Search

The basal forebrain cholinergic system is widely assumed to control cortical functions via non-synaptic transmission of a single neurotransmitter. Yet, we find that mouse hippocampal cholinergic terminals invariably establish GABAergic synapses, and their cholinergic vesicles dock at those synapses only. We demonstrate that these synapses do not co-release but co-transmit GABA and acetylcholine via different vesicles, whose release is triggered by distinct calcium channels. This co-transmission evokes composite postsynaptic potentials, which are mutually cross-regulated by presynaptic autoreceptors. Although postsynaptic cholinergic receptor distribution cannot be investigated, their response latencies suggest a focal, intra- and/or peri-synaptic localisation, while GABA A receptors are detected intra-synaptically. The GABAergic component alone effectively suppresses hippocampal sharp wave-ripples and epileptiform activity. Therefore, the differentially regulated GABAergic and cholinergic co-transmission suggests a hitherto unrecognised level of control over cortical states. This novel model of hippocampal cholinergic neurotransmission may lead to alternative pharmacotherapies after cholinergic deinnervation seen in neurodegenerative disorders.

Although endocannabinoids have emerged as essential retrograde messengers in several forms of synaptic plasticity, it remains controversial whether they mediate long-term depression (LTD) of glutamatergic synapses onto excitatory and inhibitory neurons in the hippocampus. Here, we show that parvalbumin- and somatostatin/metabotropic glutamate receptor 1(a) (mGlu(1a))-positive GABAergic interneurons express diacylglycerol lipase-α (DGL-α), a synthesizing enzyme of the endocannabinoid 2-arachidonoylglycerol (2-AG), albeit at lower levels than principal cells. Moreover, this lipase accumulates postsynaptically around afferent excitatory synapses in all three cell types. To address the role of retrograde 2-AG signaling in LTD, we investigated two forms: (1) produced by postsynaptic spiking paired with subsequent presynaptic stimulation or (2) induced by group I mGlu activation by (S)-3,5-dihydroxyphenylglycine (DHPG). Neither form of LTD was evoked in the presence of the mGlu(5) antagonist MPEP [2-methyl-6-(phenylethynyl)-pyridine], the DGL inhibitor THL [N-formyl-l-leucine (1S)-1-[[(2S,3S)-3-hexyl-4-oxo-2-oxetanyl]methyl]dodecyl ester], or the intracellularly applied Ca(2+) chelator BAPTA in CA1 pyramidal cells, fast-spiking interneurons (representing parvalbumin-containing cells) and interneurons projecting to stratum lacunosum-moleculare (representing somatostatin/mGlu(1a)-expressing interneurons). Both forms of LTD were completely absent in CB(1) cannabinoid receptor knock-out mice, whereas pharmacological blockade of CB(1) led to inconsistent results. Notably, in accordance with their lower DGL-α level, a higher stimulation frequency or higher DHPG concentration was required for LTD induction in interneurons compared with pyramidal cells. These findings demonstrate that hippocampal principal cells and interneurons produce endocannabinoids to mediate LTD in a qualitatively similar, but quantitatively different manner. The shifted induction threshold implies that endocannabinoid-LTD contributes to cortical information processing during distinct network activity patterns in a cell type-specific manner.

GABAergic inhibition plays a central role in the control of pyramidal cell ensemble activities; thus, any signaling mechanism that regulates inhibition is able to fine-tune network patterns. Here, we provide evidence that the retrograde nitric oxide (NO)-cGMP cascade triggered by NMDA receptor (NMDAR) activation plays a role in the control of hippocampal GABAergic transmission in mice. GABAergic synapses express neuronal nitric oxide synthase (nNOS) postsynaptically and NO receptors (NO-sensitive guanylyl cyclase) in the presynaptic terminals. We hypothesized that--similar to glutamatergic synapses--the Ca(2+) transients required to activate nNOS were provided by NMDA receptor activation. Indeed, administration of 5 μm NMDA induced a robust nNOS-dependent cGMP production in GABAergic terminals, selectively in the CA1 and CA3c areas. Furthermore, using preembedding, postembedding, and SDS-digested freeze-fracture replica immunogold labeling, we provided quantitative immunocytochemical evidence that NMDAR subunits GluN1, GluN2A, and GluN2B were present in most somatic GABAergic synapses postsynaptically. These data indicate that NMDARs can modulate hippocampal GABAergic inhibition via NO-cGMP signaling in an activity-dependent manner and that this effect is subregion specific in the mouse hippocampus.

Information processing in the hippocampus critically relies on its reciprocal interaction with the medial septum (MS). Synchronization of the septo-hippocampal system was demonstrated during both major hippocampal activity states, the regular theta rhythm and the large amplitude irregular activity. Previous experimental and modeling data suggest that the MS provides rhythmic drive to the hippocampus, and hippocampo-septal feedback synchronizes septal pacemaker units. However, this view has recently been questioned based on the possibility of intrahippocampal theta genesis. Previously, we identified putative pacemaker neurons expressing parvalbumin (PV) and/or the pacemaker hyperpolarization-activated and cyclic nucleotide-gated nonselective cation channel (HCN) in the MS. In this study, by analyzing the temporal relationship of activity between the PV/HCN-containing medial septal neurons and hippocampal local field potential, we aimed to uncover whether the sequence of events during theta formation supports the classic view of septal drive or the challenging theory of hippocampal pacing of theta. Importantly, by implementing a circular statistical method, a temporal lead of these septal neurons over the hippocampus was observed on the course of theta synchronization. Moreover, the activity of putative hippocampal interneurons also preceded hippocampal local field theta, but by a shorter time period compared with PV/HCN-containing septal neurons. Using the concept of mutual information, the action potential series of PV/HCN-containing neurons shared higher amount of information with hippocampal field oscillation than PV/HCN-immunonegative cells. Thus, a pacemaker neuron population of the MS leads hippocampal activity, presumably via the synchronization of hippocampal interneurons.

The medial septum (MS) is an indispensable component of the subcortical network which synchronizes the hippocampus at theta frequency during specific stages of information processing. GABAergic neurons exhibiting highly regular firing coupled to the hippocampal theta rhythm are thought to form the core of the MS rhythm-generating network. In recent studies the hyperpolarization-activated, cyclic nucleotide-gated non-selective cation (HCN) channel was shown to participate in theta synchronization of the medial septum. Here, we tested the hypothesis that HCN channel expression correlates with theta modulated firing behaviour of MS neurons by a combined anatomical and electrophysiological approach. HCN-expressing neurons represented a subpopulation of GABAergic cells in the MS partly overlapping with parvalbumin (PV)-containing neurons. Rhythmic firing in the theta frequency range was characteristic of all HCN-expressing neurons. In contrast, only a minority of HCN-negative cells displayed theta related activity. All HCN cells had tight phase coupling to hippocampal theta waves. As a group, PV-expressing HCN neurons had a marked bimodal phase distribution, whereas PV-immunonegative HCN neurons did not show group-level phase preference despite significant individual phase coupling. Microiontophoretic blockade of HCN channels resulted in the reduction of discharge frequency, but theta rhythmic firing was perturbed only in a few cases. Our data imply that HCN-expressing GABAergic neurons provide rhythmic drive in all phases of the hippocampal theta activity. In most MS theta cells rhythm genesis is apparently determined by interactions at the level of the network rather than by the pacemaking property of HCN channels alone.

Neuroscience produces a vast amount of data from an enormous diversity of neurons. A neuronal classification system is essential to organize such data and the knowledge that is derived from them. Classification depends on the unequivocal identification of the features that distinguish one type of neuron from another. The problems inherent in this are particularly acute when studying cortical interneurons. To tackle this, we convened a representative group of researchers to agree on a set of terms to describe the anatomical, physiological and molecular features of GABAergic interneurons of the cerebral cortex. The resulting terminology might provide a stepping stone towards a future classification of these complex and heterogeneous cells. Consistent adoption will be important for the success of such an initiative, and we also encourage the active involvement of the broader scientific community in the dynamic evolution of this project.

