Synaptic connections of axo-axonic (chandelier) cells in human epileptic temporal cortex

Specific and Plastic: Chandelier Cell-to-Axon Initial Segment Connections in Shaping Functional Cortical Network

Axon initial segment (AIS) is the most excitable subcellular domain of a neuron for action potential initiation. AISs of cortical projection neurons (PNs) receive GABAergic synaptic inputs primarily from chandelier cells (ChCs), which are believed to regulate action potential generation and modulate neuronal excitability. As individual ChCs often innervate hundreds of PNs, they may alter the activity of PN ensembles and even impact the entire neural network. During postnatal development or in response to changes in network activity, the AISs and axo-axonic synapses undergo dynamic structural and functional changes that underlie the wiring, refinement, and adaptation of cortical microcircuits. Here we briefly introduce the history of ChCs and review recent research advances employing modern genetic and molecular tools. Special attention will be attributed to the plasticity of the AIS and the ChC-PN connections, which play a pivotal role in shaping the dynamic network under both physiological and pathological conditions.

Chandelier cells shine a light on the formation of GABAergic synapses

Uncovering the wiring rules employed by neurons during development represents a formidable challenge with important repercussions for neurodevelopmental disorders. Chandelier cells (ChCs) are a singular GABAergic interneuron type, with a unique morphology, that have recently begun to shed light on the rules that drive the formation and plasticity of inhibitory synapses. This review will focus on the wealth of recent data charting the emergence of synapses formed by ChCs onto pyramidal cells, from the molecules involved to the plasticity of these connections during development.

Differential effects of group III metabotropic glutamate receptors on spontaneous inhibitory synaptic currents in spine-innervating double bouquet and parvalbumin-expressing dendrite-targeting GABAergic interneurons in human neocortex

Diverse neocortical GABAergic neurons specialize in synaptic targeting and their effects are modulated by presynaptic metabotropic glutamate receptors (mGluRs) suppressing neurotransmitter release in rodents, but their effects in human neocortex are unknown. We tested whether activation of group III mGluRs by L-AP4 changes GABAA receptor-mediated spontaneous inhibitory postsynaptic currents (sIPSCs) in 2 distinct dendritic spine-innervating GABAergic interneurons recorded in vitro in human neocortex. Calbindin-positive double bouquet cells (DBCs) had columnar "horsetail" axons descending through layers II-V innervating dendritic spines (48%) and shafts, but not somata of pyramidal and nonpyramidal neurons. Parvalbumin-expressing dendrite-targeting cell (PV-DTC) axons extended in all directions innervating dendritic spines (22%), shafts (65%), and somata (13%). As measured, 20% of GABAergic neuropil synapses innervate spines, hence DBCs, but not PV-DTCs, preferentially select spine targets. Group III mGluR activation paradoxically increased the frequency of sIPSCs in DBCs (to median 137% of baseline) but suppressed it in PV-DTCs (median 92%), leaving the amplitude unchanged. The facilitation of sIPSCs in DBCs may result from their unique GABAergic input being disinhibited via network effect. We conclude that dendritic spines receive specialized, diverse GABAergic inputs, and group III mGluRs differentially regulate GABAergic synaptic transmission to distinct GABAergic cell types in human cortex.

Toward Understanding the Diverse Roles of Perisomatic Interneurons in Epilepsy

Epileptic seizures are associated with excessive neuronal spiking. Perisomatic γ-aminobutyric acid (GABA)ergic interneurons specifically innervate the subcellular domains of postsynaptic excitatory cells that are critical for spike generation. With a revolution in transcriptomics-based cell taxonomy driving the development of novel transgenic mouse lines, selectively monitoring and modulating previously elusive interneuron types is becoming increasingly feasible. Emerging evidence suggests that the three types of hippocampal perisomatic interneurons, axo-axonic cells, along with parvalbumin- and cholecystokinin-expressing basket cells, each follow unique activity patterns in vivo, suggesting distinctive roles in regulating epileptic networks.

Chandelier Cells in Functional and Dysfunctional Neural Circuits

Chandelier cells (ChCs; also called axo-axonic cells) are a specialized GABAergic interneuron subtype that selectively innervates pyramidal neurons at the axon initial segment (AIS), the site of action potential generation. ChC connectivity allows for powerful yet precise modulation of large populations of pyramidal cells, suggesting ChCs have a critical role in brain functions. Dysfunctions in ChC connectivity are associated with brain disorders such as epilepsy and schizophrenia; however, whether this is causative, contributory or compensatory is not known. A likely stumbling block toward mechanistic discoveries and uncovering potential therapeutic targets is the apparent lack of rudimentary understanding of ChCs. For example, whether cortical ChCs are inhibitory or excitatory remains unresolved, and thus whether altered ChC activity results in altered inhibition or excitation is not clear. Recent studies have shed some light onto this excitation-inhibition controversy. In addition, new findings have identified preferential cell-type connectivities established by cortical ChCs, greatly expanding our understanding of the role of ChCs in the cortical microcircuit. Here we aim to bring more attention to ChC connectivity to better understand its role in neural circuits, address whether ChCs are inhibitory or excitatory in light of recent findings and discuss ChC dysfunctions in brain disorders.

