic signaling, and address the interactions between these modulatory influences on the interneuron network.

175

RESULTS Group III mGluRs depress GABA release The principal lines of evidence that group III mGluRs modulate monosynaptic evoked inhibitory postsynaptic currents (IPSCs) are described in detail elsewhere (13). In brief, pharmacologically isolated IPSCs recorded in interneurons were depressed by 50 ␮M L-AP4 to a much greater extent than IPSCs recorded in pyramidal cells, consistent with the target-cell dependence of group III mGluRs (4,6). The depression was accompanied by changes in paired-pulse depression and trial-to-trial IPSC variability implying a presynaptic depression in GABA release. Strikingly, IPSCs in interneurons were as profoundly depressed as glutamatergic EPSCs (Fig. 1). Taken together with the immunohistochemical data, this observation argues that the presynaptic location of group III mGluRs in the hippocampus, and the consequence for transmitter release, are entirely determined by the identity of the postsynaptic cell, and that whether GABA or glutamate is released from the presynaptic terminal plays no role. We asked whether glutamate could diffuse from neighboring excitatory synapses to presynaptic group III mGluRs on interneuron terminals. Two stimulating electrodes were used, one to activate a monosynaptic IPSC in the target interneuron, and another to activate Schaffer collaterals to evoke glutamate release. Because this conditioning stimulus was relatively remote, and ionotropic glutamate receptors were blocked, it did not evoke a large monosynaptic IPSC in its own right. Highfrequency bursts of stimuli delivered to the Schaffer collaterals depressed the monosynaptic test IPSC delivered after an interval of 100 ms (13). This depression of inhibition was enhanced by blocking glutamate uptake

METHODS Whole-cell recordings were obtained from either pyramidal neurons or stratum radiatum interneurons in the CA1 field of guinea pig hippocampal slices using conventional methods (13,14). For the study of group III mGluRs, the experiments were routinely performed with blocking concentrations of the following antagonists of ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA), kainate, N-methyl-D-aspartate (NMDA), and GABAB receptors: 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulphonamide (NBQX), DL-2amino-5-phosphonovalerate (APV), and CGP52432. To study kainate receptor–mediated modulation, AMPA receptors were blocked with GYKI53655, and group III mGluRs were blocked with ␣-methylserine-O-phosphate (MSOP).

FIG. 1. Group III metabotropic glutamate receptors mGluRdependent modulation of transmission depends on the identity of the target cell, and not of the neurotransmitter released. The group III mGluR agonist L-AP4 (50 µM) had little effect on transmission to pyramidal neurons, but depressed evoked excitatory postsynaptic currents (EPSCs) and inhibitory postsynaptic currents (IPSCs) recorded in interneurons to approximately the same extent. Data were normalized to the baseline amplitude of the synaptic signal, and averaged across the neurons (±SEM). Adapted from reference 13.

Epilepsia, Vol. 43, Suppl. 5, 2002

176

D. M. KULLMANN AND A. SEMYANOV

with dihydrokainate and was abolished by inhibiting group III mGluRs with the selective antagonist MSOP. These results are consistent with the hypothesis that glutamate accumulating in the extracellular space depresses GABAergic inhibition of interneurons by activating presynaptic mGluRs. Kainate receptors depolarize axons Some of the evidence that kainate receptors promote the release of GABA from the terminals of interneurons was recently published (14). We noted that low concentrations of kainate (250 nM–1 ␮M) caused a large increase in holding current in interneurons recorded with a CsCl-based pipette solution. This intracellular solution ensures that activation of either kainate or GABAA receptors causes an inward current. Bath perfusion of kainate (1 ␮M), in the presence of blocking concentrations of AMPA, NMDA, and GABAB-receptor antagonists, caused an intense barrage of IPSCs, as well as a large increase in the holding current, which was attenuated by blocking GABAA receptors with picrotoxin (100 ␮M; Fig. 2). This implies that most of the increase in holding current was not due to direct activation of kainate receptors on the neuronal membrane, but was secondary to an increase in the extracellular GABA concentration. The effect of kainate also was attenuated by preventing firing

FIG. 2. Kainate application causes a large increase in extracellular ␥-aminobutyric acid (GABA) concentration by evoking interneuron firing. Kainate application (1 µM) evoked an increase in holding current in hippocampal interneurons held at –60 mV. When applied in the presence of picrotoxin (100 µM) or tetrodotoxin (TTX, 1 µM), or together with L-AP4 (50 µM), the effect of kainate was greatly attenuated. Insets show the time course of the normalized holding current, averaged across the neurons (with gray areas showing ±SEM). Adapted from reference 14.