Intact endogenous cannabinoid signaling is involved in several aspects of drug addiction. Most importantly, endocannabinoids exert pronounced influence on primary rewarding effects of abused drugs, including exogenous cannabis itself, through the regulation of drug-induced increase in bursting activity of dopaminergic neurons in the ventral tegmental area (VTA). Previous electrophysiological studies have proposed that these dopaminergic neurons may release endocannabinoids in an activity-dependent manner to regulate their various synaptic inputs; however, the underlying molecular and anatomical substrates have so far been elusive. To facilitate understanding of the neurobiological mechanisms involving endocannabinoid signaling in drug addiction, we carried out detailed analysis of the molecular architecture of the endocannabinoid system in the VTA. In situ hybridization for sn-1-diacylglycerol lipase-alpha (DGL-alpha), the biosynthetic enzyme of the most abundant endocannabinoid, 2-arachidonoylglycerol (2-AG), revealed that DGL-alpha was expressed at moderate to high levels by most neurons of the VTA. Immunostaining for DGL-alpha resulted in a widespread punctate pattern at the light microscopic level, whereas high-resolution electron microscopic analysis demonstrated that this pattern is due to accumulation of the enzyme adjacent to postsynaptic specializations of several distinct morphological types of glutamatergic and GABAergic synapses. These axon terminal types carried presynaptic CB(1) cannabinoid receptors on the opposite side of DGL-alpha-containing synapses and double immunostaining confirmed that DGL-alpha is present on the plasma membrane of both tyrosine hydroxylase (TH)-positive (dopaminergic) and TH-negative dendrites. These findings indicate that retrograde synaptic signaling mediated by 2-AG via CB(1) may influence the drug-reward circuitry at multiple types of synapses in the VTA.

Functionally distinct subsets of hippocampal inhibitory neurons exhibit large differences in the frequency, pattern and short-term plasticity of GABA release from their terminals. Heterogeneity is also evident in the ultrastructural features of GABAergic axon terminals examined in the electron microscope, but it is not known if or how this corresponds to interneuron subtypes. We investigated the feasibility of separating morphologically distinct clusters of terminal types, using the approach of measuring several ultrastructural parameters of GABAergic terminals in the CA1 area of the rat hippocampus. Septo-hippocampal axon terminals were anterogradely labeled by biotinylated dextran amine and visualized by pre-embedding immunogold staining to delineate one homogeneous terminal population. Long series (100-150) of ultrathin sections were cut from stratum oriens and stratum radiatum of the CA1 area, and GABAergic terminals were identified by post-embedding immunogold staining. Stereologically unbiased samples of the total GABAergic axon terminal population and a random sample of the septal axon terminals were reconstructed in 3D, and several of their parameters were measured (e.g. bouton volume, synapse surface, volume occupied by vesicles, mitochondria volume). Septal terminals demonstrated significantly larger mean values for most parameters than the total population of local GABAergic terminals. There was no significant difference between terminals reconstructed in the basal and apical dendritic regions of pyramidal cells, neither for the septal nor for the local population. Importantly, almost all parameters were highly correlated, precluding the possibility of clustering the local terminals into non-overlapping subsets. Factor and cluster analysis confirmed these findings. Our results suggest that similarly to excitatory terminals, inhibitory terminals follow an "ultrastructural size principle," and that the terminals of different interneuron subtypes cannot be distinguished by ultrastructure alone.

Nitric oxide (NO) plays an important role in synaptic plasticity as a retrograde messenger at glutamatergic synapses. Here we describe that, in hippocampal pyramidal cells, neuronal nitric oxide synthase (nNOS) is also associated with the postsynaptic active zones of GABAergic symmetrical synapses terminating on their somata, dendrites, and axon initial segments in both mice and rats. The NO receptor nitric oxide-sensitive guanylyl cyclase (NOsGC) is present in the brain in two functional subunit compositions: alpha1beta1 and alpha2beta1. The beta1 subunit is expressed in both pyramidal cells and interneurons in the hippocampus. Using immunohistochemistry and in situ hybridization methods, we describe that the alpha1 subunit is detectable only in interneurons, which are always positive for beta1 subunit as well; however, pyramidal cells are labeled only for beta1 and alpha2 subunits. With double-immunofluorescent staining, we also found that most cholecystokinin- and parvalbumin-positive and smaller proportion of the somatostatin- and nNOS-positive interneurons are alpha1 subunit positive. We also found that the alpha1 subunit is present in parvalbumin- and cholecystokinin-positive interneuron terminals that establish synapses on somata, dendrites, or axon initial segments. Our results demonstrate that NOsGC, composed of alpha1beta1 subunits, is selectively expressed in different types of interneurons and is present in their presynaptic GABAergic terminals, in which it may serve as a receptor for NO produced postsynaptically by nNOS in the very same synapse.

Cannabinoid ligands show therapeutic potential in a variety of disorders including anxiety. However, the anxiety-related effects of cannabinoids remain controversial as agonists show opposite effects in mice and rats. Here we compared the effects of the cannabinoid agonist WIN-55,212 and the CB1 antagonist AM-251 in CD1 mice and Wistar rats. Special attention was paid to antagonist-agonist interactions, which had not yet been studied in rats. In mice, WIN-55,212 decreased whereas AM-251 increased anxiety. The antagonist abolished the effects of the agonist. In contrast, WIN-55,212 increased anxiety in rats. Surprisingly, the antagonist potentiated this effect. Cannabinoids affect both GABAergic and glutamatergic functions, which play opposite roles in anxiety. We hypothesized that discrepant findings resulted from species differences in the relative responsiveness of the two transmitter systems to cannabinoids. We investigated this hypothesis by studying the effects of WIN-55,212 on evoked hippocampal inhibitory and excitatory postsynaptic currents (IPSCs and EPSCs). IPSCs were one order of magnitude more sensitive to WIN-55,212 in mice than in rats. In mice, IPSCs were more sensitive than EPSCs to WIN-55,212. This is the first study showing that the relative cannabinoid sensitivity of GABA and glutamate neurotransmission is species-dependent. Based on behavioural and electrophysiological findings, we hypothesize that WIN-55,212 reduced anxiety in mice by affecting GABA neurotransmission whereas it increased anxiety in rats via glutamatergic mechanisms. In rats, AM-251 potentiated this anxiogenic effect by inhibiting the anxiolytic GABAergic mechanism. We suggest that the anxiety-related effects of cannabinoids depend on the relative cannabinoid responsiveness of GABAergic and glutamatergic neurotransmission.

Gamma frequency (30-100 Hz) network oscillations occur in the intact hippocampus during awake, attentive behavior. Here, we explored the underlying cellular mechanisms in an in vitro model of persistent gamma-frequency oscillations, induced by bath application of 20 microm carbachol in submerged hippocampal slices at 30 +/- 1 degrees C. Current-source density analysis of the field oscillation revealed a prominent alternating sink-source pair in the perisomatic and apical dendritic regions of CA3. To elucidate the active events generating these extracellular dipoles, we examined the firing properties of distinct neuron types. Visually guided unit recordings were obtained from individual CA3 neurons followed by intracellular labeling for anatomical identification. Pyramidal cells fired at 2.82 +/- 0.7 Hz, close to the negative peak of the oscillation (0.03 +/- 0.65 msec), and often in conjunction with a negative spike-like component of the field potential. In contrast, all phase-coupled interneurons fired after this negative peak. Perisomatic inhibitory interneurons fired at high frequency (18.1 +/- 2.7 Hz), shortly after the negative peak (1.97 +/- 0.95 msec) and were strongly phase-coupled. Dendritic inhibitory interneurons fired at lower frequency (8.4 +/- 2.4 Hz) and with less fidelity and a longer delay after the negative peak (4.3 +/- 1.1 msec), whereas interneurons with cell body in the stratum radiatum often showed no phase relationship with the field oscillation. The phase and spike time data of individual neurons, together with the current-source density analysis, support a synaptic feedback model of gamma oscillations primarily involving pyramidal cells and inhibitory cells targeting their perisomatic region.

GABAergic interneurons innervating the perisomatic region of pyramidal cells control population discharge patterns, and thereby all cognitive operations in the cerebral cortex. A striking dichotomy in the function of this interneuron population seems to emerge from the synthesis of recent molecular, anatomical and electrophysiological data. Synaptically and electrically coupled networks of parvalbumin-containing basket cells operate as a non-plastic, precision clockwork for gamma and theta oscillations, and are indispensable for basic cortical processing. By contrast, a highly modifiable interneuron syncytium containing cholecystokinin carries information from subcortical pathways about the emotional, motivational and general physiological state of the animal, and appears to be involved in the fine-tuning of network cooperativity. Impairment of this inhibitory mechanism is likely to result in mood disorders such as anxiety.