Molecular and Electrophysiological Characterization of GABAergic Interneurons Expressing the Transcription Factor COUP-TFII in the Adult Human Temporal Cortex

Transcription factors contribute to the differentiation of cortical neurons, orchestrate specific interneuronal circuits, and define synaptic relationships. We have investigated neurons expressing chicken ovalbumin upstream promoter transcription factor II (COUP-TFII), which plays a role in the migration of GABAergic neurons. Whole-cell, patch-clamp recording in vitro combined with colocalization of molecular cell markers in the adult cortex differentiates distinct interneurons. The majority of strongly COUP-TFII-expressing neurons were in layers I–III. Most calretinin (CR) and/or cholecystokinin- (CCK) and/or reelin-positive interneurons were also COUP-TFII-positive. CR-, CCK-, or reelin-positive neurons formed 80%, 20%, or 17% of COUP-TFII-positive interneurons, respectively. About half of COUP-TFII-/CCK-positive interneurons were CR-positive, a quarter of them reelin-positive, but none expressed both. Interneurons positive for COUP-TFII fired irregular, accommodating and adapting trains of action potentials (APs) and innervated mostly small dendritic shafts and rarely spines or somata. Paired recording showed that a calretinin-/COUP-TFII-positive interneuron elicited inhibitory postsynaptic potentials (IPSPs) in a reciprocally connected pyramidal cell. Calbindin, somatostatin, or parvalbumin-immunoreactive interneurons and most pyramidal cells express no immunohistochemically detectable COUP-TFII. In layers V and VI, some pyramidal cells expressed a low level of COUP-TFII in the nucleus. In conclusion, COUP-TFII is expressed in a diverse subset of GABAergic interneurons predominantly innervating small dendritic shafts originating from both interneurons and pyramidal cells.

The distribution of chandelier cell axon terminals that express the GABA plasma membrane transporter GAT-1 in the human neocortex

Chandelier cells represent a unique type of cortical GABAergic interneuron whose axon terminals (Ch-terminals) form synapses exclusively with the axon initial segments of pyramidal cells. In this study, we have used immunocytochemistry for the high-affinity plasma membrane transporter-1 (GAT-1) to analyze the distribution and density of Ch-terminals in various cytoarchitectonic and functional areas of the human neocortex. The lowest density of GAT-1-immuoreactive (-ir) Ch-terminals was detected in the primary and secondary visual (areas 17 and 18) and in the somatosensory areas (areas 3b and 1). In contrast, an intermediate density was observed in the motor area 4 and the associative frontolateral areas 45 and 46, whereas the associative frontolateral areas 9 and 10, frontal orbitary areas 11, 12, 13, 14, and 47, associative temporal areas 20, 21, 22, and 38, and cingulate areas 24 and 32 displayed the highest density of GAT-1-ir Ch-terminals. Despite these differences, the laminar distribution of GAT-1-ir Ch-terminals was similar in most cortical areas. Hence, the highest density of this transporter was observed in layer II, followed by layers III, V, VI, and IV. In most cortical areas, the density of GAT-1-ir Ch-terminals was positively correlated with the neuronal density, although a negative correlation was detected in layer III across all cortical areas. These results indicate that there are substantial differences in the distribution and density of GAT-1-ir Ch-terminals between areas and layers of the human neocortex. These differences might be related to the different functional attributes of the cortical regions examined.

Lighting the chandelier: new vistas for axo-axonic cells

Chandelier or axo-axonic cells are the most selective of all cortical GABAergic interneurons, because they exclusively contact axon initial segments of cortical glutamatergic neurons. Owing to their privileged location on initial segments, axo-axonic cells have often been assumed to have the ultimate control of pyramidal cell output. Recently, key molecules expressed at the initial-segment synapses have been identified, and novel in vitro and in vivo electrophysiological studies have revealed unexpectedly versatile functional effects exerted by axo-axonic cells on their postsynaptic targets. In addition, there is also emerging recognition of the mechanistic involvement of these unique cells in several neurological diseases, including epilepsy and schizophrenia.