Epilepsia, Vol. 43, Suppl. 5, 2002

FIG. 3. Schematic illustration of the location and effect of group III metabotropic glutamate receptors (mGluRs) and kainate receptors on hippocampal interneurons. Two synaptically linked ␥-aminobutyric acid (GABA)ergic interneurons are shown, one projecting to a pyramidal neuron (left). A Schaffer collateral releases glutamate (illustrated as clouds) into the extracellular space, where it can activate the two classes of receptors. Group III mGluRs selectively decrease neurotransmitter release at synapses among interneurons. Kainate receptors on interneurons depolarize their axons and may either initiate or facilitate the propagation of action potentials.

of surrounding neurons with tetrodotoxin (1 ␮M), confirming that the source of GABA elevation was action potential–dependent release from interneurons. Moreover, consistent with the hypothesis that group III metabotropic glutamate receptors depress GABA release, L-AP4 (50 ␮M) also attenuated the effect of kainate. These results confirm that kainate exerts a powerful drive to interneurons, causing them to release GABA. How do kainate receptors cause interneurons to fire? It has been proposed that kainate activates somatodendritic synaptic receptors in interneurons, the function of which is normally to mediate a kainate receptor–mediated EPSC (11,12,15). Surprisingly, when we looked for kainate receptor–mediated EPSCs in interneurons, we were unable to evoke them, even with high-intensity and high-frequency bursts of stimuli (14). Previous reports of kainate receptor–mediated EPSCs in interneurons have been obtained in rat slices, so this may reflect a species difference. We asked whether kainate receptors can depolarize axons directly and cause ectopic action potentials. We recorded from stratum radiatum interneurons (with AMPA, NMDA, group III mGlu, GABAA, and GABAB receptors blocked), and positioned a stimulating electrode in stratum oriens to trigger antidromic action potentials (14). Because we recorded in voltage-clamp mode with a KCl-based pipette solution, these appeared as action currents. We then adjusted the stimulus intensity to obtain intermittent failures to trigger an action current. Application of 250 mM or 1 ␮M kainate produced a robust increase in success rate, consistent with kainate receptor–mediated axonal depolarization. We asked whether an increase in extracellular K+ concentration, secondary to kainate-induced depolarization of neighboring neurons, could explain the enhanced axonal

GLUTAMATERGIC MODULATION OF INHIBITION excitability (16). Increasing the K+ concentration from 2.5 to 5 mM had no effect on axonal excitability, despite an increase in holding current that was larger than that obtained with kainate. Thus, the effect of kainate is unlikely to be mediated indirectly by promoting K+ accumulation. We asked whether synaptically released glutamate could mimic the effect of exogenous kainate. We examined the success rate of a just-threshold antidromic stimulus delivered either on its own, or 50 ms after a train of stimuli delivered to Schaffer collaterals, designed to produce an extracellular glutamate surge. The success rate after the conditioning stimulus was significantly enhanced, and this enhancement was abolished by blocking kainate receptors. This implies that glutamate can indeed diffuse from excitatory afferent terminals to depolarize axons of interneurons via kainate receptors (Fig. 3). What is the consequence of axonal depolarization? Kainate application not only lowered the threshold for antidromic action potential initiation, but also evoked spontaneous action currents in the majority of interneurons held in voltage clamp. These action currents could not be prevented by hyperpolarizing the cell body. We therefore conclude that kainate evokes ectopic action potentials, which propagate antidromically. Given that kainate receptors depolarize the axons of interneurons, can these receptors also enhance GABA release through direct depolarization of terminals? Conflicting data have been reported on the effect of kainate on GABA release in the absence of propagated action potentials. Cossart et al. (23) observed that kainate (250 nM) enhanced the frequency of spontaneous action potential–independent miniature IPSCs (mIPSCs) in interneurons. Surprisingly, they did not see this in pyramidal neurons. Kainate (300 nM) has, however, been reported to enhance unitary IPSCs in pyramidal neurons (17). We saw no effect of 250 nM to 1 ␮M kainate on mIPSCs in guinea-pig interneurons. This may reflect a species difference: when we examined rat slices, 1 ␮M kainate (but not 250 nM kainate) did enhance the frequency of mIPSCs in interneurons (14). A possible explanation for the results is that kainate can cause axonal depolarization, which, if sufficiently intense, can propagate passively to GABAergic terminals to enhance transmitter release probability. (Ca2+ influx through Ca2+-permeable kainate receptors may also trigger exocytosis.) mGluR and kainate receptor effects are not opposite How do kainate and mGlu receptors modulate spontaneous GABAergic signaling among interneurons? Group III mGlu and kainate receptors apparently have opposite actions on the inhibitory drive to hippocampal stratum radiatum interneurons. The two effects might cancel out under physiologic conditions. However, their