Thalamocortical circuits that govern cortical rhythms and ultimately effect sensory transmission consist of three major interconnected elements: excitatory thalamocortical and corticothalamic neurons and GABAergic cells in the reticular thalamic nucleus. Based on the present results, a fourth component has to be added to this scheme. GABAergic fibres from an extrareticular diencephalic source were found to selectively innervate relay cells located mainly in higher-order thalamic nuclei. The origin of this pathway was localized to zona incerta (ZI), known to receive collaterals from corticothalamic fibres. First-order nuclei were innervated only in zones showing a high density of calbindin-positive neurons. The large GABA-immunoreactive incertal terminals established multiple contacts preferentially on the proximal dendrites of relay cells via symmetrical synapses with multiple release sites. The distribution, ultrastructural characteristics and postsynaptic target selection of extrareticular terminals were similar to type II muscarinic acetylcholine receptor-positive boutons, which constituted up to 49% of all GABAergic terminals in the posterior nucleus. This suggests that a significant proportion of the GABAergic input into certain thalamic territories involved in higher-order functions may have extrareticular origin. Unlike the reticular nucleus, ZI receives peripheral and layer V cortical input but no thalamic feedback; it projects to brainstem centres and has extensive intranuclear recurrent collaterals. This indicates that ZI exerts a conceptually new type of inhibitory control over the thalamus. The proximally situated, multiple active zones of ZI terminals indicate a powerful influence on the firing properties of thalamic neurons, which is conveyed to multiple cortical areas via relay cells which have widespread projections to neocortex.

Cannabinoids are the most popular illicit drugs used for recreational purposes worldwide. However, the neurobiological substrate of their mood-altering capacity has not been elucidated so far. Here we report that CB1 cannabinoid receptors are expressed at high levels in certain amygdala nuclei, especially in the lateral and basal nuclei, but are absent in other nuclei (e.g., in the central nucleus and in the medial nucleus). Expression of the CB1 protein was restricted to a distinct subpopulation of GABAergic interneurons corresponding to large cholecystokinin-positive cells. Detailed electron microscopic investigation revealed that CB1 receptors are located presynaptically on cholecystokinin-positive axon terminals, which establish symmetrical GABAergic synapses with their postsynaptic targets. The physiological consequence of this particular anatomical localization was investigated by whole-cell patch-clamp recordings in principal cells of the lateral and basal nuclei. CB1 receptor agonists WIN 55,212-2 and CP 55,940 reduced the amplitude of GABA(A) receptor-mediated evoked and spontaneous IPSCs, whereas the action potential-independent miniature IPSCs were not significantly affected. In contrast, CB1 receptor agonists were ineffective in changing the amplitude of IPSCs in the rat central nucleus and in the basal nucleus of CB1 knock-out mice. These results suggest that cannabinoids target specific elements in neuronal networks of given amygdala nuclei, where they presynaptically modulate GABAergic synaptic transmission. We propose that these anatomical and physiological features, characteristic of CB1 receptors in several forebrain regions, represent the neuronal substrate for endocannabinoids involved in retrograde synaptic signaling and may explain some of the emotionally relevant behavioral effects of cannabinoid exposure.

Psychoactive effects of cannabinoids are thought to be mediated, at least in part, by suppression of both glutamate and GABA release via CB1 cannabinoid receptor. Two types of cannabinoid receptor (CB1 and CB2) have been cloned so far. The CB1 receptors are abundantly expressed in the nervous system, whereas CB2 receptors are limited to lymphoid organs (Matsuda et al., 1990; Munro et al., 1993). Immunocytochemical and electrophysiological studies revealed that in the hippocampus CB1 receptors are expressed on axon terminals of GABAergic inhibitory interneurons (Tsou et al., 1999; Katona et al., 1999) and activation of these receptors decreases GABA release (Hájos et al., 2000). Other physiological studies pointed out the involvement of CB1 receptors in the modulation of hippocampal glutamatergic synaptic transmission and long-term potentiation (Stella et al., 1997; Misner and Sullivan, 1999), but anatomical studies could not confirm the existence of CB1 receptors on glutamatergic terminals. Here we examined cannabinoid actions on both glutamatergic and GABAergic synaptic transmission in the hippocampus of wild type (CB1+/+) and CB1 receptor knockout mice (CB1-/-). The synthetic cannabinoid agonist WIN55,212-2 reduced the amplitudes of excitatory postsynaptic currents in both wild type and CB1-/- mice, while inhibitory postsynaptic currents were decreased only in wild type mice, but not in CB1-/- animals. Our findings are consistent with a CB1 cannabinoid receptor-dependent modulation of GABAergic postsynaptic currents, but a novel cannabinoid-sensitive receptor must be responsible for the inhibition of glutamatergic neurotransmission.

Using a new antibody developed against the C-terminus of the cannabinoid receptor (CB1), the immunostaining in the hippocampus revealed additional axon terminals relative to the pattern reported previously with an N-terminus antibody. Due to a greater sensitivity of this antibody, a large proportion of boutons in the dendritic layers displaying symmetrical (GABAergic) synapses were also strongly immunoreactive for CB1 receptors, as were axon terminals of perisomatic inhibitory cells containing cholecystokinin. Asymmetrical (glutamatergic) synapses, however, were always negative for CB1. To investigate the effect of presynaptic CB1 receptor activation on hippocampal inhibition, we recorded inhibitory postsynaptic currents (IPSCs) from principal cells. Bath application of CB1 receptor agonists (WIN55,212-2 and CP55,940) suppressed IPSCs evoked by local electrical stimulation, which could be prevented or reversed by the CB1 receptor antagonist SR141716A. Action potential-driven IPSCs, evoked by pharmacological stimulation of a subset of interneurons, were also decreased by CB1 receptor activation. We also examined the effects of CB1 receptor agonists on Ca2+-independent miniature IPSCs (mIPSC). Both agonists were without significant effect on the frequency or amplitude of mIPSCs. Synchronous gamma oscillations induced by kainic acid in the CA3 region of hippocampal slices were reversibly reduced in amplitude by the CB1 receptor agonist CP 55,940, which is consistent with an action on IPSCs. We used CB1-/- knock-out mice to confirm the specificity of the antibody and of the agonist (WIN55,212-2) action. We conclude that activation of presynaptic CB1 receptors decreases Ca2+-dependent GABA release, and thereby reduces the power of hippocampal network oscillations.

Conventional uptake of neurotrophins takes place at axon terminals via specific receptors, and is followed by retrograde transport. Recent studies demonstrated that, with the exception of nerve growth factor, other neurotrophins may be delivered anterogradely to the region containing the receptor expressing neurons. In this study we used a triple labeling method that combines retrograde tract tracing, in situ hybridization and immunocytochemistry to examine whether non-principal cells projecting from the hippocampus to the septum synthesize nerve growth factor. Our results show that, on average, 59% of the horseradish peroxidase-labeled hippocamposeptal nonpyramidal neurons also display nerve growth factor messenger RNA hybridization signal. The ratio was slightly higher in the CA1 stratum oriens and the hilus of the dentate gyrus (64 and 62%, respectively) compared to stratum oriens of the CA3 region (58%). In addition, we demonstrated that many nerve growth factor-positive septally projecting neurons also contain the calcium-binding protein calbindin D-28K, whereas nerve growth factor-negative projecting cells mostly lack this neurochemical marker. In contrast to nerve growth factor, neurotrophin-3 has never been found in hippocamposeptal cells. Hippocamposeptal GABAergic cells are reciprocally connected with the medial septum, thus they are in a key position to regulate nerve growth factor release as a function of the activity level in the septohippocampal system. Furthermore, our results raise the intriguing possibility that nerve growth factor may be transported also in an anterograde manner. Regardless of the direction of transport, the presence of nerve growth factor in hippocamposeptal cells suggests that long distance fast synaptic mechanisms and slow neurotrophin action are coupled in these neurons.

Cannabinoids have been shown to disrupt memory processes in mammals including humans. Although the CB1 neuronal cannabinoid receptor was identified several years ago, neuronal network mechanisms mediating cannabinoid effects are still controversial in animals, and even more obscure in humans. In the present study, the localization of CB1 receptors was investigated at the cellular and subcellular levels in the human hippocampus, using control post mortem and epileptic lobectomy tissue. The latter tissue was also used for [3H]GABA release experiments, testing the predictions of the anatomical data. Detectable expression of CB1 was confined to interneurons, most of which were found to be cholecystokinin-containing basket cells. CB1-positive cell bodies showed immunostaining in their perinuclear cytoplasm, but not in their somadendritic plasmamembrane. CB1-immunoreactive axon terminals densely covered the entire hippocampus, forming symmetrical synapses characteristic of GABAergic boutons. Human temporal lobectomy samples were used in the release experiments, as they were similar to the controls regarding cellular and subcellular distribution of CB1 receptors. We found that the CB1 receptor agonist, WIN 55,212-2, strongly reduced [3H]GABA release, and this effect was fully prevented by the specific CB1 receptor antagonist SR 141716A. This unique expression pattern and the presynaptic modulation of GABA release suggests a conserved role for CB1 receptors in controlling inhibitory networks of the hippocampus that are responsible for the generation and maintenance of fast and slow oscillatory patterns. Therefore, a likely mechanism by which cannabinoids may impair memory and associational processes is an alteration of the fine-tuning of synchronized, rhythmic population events.

To understand the functional significance and mechanisms of action in the CNS of endogenous and exogenous cannabinoids, it is crucial to identify the neural elements that serve as the structural substrate of these actions. We used a recently developed antibody against the CB1 cannabinoid receptor to study this question in hippocampal networks. Interneurons with features typical of basket cells showed a selective, intense staining for CB1 in all hippocampal subfields and layers. Most of them (85.6%) contained cholecystokinin (CCK), which corresponded to 96.9% of all CCK-positive interneurons, whereas only 4.6% of the parvalbumin (PV)-containing basket cells expressed CB1. Accordingly, electron microscopy revealed that CB1-immunoreactive axon terminals of CCK-containing basket cells surrounded the somata and proximal dendrites of pyramidal neurons, whereas PV-positive basket cell terminals in similar locations were negative for CB1. The synthetic cannabinoid agonist WIN 55,212-2 (0.01-3 microM) reduced dose-dependently the electrical field stimulation-induced [3H]GABA release from superfused hippocampal slices, with an EC50 value of 0. 041 microM. Inhibition of GABA release by WIN 55,212-2 was not mediated by inhibition of glutamatergic transmission because the WIN 55,212-2 effect was not reduced by the glutamate blockers AP5 and CNQX. In contrast, the CB1 cannabinoid receptor antagonist SR 141716A (1 microM) prevented this effect, whereas by itself it did not change the outflow of [3H]GABA. These results suggest that cannabinoid-mediated modulation of hippocampal interneuron networks operate largely via presynaptic receptors on CCK-immunoreactive basket cell terminals. Reduction of GABA release from these terminals is the likely mechanism by which both endogenous and exogenous CB1 ligands interfere with hippocampal network oscillations and associated cognitive functions.

Ascending subcortical pathways effectively modulate hippocampal information processing. Two components, the cholinergic and serotonergic pathways have been demonstrated to play an important role in the generation of behaviour-dependent hippocampal EEG patterns. Several findings suggest that the above projections influence the activity of hippocampal interneurons. Here we review the available data from physiological, pharmacological and receptor localization experiments, drawing attention to the crucial role of interneurons in the transfer and amplification of subcortical effects on cortical information processing. We hypothesize that, by exerting diverse actions on different subsets of interneurons, the cholinergic and serotonergic systems might change the balance of somatic and dendritic inhibition, and consequently change the integrative properties of hippocampal principal cells.

The Ca2+/calmodulin-dependent protein phosphatase 2B or calcineurin (CN) participates in several Ca2+-dependent signal transduction cascades and, thus, contributes to the short and long term regulation of neuronal excitability. By using a specific antibody to CN, we demonstrate its absence from hippocampal interneurons and illustrate a physiological consequence of such CN deficiency. Consistent with the lack of CN in interneurons as detected by immunocytochemistry, the CN inhibitors FK-506 or okadaic acid significantly prolonged N-methyl-D-aspartate channel openings recorded in the cell-attached mode in hippocampal principal cells but not those recorded in interneurons. Interneurons were also devoid of Ca2+/calmodulin-dependent protein kinase IIalpha, yet many of their nuclei contained the cyclic AMP-responsive element binding protein. On the basis of the CN and Ca2+/calmodulin-dependent protein kinase IIalpha deficiency of interneurons, entirely different biochemical mechanisms are expected to govern Ca2+-dependent neuronal plasticity in interneurons versus principal cells.

In the present study, the connectivity between two subcortical nuclei involved in hippocampal theta activity, the supramammillary nucleus and the medial septum-diagonal band of Broca complex, was examined. Targets of the supramammillary afferents in the medial septum-diagonal band of Broca complex were identified by combining anterograde transport of Phaseolus vulgaris leucoagglutinin with immunostaining for putative postsynaptic neurons, i.e. for parvalbumin and choline acetyltransferase that are known to label the GABAergic and cholinergic neurons, respectively, of the medial septum-diagonal band of Broca complex. Double retrograde transport experiments using the tracers horseradish peroxidase and wheat germ agglutinin-conjugated colloidal gold were employed to identify supramammillary neurons that project both to the hippocampus and the medial septum-diagonal band of Broca complex. Phaseolus vulgaris leucoagglutinin injections into the supramammillary nucleus of the rat resulted in dense fibre and terminal labelling in the medial septum-diagonal band of Broca complex. Labelled terminals formed asymmetric synapses mainly on distal dendrites of medial septal neurons. Proximal dendrites and somata were rarely contacted. The supramammillary afferents showed no target selectivity for a particular cell type; they innervated both cholinergic and GABAergic cells. Occasionally, perisomatic, basket-like terminals of supramammillary origin were found around parvalbumin-containing neurons. Double-retrograde experiments revealed that at least 25% of the supramammillo-hippocampal cells also projected to the medial septum-diagonal band of Broca. These data suggest that the nucleus, known to modulate the hippocampal electrical activity directly by the supramammillo-hippocampal pathway, also has the potential for an indirect action via the innervation of both the GABAergic and cholinergic septohippocampal neurons. This dual modulation may originate, at least in part, from the same population of supramammillary neurons.

Interneurons immunoreactive for vasoactive intestinal polypeptide (VIP) are integral elements of columnar organization patterns in the rat cerebral cortex. By application of the sensitive mirror technique, the co-localization of VIP with the classical inhibitory neurotransmitter gamma-aminobutyric acid (GABA) and the acetylcholine-synthesizing enzyme, choline acetyltransferase (ChAT), was investigated in neocortical neurons. Furthermore, the frequency of co-localization of ChAT with GABA was determined. In a sample of 118 VIP-immunoreactive neurons, mostly from the primary somatosensory cortex, it was demonstrated that virtually all of them reveal immunoreactivity for GABA and, therefore, are to be GABAergic. Moreover, 34% of mostly bipolar, VIP-positive neurons contained ChAT and are, thus, supposedly cholinergic as well. Co-localization of VIP and ChAT varied according to cortical laminae. Finally, 88% of a total of 60 ChAT-immunoreactive neurons were also immunostained for GABA. It is concluded that almost all VIP-immunoreactive neurons and most of the cholinergic neurons in rat neocortex represent partly overlapping subpopulations of inhibitory interneurons utilizing GABA.

1. Slices were prepared from rat forebrain to include both the septum and the hippocampus in order to examine the effects of septal stimulation on hippocampal inhibitory circuits. 2. Repetitive stimulation of septo-hippocampal fibres caused a maintained decrease in the frequency of spontaneous IPSPs recorded from CA3 pyramidal cells in the presence of antagonists of excitatory amino acid receptors and of muscarine receptors. 3. In records made from pyramidal cells with CsCl-filled electrodes, IPSPs were examined at potentials both more positive and more negative than their reversal potential. Single septal stimuli hyperpolarized pyramidal cells when IPSPs were depolarizing events and depolarized them when IPSPs were hyperpolarizing. The GABAA receptor antagonist picrotoxin abolished the effects of septal stimulation. 4. Activation of septal afferents initiated an IPSP in hippocampal inhibitory cells but not in pyramidal cells. Septal IPSPs had similar kinetics to those initiated by local hippocampal stimulation and could suppress inhibitory cell discharge. 5. In pyramidal cells recorded with potassium acetate-filled electrodes, septal stimuli initiated a depolarization that increased with the driving force for Cl- and that could cause firing. 6. Rhythmic stimulation of septo-hippocampal fibres at 5 Hz initiated, in the hippocampus, a maintained out-of-phase oscillation of pyramidal cell discharge and inhibitory cell firing, as detected by the occurrence of spontaneous IPSPs. 7. These results suggest that GABAergic septo-hippocampal afferents selectively inhibit hippocampal inhibitory cells and so disinhibit pyramidal cells. This disinhibition could contribute to the transmission of the theta rhythm from the septum to the hippocampus.