177

effector mechanisms are distinct. The action of group III mGlu receptors is most simply explained by a direct action on transmitter release, which should depress both action potential–dependent and action potential– independent IPSCs. In contrast, a major effect of kainate receptors is to depolarize axons and possibly initiate ectopic action potentials, with very little effect on action potential–independent GABA release. Synchronous activation of both classes of receptors by endogenous glutamate might therefore selectively enhance the propagation of action potential–dependent IPSCs, with a lesser effect on the background action potential– independent GABAergic inhibition. We examined the effect of synchronous activation of group III mGlu and kainate receptors on spontaneous GABAergic IPSCs in interneurons, by applying both agonists in the continued presence of AMPA, NMDA, and GABAB receptor blockers (Fig. 4, A. Semyanov and D. M. Kullmann, unpublished observations). L-AP4 (50 ␮M) depressed the frequency of spontaneous IPSCs, whereas kainate (1 ␮M) had the opposite effect. When applied together, the kainate-evoked enhancement in frequency dominated over the L-AP4 effect. We attempted to distinguish between action potential–dependent multiquantal IPSCs and background action potential–

FIG. 4. Effect of separate or simultaneous activation of group III metabotropic glutamate receptors (mGluRs) and kainate receptors on the frequency of spontaneous inhibitory postsynaptic currents (sIPSCs) recorded in interneurons. Multiquantal action potential–dependent IPSCs were tentatively identified by their large amplitude (>50 pA). Group III mGluR activation decreases the frequency of both large and small sIPSCs, consistent with an effect on ␥-aminobutyric acid (GABA) release. Kainate increases the frequency of sIPSCs, and has a greater effect on large, possibly action potential–dependent multiquantal IPSCs. When both agonists were applied together, the frequency of large IPSCs underwent a relatively larger increase.

Epilepsia, Vol. 43, Suppl. 5, 2002

178

D. M. KULLMANN AND A. SEMYANOV

independent IPSCs, by subdividing them into small and large events (arbitrarily cut off at 50-pA amplitude). Coapplication of both agonists produced a relatively selective increase in large, putatively multiquantal IPSCs, consistent with the hypothesis that the two agonists do not have directly opposite roles on inhibition in the hippocampus. DISCUSSION The main results summarized here are as follows. Presynaptic group III mGluRs depress GABA release relatively selectively at synapses among interneurons. Kainate receptors, on the other hand, depolarize the axons of interneurons, trigger ectopic axonal action potentials, and enhance GABA release through multiple mechanisms. Synaptically released glutamate can reach both receptors, potentially playing a role in modulating of GABAergic signaling during ictogenesis. These results may shed some light on the observation that exogenous group III mGluR agonists have antiepileptic activity in experimental rodent seizure models (18–21). A reduction in inhibition of interneurons might cause them to fire at higher rates, and this phenomenon could increase the GABAergic drive to pyramidal neurons, thereby reducing the spread or initiation of a seizure. Although we have demonstrated that axonal kainate receptors trigger ectopic action potentials, we have not been able to show this phenomenon with synaptically released glutamate. If this phenomenon occurs, it would represent a totally novel mechanism for the integration of excitatory traffic in the hippocampus. Clearly much work remains to be done before the modulation of GABAergic transmission by glutamate receptors is fully understood. Although our recent work has uncovered some of the mechanisms underlying this phenomenon, a major task will be to assess how it affects the spontaneously firing intact network. The relative importance of group III mGlu and kainate receptor– dependent modulation of GABAergic transmission must also be seen in the context of the rich diversity of interneuron types, many of which have a disinhibitory role (22). The distinction between dendritic and somatic inhibition, which may have opposite effects on the initiation and termination of seizures (23), further complicates the interpretation of alterations in GABAergic signaling to principal cells. Acknowledgment: This work was supported by the Medical Research Council.