A specific antiserum against substance P receptor (SPR) labels nonprincipal neurons in the cerebral cortex of the rat (T. Kaneko et al. [1994], Neuroscience 60:199-211; Y. Nakaya et al. [1994], J. Comp. Neurol. 347:249-274). In the present study, we aimed to identify the types of SPR-immunoreactive neurons in the hippocampus according to their content of neurochemical markers, which label interneuron populations with distinct termination patterns. Markers for perisomatic inhibitory cells, parvalbumin and cholecystokinin (CCK), colocalized with SPR in pyramidallike basket cells in the dentate gyrus and in large multipolar or bitufted cells within all hippocampal subfields respectively. A dense meshwork of SPR-immunoreactive spiny dendrites in the hilus and stratum lucidum of the CA3 region belonged largely to inhibitory cells terminating in the distal dendritic region of granule cells, as indicated by the somatostatin and neuropeptide Y (NPY) content. In addition, SPR and NPY were colocalized in numerous multipolar interneurons with dendrites branching close to the soma. Twenty-five percent of the SPR-immunoreactive cells overlapped with calretinin-positive neurons in all hippocampal subfields, showing that interneurons specialized to contact other gamma-aminobutyric acid-ergic cells may also contain SPR. On the basis of the known termination pattern of the colocalized markers, we conclude that SPR-positive interneurons are functionally heterogeneous and participate in different inhibitory processes: (1) perisomatic inhibition of principal cells (CCK-containing cells, and parvalbumin-positive cells in the dentate gyrus), (2) feedback dendritic inhibition in the entorhinal termination zone (somatostatin and NPY-containing cells), and (3) innervation of other interneurons (calretinin-containing cells).

The spatial synaptic pattern formed by boutons, originating in the ventroposteromedial thalamic nucleus, with GABAergic neurons in the rat barrel cortex was mapped. The aim was to shed light on the structural basis by which inhibitory circuits may be activated at the first stage of cortical information processing. The thalamic afferent projection was labelled by anterograde transport of Phaseolus vulgaris leucoagglutinin (PHA-L), whereas the GABAergic targets in layer IV of the rat barrel cortex was visualized by postembedding GABA immunogold-labelling or by pre-embedding parvalbumin immunocytochemistry. In the first set of experiments, we mapped barrels, contained in single ultrathin sections, by means of a computer-controlled electron microscope stage in their entire layer IV representation. From a total of 1199 asymmetric PHA-L-labelled synapses, only 98 were on GABAergic elements, mainly on dendritic shafts. This corresponded to 8.2% of all synapses counted. These synapses on GABAergic targets were essentially homogeneously distributed without a reliable relationship to barrel subdivisions, i.e., hollow versus wall; or layer IVa versus layer IVb. In the second part of the study, we demonstrated that parvalbumin-containing neurons represent the major GABAergic cell type targetted by thalamic afferents in layer IV of the barrel cortex, since all parvalbumin-positive cells investigated received multiple synaptic contacts (up to eight synapses per neuron) from the ventroposteromedial thalamic nucleus. These results imply that interneurons responsible for perisomatic inhibition (basket and chandelier cells known to contain parvalbumin) are likely to be strongly excited by thalamic afferents, despite the relatively low proportion of thalamic synapses on GABAergic elements compared to spines of principal cells, and participate in the early stages of cortical sensory information processing.

The formation of the GABAergic septohippocampal projection was studied in vitro. Slice cultures of the septal complex from young postnatal rats were prepared and co-cultivated with hippocampal slices for up to four weeks. Then, the anterogradely transported tracer Phaseolus vulgaris leucoagglutinin was injected into the septal culture and the labeled fibers were traced into the hippocampal culture. Some fibers were identified as originating from GABAergic septal cells by double-labeling with an antiserum against GABA using the postembedding immunogold procedure. Our results showed that double-labeled terminals of GABAergic septohippocampal neurons established symmetric synapses exclusively with GABA-positive dendrites in one out of five co-cultures, but also contacted numerous GABA-negative structures in the remaining four co-cultures. These findings, together with light microscopic data from sections double-stained for Phaseolus and parvalbumin, indicate that the high target selectivity of the GABAergic septohippocampal pathway for GABAergic interneurons in vivo is lost in most cases, at least under the present in vitro conditions. It is hypothesized that this may be due to an immaturity of the connection, the lack of axon-guiding factors or an expansion of the septohippocampal GABAergic fibers in the absence of many extrinsic afferents, including GABAergic fibers. The simultaneous occurrence of anterogradely labeled, but GABA-negative, septohippocampal terminals in the hippocampal target culture also suggests that the septohippocampal cholinergic projection developed in vitro, as was shown before in other studies. Since most septohippocampal neurons have to be axotomized for culture preparation, the present results indicate that GABAergic septohippocampal neurons from young postnatal rats survive axotomy and are capable of regenerating a septohippocampal projection, including the formation of characteristic GABAergic synapses on co-cultured hippocampal neurons. However, the characteristic target selectivity is rarely preserved.

The termination pattern of hippocamposeptal nonpyramidal cells was investigated by injecting Phaseolus vulgaris leucoagglutinin (PHAL) into stratum oriens of the CA1 region. Electron microscopic analysis showed that the majority of the anterogradely labeled boutons formed symmetric synapses with dendrites and occasionally with cell bodies located in the medial septum-diagonal band of Broca complex. We have demonstrated with postembedding GABA immunocytochemistry that the majority of PHAL-labeled axon terminals were GABAergic. The neurochemical character of the postsynaptic target cells was also investigated using immunocytochemical double staining. Our data showed that the majority of the labeled hippocamposeptal axons innervated parvalbumin-immunoreactive cells representing GABAergic projection neurons, and a smaller number of contacts were found on ChAT-positive neurons. Septohippocampal neurons identified by retrograde HRP transport were also shown to reactive direct hippocamposeptal input. According to recent results, the lateral septum is unlikely to relay the hippocampal feedback to the medial septum; therefore, the direct hippocampal projection to the medial septum, arising from GABAergic nonpyramidal cells, seems to be the only feedback pathway to the area containing septohippocampal neurons. A novel circuit diagram, based on our recent morphological-immunocytochemical findings, is shown for the synaptic organization of the septo-hippocampo-septal loop. We suggest that the GABAergic hippocamposeptal feedback controls the activity of septal (mostly GABAergic) projection neurons as a function of hippocampal synchrony. The newly discovered reciprocal interactions may give a better insight into septohippocampal physiology.

The basal forebrain projection to the cerebral cortex was studied in the lizard Podarcis hispanica by anterograde transport of Phaseolus vulgaris leucoagglutinin. After injections of the lectin into the septal-basal forebrain area, Phaseolus vulgaris leucoagglutinin-labelled fibres were mainly detected in the outer plexiform layer of the medial cortex and in the inner plexiform layer of the dorsal and dorsomedial cortices. Ultrathin sections from these areas were obtained and processed for postembedding immunogold staining for GABA. Most of the Phaseolus vulgaris leucoagglutinin-labelled boutons in the dorsal and dorsomedial cortical areas were GABA immunoreactive and all the double-labelled boutons established symmetric synaptic contacts on cell bodies and dendrites that were also found to be GABA immunoreactive in all cases. In contrast, Phaseolus vulgaris leucoagglutinin-labelled varicosities in the outer plexiform layer of the medial cortex made asymmetric synaptic contacts on GABA-immunonegative profiles and they were themselves negative for GABA. In double-labelled sections, GABA-, calbindin D28k- and neuropeptide Y-immunoreactive neurons were found to be innervated by multiple Phaseolus vulgaris leucoagglutinin-labelled varicosities in the dorsal and dorsomedial cortical areas, whereas in the medial cortex Phaseolus vulgaris leucoagglutinin-labelled fibres were not observed in contact with any subpopulation of GABAergic cells. The results demonstrate that in lizards the septal-basal forebrain projection to the cortex has a GABAergic component, which selectively terminates on GABAergic non-pyramidal cells including the neuropeptide Y- and the calbindin D28k-containing subpopulations. This synaptic organization is remarkably similar to that in mammals, and suggests that the mechanisms of control of the cortical activity by the basal forebrain have been highly preserved during phylogeny.

The types and distribution of cells containing three calcium-binding proteins, calretinin, calbindin D28K, and parvalbumin, have been studied by immunocytochemistry in different areas of the cerebral cortex of lizards. Cross-reactivity of the antisera has been excluded by demonstrating the existence of several cell groups immunoreactive for one but not the other two calcium-binding proteins. In the dorsal and dorsomedial cortices all three proteins coexist in a single subpopulation of gamma-aminobutyric acid (GABA)ergic neurons, the terminals of which form pericellular baskets around cell bodies of bipyramidal neurons. The somata of these neurons are largely restricted to the cellular and inner plexiform layers, but the dendrites usually penetrate all layers, allowing the neurons to sample input from all possible sources. A small number of parvalbumin-containing neurons in the outer plexiform layer do not contain the other two proteins. The medial cortex, which is likely to be homologous to the mammalian dentate gyrus, only contains parvalbumin-immunoreactive neurons. The dendritic trees of these cells appear to avoid the Timm-positive fields receiving input from zinc-rich fiber collaterals, originating from principal cells. The lateral cortex contains calbindin D28K-immunoreactive GABAergic neurons, which lack the other two calcium-binding proteins. These neurons have horizontally running dendrites in the outer plexiform layer, but their axon terminals could not be visualized. The present study uncovered important similarities and differences between the lizard and the mammalian archicortex in the types of neurons containing calcium-binding proteins. As in mammals, different cell types evolved in the lizard to inhibit the perisomatic versus the distal dendritic region of principal cells, the calcium-binding protein-containing neurons being responsible for the former, and neuropeptide-containing neurons for the latter. The results also suggest that further neurochemical diversion of GABAergic interneurons coupled to a functional specialization took place during phylogenetic development from reptiles to mammals.

Calbindin D28k-containing non-pyramidal cells were found in all layers and subfields of the hippocampus, with the highest frequency in stratum radiatum of the CA1-CA3 subfields. A large number of these neurons had a vertically oriented dendritic tree, often restricted to to stratum radiatum. In stratum oriens and near to the border of strata radiatum and lacunosum moleculare cells with horizontally running dendrites were also found. Multipolar cells were most common in stratum radiatum of the CA3 region. The GABAergic nature of the calbindin D28k-containing non-pyramidal cells was studied using the "mirror" technique. Adjacent thick sections were immunostained for calbindin D28k and GABA, and halved neurons were identified on the common surfaces. The majority of calbindin D28k-containing non-pyramidal cells were shown to be GABAergic. The GABA-negative calbindin cells were found in relatively large numbers in stratum oriens of the CA1-CA3 region, and occasionally in strata radiatum and pyramidale of CA2, and in stratum radiatum of the CA3c region near to the border of the dentate hilus. However, even in these cells a weak immunostaining, only slightly but consistently above background level, was always observed. Earlier studies have demonstrated that the somata of GABAergic neurons with distant projections may contain a level of GABA that is below the detection threshold of immunocytochemistry. Here we provide direct evidence that the calbindin-containing non-pyramidal cells were among those projecting to the medial septum. Following horseradish peroxidase injections into the medial septum 80% of the retrogradely labelled non-pyramidal cells were found to be immunoreactive for calbindin D28k, and 20% contained neuropeptide Y. These results suggest that the calbindin D28k-containing and apparently GABA-immunonegative non-pyramidal cells in stratum oriens of the CA1-CA3 regions may also be GABAergic, but have a distant projection, that is, to the medial septum.

The basal forebrain-neocortex pathway--involved in higher cognitive processing, selective attention, and arousal--is considered one of the functionally most important ascending subcortical projections. The mechanism by which this relatively sparse subcortical pathway can control neuronal activity patterns in the entire cortical mantle is still unknown. The present study in the cat provides evidence that gamma-aminobutyric acid-containing basal forebrain neurons participate in the neocortical projection and establish multiple synaptic connections with gamma-aminobutyric acid-releasing interneurons containing somatostatin or parvalbumin. We propose that a mechanism by which the numerically small ascending pathways can exert a powerful global effect in the neocortex is by the selective innervation of gamma-aminobutyric acid-releasing interneurons, which, in turn, control the activity of large populations of pyramidal cells through their extensive axon arborizations. Finally, these results demonstrate a direct anatomical link between two cell populations implicated in Alzheimer disease pathology: basal forebrain neurons and cortical somatostatin cells.

Lesions causing loss of hippocampal theta activity have been shown to result in spatial memory deficits in rats. On the other hand, hippocampal theta activity is thought to be associated only with motor activity, and its role in learning/memory is not clear. Vigabatrin, an inhibitor of GABA-tranasminase, causes elevation of brain GABA levels. Previously, we have found that subchronic administration of vigabatrin did not impair spatial learning/memory in a water maze task. This experiment was carried out to further examine the hippocampal effects of vigabatrin by studying whether vigabatrin at antiepileptic doses affects mobility-related hippocampal EEG. Administration of vigabatrin (100 mg/kg or 500 mg/kg, IP) to nonepileptic rats caused no significant changes in mobility-related rhythmic theta activity, and the relative spectral power of theta frequency had a slight increasing tendency. These results suggest that the vigabatrin-induced enhanced GABAergic inhibition does not disturb normal mobility-related hippocampal theta activity.

The possible co-existence of calretinin with other calcium binding proteins, parvalbumin and calbindin D28k, and with GABA, was studied in non-pyramidal cells of the rat dorsal hippocampal formation, using the mirror technique. The majority of the calretinin-containing neurons (83%) were found to be immunoreactive for GABA (79% in the dentate gyrus, 84% in the CA2-3, and 88% in the CA1 subfield). Most of the GABA-negative calretinin-immunoreactive neurons were located in the hilus of the dentate gyrus and in stratum lucidum of the CA3 subfield. Detailed analysis of the calretinin-immunoreactive cells of these subfields revealed that the two morphologically distinct types of calretinin neurons, i.e. the spiny and the spine-free cells, differ in their immunoreactivity for GABA. The overwhelming majority (92%) of the spine-free neurons were GABA-positive, whereas the immunoreactivity of spiny cells was ambiguous. At the sensitivity threshold of the immunocytochemical techniques used in the present study, most of the spiny cells (89%) had to be considered as GABA-negative, although the staining intensity in their cell bodies was somewhat above background level. Colchicine treatment resulted in a degeneration of calretinin-immunoreactive neurons; therefore, its effect on the GABA content of spiny neurons could not be evaluated. Nevertheless, the observations suggest that calretinin-containing neurons are heterogeneous both morphologically and neurochemically. Examination of the co-existence of calcium binding proteins revealed that none of the hippocampal cells contained both calretinin and parvalbumin in any regions of the hippocampal formation. Some overlap was detected between the calretinin- and the calbindin D28k-containing cell populations, 5.1% of the former and 6.2% of the latter were immunoreactive for both calcium binding proteins. This may be due to a small degree of cross-reactivity of the calbindin D28k antiserum with calretinin. Thus, our results demonstrate that the majority of calretinin-immunoreactive neurons are GABAergic and represent a subpopulation of non-pyramidal cells with no or only a negligible overlap with the subpopulations containing the other calcium binding proteins, parvalbumin and calbindin.

Postsynaptic targets of the GABAergic septohippocampal and the serotonergic raphe-hippocampal pathways were studied using anterograde tracing with Phaseolus vulgaris leucoagglutinin combined with pre- and postembedding immunocytochemistry in the rat. Two types of afferents were labeled in the hippocampus and dentate gyrus from the medial septum-diagonal band of Broca complex, one with large diameter varicosities and another with smaller terminals. The former type was shown to be immunoreactive for gamma-aminobutyric acid (GABA), and to innervate predominantly GABA-immunoreactive interneurons. Subsequently, these target interneurons were demonstrated to include all subpopulations of GABAergic cells which could be visualized by antisera against parvalbumin, calbindin D28k, calretinin, cholecystokinin, somatostatin, neuropeptide Y and vasoactive intestinal polypeptide. These types of interneurons have different afferent and efferent connections, and thus participate in different inhibitory processes in the hippocampal formation. The other subcortical pathway, the serotonergic projection from the median raphe nucleus, was also shown to establish synapses predominantly with GABAergic interneurons both in the hippocampus and in the dentate gyrus. In contrast to the septohippocampal projection, this pathway did not innervate all types of GABAergic neurons. They selected a particular subpopulation, i.e. those which contain calbindin D28k, and ignored those which contained parvalbumin or the other neurochemical markers. This suggests a strong functional specialization among local inhibitory circuits, as well as among the subcortical afferents originating in the septum and raphe. These findings suggest that a mechanism by which numerically small afferent pathways may have a profound global effect on the electrical activity of the hippocampal formation is the selective innervation of local interneurons. These GABAergic inhibitory cells, in turn, control the activity of large populations of principal cells. The level of GABAergic inhibition determines the degree of population synchrony and influences N-methyl-D-aspartate receptor-mediated epileptiform burst-firing. Thus, the specific subcortical modulation of hippocampal inhibitory circuits may also have fundamental implications for epileptogenesis.