Epilepsia, Vol. 43, Suppl. 5, 2002

REFERENCES 1. Cobb SR, Buhl EH, Halasy K, et al. Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons. Nature 1995;378:75–8. 2. Whittington MA, Traub RD, Jefferys JG. Synchronized oscillations in interneuron networks driven by metabotropic glutamate receptor activation [see comments]. Nature 1995;373:612–5. 3. Kullmann DM. Spillover and synaptic cross talk mediated by glutamate and GABA in the mammalian brain. Prog Brain Res 2000;125:339–51. 4. Shigemoto R, Kulik A, Roberts JD, et al. Target-cell-specific concentration of a metabotropic glutamate receptor in the presynaptive active zone. Nature 1996;381:523–5. 5. Scanziani M, Gahwiler BH, Charpak S. Target cell-specific modulation of transmitter release at terminals from a single axon. Proc Natl Acad Sci U S A 1998;95:12004–9. 6. Shigemoto R, Kinoshita A, Wada E, et al. Differential presynaptic localization of metabotropic glutamate receptor subtypes in the rat hippocampus. J Neurosci 1997;17:7503–22. 7. Fisher RS, Alger BE. Electrophysiological mechanisms of kainic acid-induced epileptiform activity in the rat hippocampal slice. J Neurosci 1984;4:1312–23. 8. Rodriguez-Moreno A, Herreras O, Lerma J. Kainate receptors presynaptically downregulate GABAergic inhibition in the rat hippocampus. Neuron 1997;19:893–901. 9. Rodriguez-Moreno A, Lerma J. Kainate receptor modulation of GABA release involves a metabotropic function. Neuron 1998;20:1211–8. 10. Rodriguez-Moreno A, Lopez-Garcia JC, Lerma J. Two populations of kainate receptors with separate signaling mechanisms in hippocampal interneurons. Proc Natl Acad Sci U S A 2000;97:1293–8. 11. Cossart R, Esclapez M, Hirsch JC, et al. GluR5 kainate receptor activation in interneurons increases tonic inhibition of pyramidal cells [see comments]. Nat Neurosci 1998;1:470–8. 12. Frerking M, Malenka RC, Nicoll RA. Synaptic activation of kainate receptors on hippocampal interneurons. Nat Neurosci 1998; 1:479–86. 13. Semyanov A, Kullmann DM. Modulation of GABAergic signaling among interneurons by metabotropic glutamate receptors. Neuron 2000;25:663–72. 14. Semyanov A, Kullmann DM. Kainate receptor-dependent axonal depolarization and action potential initiation in interneurons. Nat Neurosci 2001;4:718–23. 15. Frerking M, Petersen CC, Nicoll RA. Mechanisms underling kainate receptor-mediated disinhibition in the hippocampus. Proc Natl Acad Sci U S A 1999;96:12917–22. 16. Schmitz D, Frerking M, Nicoll RA. Synaptic activation of presynaptic kainate receptors on hippocampal mossy fiber synapses. Neuron 2000;27:327–38. 17. Jiang L, Xu J, Nedergaard M, et al. A kainate receptor increases the efficacy of GABAergic synapses. Neuron 2001;30:503–13. 18. Abdul-Ghani AS, Attwell PJ, Singh Kent N, et al. Anti-epileptogenic and anticonvulsant activity of L-2-amino-4-phosphonobutyrate, a presynaptic glutamate receptor agonist. Brain Res 1997;755:202–12. 19. Chapman AG, Nanan K, Yip P, et al. Anticonvulsant activity of a metabotropic glutamate receptor 8 preferential agonist, (R,S)-4phosphonophenylglycine. Eur J Pharmacol 1999;383:23–7. 20. Ghauri M, Chapman AG, Meldrum BS. Convulsant and anticonvulsant actions of agonists and antagonists of group III mGluRs. Neuroreport 1996;7:1469–74. 21. Tizzano JP, Griffey KI, Schoepp DD. Induction or protection of limbic seizures in mice by mGluR subtype selective agonists. Neuropharmacology 1995;34:1063–7. 22. Freund TF, Buzsaki G. Interneurons of the hippocampus. Hippocampus 1996;6:347–470. 23. Cossart R, Tyzio R, Dinocourt C, et al. Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy. Nat Neurosci 2001;4:52–62.