The arborization pattern and postsynaptic targets of the GABAergic component of the basal forebrain projection to neo- and mesocortical areas have been studied by the combination of anterograde tracing and pre- and postembedding immunocytochemistry. Phaseolus vulgaris leucoagglutinin (PHAL) was iontophoretically delivered into the region of the diagonal band of Broca, with some spread of the tracer into the substantia innominata and ventral pallidum. A large number of anterogradely labelled varicose fibres were visualized in the cingulate and retrosplenial cortices, and a relatively sparse innervation was observed in frontal and occipital cortical areas. Most of the labelled axons were studded with large en passant varicosities (Type 1), whereas the others (Type 2) had smaller boutons often of the drumstick type. Type 1 axons were distributed in all layers of the mesocortex with slightly lower frequency in layers 1 and 4. In the neocortex, layer 4, and to a smaller extent upper layer 5 and layer 6 contained the largest number of labelled fibres, whereas only a few fibres were seen in the supragranular layers. Characteristic type 2 axons were very sparse but could be found in all layers. Most if not all boutons of PHAL-labelled type 1 axons were shown to be GABA-immunoreactive by immunogold staining for GABA. Altogether 73 boutons were serially sectioned and found to make symmetrical synaptic contacts mostly with dendritic shafts (66, 90% of total targets), cell bodies (6, 8.2% of total), and with one spine. All postsynaptic cell bodies, and the majority of the dendritic shafts (44, 60.3% of total targets) were immunoreactive for GABA. Thus at least 68.5% of the total targets were GABA-positive, but the majority of the dendrites not characterized immunocytochemically for technical reasons (15.1%) also showed the fine structural characteristics of nonpyramidal neurons. The target interneurons included some of the somatostatin- and calbindin-containing subpopulations, and a small number of parvalbumin-containing neurons, as shown by double immunostaining for PHAL and calcium-binding proteins or neuropeptides. We suggest that the innervation of inhibitory interneurons having extensive local axon arborizations may be a mechanism by which basal forebrain neurons-most notably those containing GABA--have a powerful global effect on the majority of principal cells in the entire cortical mantle.

The possible coexistence of calbindin D28k with parvalbumin and of calbindin D28k with cholecystokinin was studied in nonpyramidal cells of the rat dorsal hippocampal formation. Neighbouring Vibratome sections were immunostained either for calbindin D28k and parvalbumin or for calbindin D28k and cholecystokinin. The cells, halved during sectioning, were identified in both sections immunostained for different antigens. The coexistence of calbindin D28k and parvalbumin in the same neuron was rare throughout the hippocampal formation with the exception of stratum oriens of the CA1 region, where 9.6% of the parvalbumin-immunoreactive cells also contained calbindin D28k. In stratum radiatum of the CA3 region, calbindin D28k and cholecystokinin coexisted in 12.5% and 21.2% of the calbindin D28k and cholecystokinin-immunoreactive cells, respectively. In other regions of the hippocampal formation, the two markers coexisted in less than 5% of the cells of either type. The present results demonstrate that calbindin D28k-, parvalbumin- and cholecystokinin-containing nonpyramidal cells represent largely nonoverlapping cell populations and may thus be involved in different inhibitory circuits.

Extensive electrical stimulation of the perforant pathway input to the hippocampus results in a characteristic pattern of neuronal death, which is accompanied by an impairment of cognitive functions similar to that seen in human temporal lobe epilepsy. The excitotoxic hypothesis of epileptic cell death [Olney, J. W. (1978) in Kainic Acid as a Tool in Neurobiology, eds. McGeer, E., Olney, J. W. & McGeer, P. (Raven, New York), pp. 95-121; Olney, J. W. (1983) in Excitotoxins, eds. Fuxe, K., Roberts, P. J. & Schwartch, R. (Wenner-Gren International Symposium Series, Macmillan, London), Vol. 39, pp. 82-96; and Rothman, S. M. & Olney, J. W. (1986) Ann. Neurol. 19, 105-111] predicts an imbalance between excitation and inhibition, which occurs probably as a result of hyperactivity in afferent pathways or impaired inhibition. In the present study, we investigated whether the enhancement of gamma-aminobutyric acid (GABA)-mediated (GABAergic) inhibition of neurotransmission by blocking the GABA-metabolizing enzyme, GABA transaminase, could influence the histopathological and/or the behavioral outcome in this epilepsy model. We demonstrate that the loss of pyramidal cells and hilar somatostatin-containing neurons can be abolished by enhancing the level of synaptically released GABA, and that the preservation of hippocampal structure is accompanied by a significant sparing of spatial memory as compared with placebo-treated controls. These results suggest that enhanced GABAergic inhibition can effectively block the pathophysiological processes that lead to excitotoxic cell death and, as a result, protect the brain from seizure-induced cognitive impairment.

The septohippocampal projection was visualized in three Macaca mulatta monkeys by anterograde transport of Phaseolus vulgaris leucoagglutinin. Following injections of the lectin into the medial septal nucleus, P. vulgaris leucoagglutinin-labelled fibres were found in the hippocampal complex, mainly in stratum oriens of the CA1 subfield, throughout the CA3 subfield, and in the hilus and stratum moleculare of the dentate gyrus. The majority of labelled axons were varicose, and formed multiple contacts with cell bodies and dendrites of calbindin D28k- and parvalbumin-immunoreactive non-pyramidal cells. GABA immunoreactivity of P. vulgaris leucoagglutinin-labelled axons and their postsynaptic targets was investigated by sectioning varicose axon segments for correlated light and electron microscopy, and processing alternate ultrathin sections for postembedding immunogold staining for GABA. All P. vulgaris leucoagglutinin-labelled boutons examined were GABA-immunoreactive and the majority of them formed symmetrical synapses with GABA-immunoreactive cell bodies and dendrites. The results demonstrate that a GABAergic septohippocampal pathway exists in the monkey, and, similar to the rat, terminates on different types of GABAergic neurons, including the parvalbumin- and calbindin D28k-containing non-pyramidal cells.

The termination pattern of septohippocampal axons visualized by anterograde transport of Phaseolus vulgaris leucoagglutinin was studied in the hippocampal formation in the rat, with special reference to the innervation of neurons immunoreactive for the neuroactive peptides cholecystokinin, somatostatin or vasoactive intestinal polypeptide. The type I, GABAergic, septohippocampal afferents were shown to terminate on neurons immunoreactive for each of the three peptides. The cholecystokinin-like immunoreactive neurons in all regions, and the somatostatin-immunoreactive cells in stratum oriens of CA1 region were the most preferred targets. Cholecystokinin-immunoreactive cells, especially those in the granule cell layer of the dentate gyrus, were often seen to be contacted by type II (presumed cholinergic) axons as well. The somatostatin-immunoreactive cells in the hilus were also innervated by type I septohippocampal axons, although less frequently than those in stratum oriens of the CA1 subfield. Each type of peptidergic neuron received multiple symmetrical synaptic input from the Phaseolus vulgaris leucoagglutinin-labelled septal afferents, as confirmed by correlated electron microscopy. The majority of these neuropeptide-containing cells are known to be GABAergic, and to have distinct input and output relationships. Thus, the present results demonstrate that the GABAergic septohippocampal pathway can control a wide range of putative inhibitory circuits, and thereby influence the pattern of electrical activity in the hippocampal formation.

Septal axons were visualized in the hippocampus by anterograde transport of Phaseolus vulgaris leucoagglutinin (PHAL) injected into the medial septal region, which contains large numbers of parvalbumin (PV)-immunoreactive somata. A proportion of the PHAL-labelled afferents in the hippocampus were shown to be immunoreactive for PV by immunostaining alternate sections. This population of septohippocampal axons alone was found to be immunoreactive for gamma-aminobutyric acid (GABA) following post-embedding immunogold staining of ultrathin sections cut from the same material.

The monosynaptic targets of different functional types of geniculocortical axons were compared in the primary visual cortex of monkeys. Single thalamocortical axons were recorded extracellularly in the white matter by using horseradish-peroxidase-filled pipettes. Their receptive fields were mapped and classified as corresponding to those of parvi- or magnocellular neurons in the lateral geniculate nucleus. The axons were then impaled and injected intraaxonally with horseradish peroxidase. Two magnocellular (MA) and two parvicellular (PA) axons were successfully recovered and reconstructed in three dimensions. The two MA axons arborised mainly in layer 4C alpha, as did the two PA axons in layer 4C beta. Few collaterals formed varicosities in layer 6. Both MA axons had two large, elongated clumps of bouton (approx. 300-500 x 600-1,200 microns each) and a small clump. One PA axon had two clumps (each with a core appr. 200 microns in diameter); the other had only one (appr. 150-200 microns in axon had 1,380; one MA axon had 3,200 boutons; and those of the more extensive MA axon were not counted. The distribution of postsynaptic targets as well as the number of synapses per bouton has been established for a sample of 150 PA boutons and 173 MA boutons from serial ultrathin sections. The MA axons made on average 2.1 synapses per bouton compared to 1.79 for one PA axon and 2.6 for the other. The sample of boutons taken from the two physiological types of axons contacted similar proportions of dendritic spines (52-68%), shafts (33-47%), and somata (0-3%). The postsynaptic elements were further characterized by immunostaining for GABA. All postsynaptic perikarya and some of the dendrites (4.5-9.5% of all targets) were positive for the amino acid. Near the thalamic synapse GABA-negative dendritic shafts frequently contained lamellar bodies, an organelle identical in structure to spine apparatus. Dendritic shafts and spines postsynaptic to the thalamocortical boutons frequently received an adjacent synapse from GABA-immunoreactive boutons. The similarity between the magno-and parvicellular axons in their targeting of postsynaptic elements, including the GABAergic neurons, suggests that the structural basis of the physiological differences between 4C alpha and 4C beta neurons should be sought in other aspects of the circuitry of layer 4C, such as local cortical circuits, or in the far greater horizontal extent of the thalamocortical and GABAergic axons in layer 4C alpha compared to those in the beta subdivision.

Labelled neurons were identified by autoradiography following tangential intracortical injection of [3H]-gamma-aminobutyrate (GABA). The addition of cis-1,3-aminocyclohexane carboxylic acid to the GABA solution prevented perikaryal labelling. Labelled neurons were found in each injected layer and in addition they were always present directly above the injection track. The labelling of neurons in layer II, and upper III, following injections in layers V. and VI. can be explained by retrograde axonal transport and indicates that some GABA-ergic neurons project vertically. Ninety neurons of different types were Golgi impregnated and examined for selective [3H]-GABA uptake. Sixteen of these were labelled. On the basis of dendritic characteristics they were classified as aspiny multipolar neurons with small, medium or large dendritic fields, sparsely spiny multipolar neurons and one neuron was a bipolar cell. Thus Golgi impregnation of their processes reveals that cortical GABA-ergic neurons are a heterogeneous population. A [3H]-GABA accumulating, aspiny neuron with profoundly branching, "bushy" dendrites and locally arborizing axon in layer VI, was studied in the electron microscope. Its fine structural characteristics were similar to those of other identified non-pyramidal neurons. The existence of several types of cortical GABA-ergic neurons differing in their synaptic connections is discussed.

Solid pieces of embryonic hippocampal tissue were implanted in a cavity formed by aspiration of the fimbria-fornix and the overlying cingulate cortex in adult rats. Six to 8 months after the transplantation, chronic recording electrodes were implanted into the graft and the host hippocampi for the recording of electroencephalogram and unit activity in the freely moving animal. Irregularly occurring sharp waves or electroencephalogram spikes and concurrent synchronous discharge of large groups of neurons dominated the electrical activity of the grafts, in contrast to the situation in normal animals. Light microscopy and GABA immunocytochemistry in the grafts revealed that the three major cell types of the hippocampal formation, i.e. pyramidal neurons, dentate granule cells and GABA-immunoreactive interneurons were present in the hippocampal grafts. At the ultrastructural level, however, significant alterations in connectivity were observed. The most striking finding was the absence or sparse occurrence of synapses on the axon initial segments of pyramidal neurons. The axon initial segments are normally densely covered by GABAergic synapses derived from a specialized type of interneuron, the chandelier or axo-axonic cell. On the other hand, numerous GABA-immunoreactive terminals were found in synaptic contact with somata of pyramidal neurons, suggesting that other types of GABAergic interneurons and their efferent connections may have developed in a normal manner. The cell bodies of pyramidal neurons received, in addition, several asymmetric synapses from GABA-negative terminals. These presumably excitatory synapses are not present on the somata of pyramidal cells in the normally developing hippocampus. We hypothesize that the somatic excitatory synapses originate, at least in part, from the axon collaterals of the neighbouring pyramidal cells in the graft. We suggest that the hyperexcitability of the neuronal circuitry within the graft is due to reduced inhibition (lack of axo-axonic synapses) coupled with increased collateral excitation of the pyramidal neurons.

Chandelier or axo-axonic cells (AACs) are specialized interneurons terminating on the axon initial segments of pyramidal neurons. Two AACs have been localized by Golgi impregnation, one in the CA1 region of the hippocampus and one in the visual cortex of cat, for structural analysis and for the identification of their transmitter. They had 323 and 268 terminal bouton rows, respectively, probably making synapses with an equal number of initial segments. The distribution of the dendrites of the hippocampal cell was strikingly similar to that of pyramidal cells suggesting a similar input. Using an antiserum to GABA and postembedding GABA-immunocytochemistry, developed for Golgi-impregnated neurons, both cells were found to be GABA-immunoreactive. The strategic location of their synapses and the presence of GABA in AACs suggest that in normal cortical tissue they play a major role in GABA-mediated inhibition.

3H-GABA was injected into restricted regions of visual areas 1 and 2 (cortical areas 17 and 18) on the lateral surface of the occipital lobe in monkeys. The injected tissue was processed for Golgi impregnation and gold toning. Sections containing Golgi-impregnated neurons were re-embedded, sectioned at 1 micron, and prepared for autoradiography to reveal neurones that had selectively accumulated 3H-GABA. Golgi-impregnated pyramidal, spiny stellate and aspiny nonpyramidal neurons were examined for 3H-GABA accumulation. Out of 47 aspiny non-pyramidal neurons 16 were labelled by 3H-GABA. The other cell types did not accumulate the amino acid. Twelve of the labelled neurons were drawn. Eight were bitufted neurons with their dendrites oriented predominantly radially, three were small multipolar neurons, and one could be reconstructed only partially. One neuron had a locally arborizing axon in layer III. Two bitufted, Golgi-impregnated neurons in layer II and upper III of area 18 were labelled from GABA injection radially beneath in layer VI, providing evidence for earlier suggestions that in the monkey's visual cortex the cells in the upper layers which project radially and accumulate 3H-GABA are aspiny non-pyramidal cells. The results indicate the existence of different types of putative GABA-ergic interneurons.

The precise location of physiologically identified specific afferent input on the different types of cell in the visual cortex and the identification of the neurotransmitters of these cells are essential to a better understanding of the first stage of cortical processing. A combination of anatomical, neurochemical, and physiological methods was used to identify the cortical neurones that receive synaptic input from X- and Y-type afferents, which are thought to originate from cells of the lateral geniculate nucleus. One method relied on chance contacts made between single physiologically characterised axons, which had been injected with horseradish peroxidase (HRP), and the processes of cells impregnated by the Golgi method. These experiments revealed that both X and Y axons formed synapses on the dendrites of spiny stellate cells in layer 4. Y axons in both areas 17 and 18 established multiple synaptic contacts on basal dendrites of layer 3 pyramidal cells. One X axon contacted the apical dendrite of a layer 5 pyramidal cell and one Y axon contacted the dendrite of a large cell with smooth dendrites in layer 3. The maximum number of synapses made between one axon and a single postsynaptic cell was eight, although in most cases it was only one. It was concluded that one axon only provides a small fraction of the geniculate afferent input to an individual cell. A second method revealed that the somata in layer 4 in synaptic contact with the HRP-filled axon terminals were GABA-immunoreactive, and therefore might be involved in inhibitory processes. From light microscopic data it was found that somata receiving contacts from X axons in area 17 were significantly smaller (average diameter 15 microns) than those contacted by the Y axons in areas 17 and 18 (average diameter 24 microns). Somatic contacts were extremely rare in layer 6. These data show that the X and Y afferents may activate separate subsets of inhibitory neurones.

Several lines of evidence suggest that gamma-aminobutyric acid is an inhibitory neurotransmitter in the cerebral cortex. To study the intracortical projection of neurons that selectively accumulate this amino acid, we injected radioactive gamma-aminobutyric acid into the upper layers of the striate cortex of monkeys along tracks at an oblique angle to the pia. Sections from the injected area were then processed by a combination of autoradiography and Golgi impregnation to reveal the distribution of labeled neurons and their morphological characteristics. Labeled neurons always occurred around the injection site in each layer. In addition, a consistent radial pattern of perikaryal labeling was observed in layers IVc-VI below the injection track in layers I-IVa. The closer the injection track was to the pia the deeper the peak density of labeled cells appeared. After injection in layers IVa and the lower part of III, the highest number of labeled neurons was in layer IVc; after injection in the upper part of layer III, most labeled neurons were in layer V; and, after injection in layers I and II, the proportion of labeled neurons increased in the lower part of layer V and in layer VI. All these neurons in the infragranular layers are presumably labeled by retrograde axonal transport via the labeled fiber bundles that extended from upper to lower layers. Thirty-four Golgi-stained neurons of various types were also examined for retrograde labeling. Two were labeled, and both were aspiny stellate cells in layer V. The arrangement of these putative GABAergic neurones, with axons that ascend from lower to upper layers in a regular pattern and arborize locally, would enable them to mediate inhibition within cortical columns and between neighboring columns.