A comparison of possible markers for chandelier cartridges in rat medial prefrontal cortex and hippocampus

Chandelier neurons and their characteristic arrays of axonal terminals, known as cartridges, have been implicated in a variety of psychiatric and neurological disorders including schizophrenia and epilepsy. As a result, these neurons have been extensively examined in the brains of several species using a range of markers. However, these markers have not been systematically compared in a single species for their robustness in labelling chandelier cell cartridges. We have therefore examined several markers, reported to label chandelier arrays in primates, for their capacity to mark these structures in rat medial prefrontal cortex and hippocampus. These studies revealed that cartridge-like structures were labelled by parvalbumin and GAT-1 immunohistochemistry in both medial prefrontal cortex and hippocampus of the rat brain. Additionally, GAD65 immunohistochemistry labelled array-like structures preferentially in the dentate gyrus. In contrast, PSA-NCAM, calbindin and GAD67 immunohistochemistry did not reveal any array-like structures in either region of rat brain. These observations indicate that the various immunological markers previously used to visualise chandelier cell cartridges in primates are not equally efficient in labelling these structures in the rat brain, and that GAT-1 immunohistochemistry is the most robust means of visualising chandelier cell cartridges in the regions examined. These are important considerations for quantitative studies in animal models of neurological disorders where chandelier neurons are implicated.

RC3/neurogranin is expressed in pyramidal neurons of motor and somatosensory cortex in normal and denervated monkeys

RC3/neurogranin is a neuron-specific calpacitin located in the cytoplasm and, especially, in dendrites and dendritic spines of cortical neurons, involved in many aspects of excitatory transmission and long-term potentiation. We investigated RC3 expression in pyramidal cortical neurons and interneurons of the motor and somatosensory cortex of normal Macaca fascicularis by means of double immunofluorescence and with techniques that combine immunohistochemistry and radioactive in situ hybridization. We show that RC3 is expressed in virtually all pyramidal neurons and spiny stellate neurons of neocortical areas 4, 3b, 1, 2, 5, 7, and SII, but not in the majority of cortical interneurons. RC3 protein and mRNA are tightly colocalized with the alpha subunit of CaM kinase II and the 200-kD, nonphosphorylated neurofilament, whereas they are absent from cells expressing the 27-kD, vitamin D-dependent calbindin and parvalbumin. In order to investigate possible activity-dependent regulation of the expression of RC3, we compared these results with those obtained from monkeys subjected to chronic peripheral cutaneous denervation of the first finger. We found that the pattern of distribution of RC3 in motor and somatosensory cortices after nerve cut did not differ from normal.

Chandelier cells and epilepsy

The main goal of this article is to review certain aspects of the circuitry of the human cerebral cortex that may be particularly relevant for the development, maintenance or spread of seizures. There are a number of different structural abnormalities that are commonly found in the cortex of epileptic patients, but these abnormalities do not appear to be intrinsically epileptogenic, since some patients displaying them are epileptic (after variable delays) whereas others are not. Therefore, cortical circuits in an affected brain may undergo a series of changes that finally cause epilepsy. In this article, it is proposed that the chandelier cell, which is considered to be the most powerful cortical GABAergic inhibitory interneuron, is probably a key component of cortical circuits in the establishment of human intractable temporal lobe epilepsy. These cells (among other types) have been found to be lost or reduced at epileptic foci in both experimental animals and epileptic patients. A hypothesis is presented by which the normal variability in the number of interneurons might explain the predisposition of some individuals to develop epilepsy more than others as a result of a lesion or other precipitating factors that lead to loss of neurons. The sources of GABAergic input on dendrites and somata of cortical pyramidal cells originate from many and diverse types of interneurons but, at the level of the axon initial segment of these cells, all synapses come from a few chandelier cells (five or less). Loss of one class of interneurons ending on soma and dendrites might have relatively little impact on the inhibitory control of the pyramidal cell. However, if chandelier cells were affected, it would have serious consequences for the inhibitory control of the pyramidal cells. Evidence suggests that the loss of chandelier cells may be non-specific and that when this occurs epilepsy may develop. Therefore, these cells might represent a key component in the aetiology of human temporal lobe epilepsy.

Auto-associative memory produced by disinhibition in a sparsely connected network

Algorithms for auto-associative memory in sparsely connected networks are commonly brittle with respect to relationships between such parameters as training set size, connectivity, synaptic strength, threshold, or inhibitory feedback. This paper describes an algorithm for auto-associative memory based on depotentiation of inhibitory synapses (disinhibition) rather than potentiation of excitatory synapses. All parameter values are robust, largely independent of one another, and independent of network architecture over a large range of random and structured architectures. The algorithm does not require thresholds which depend on the number of active neurons or the number of synapses per neuron. Architectures to which the algorithm is applicable include randomized variants of projective networks in which all links between neurons in the same layer are of path length two (i.e., through some other layer). The resulting fanout produces prodigiously large networks with correspondingly large information storage capacities relative to the number of synapses per neuron.

A fixation procedure for ultrastructural investigation of synaptic connections in resected human cortex

Electron microscopic investigations of the fine circuitry of human central nervous system require a well-preserved tissue ultrastructure. Because the deterioration of subcellular structures occurs rapidly in postmortem human brain, the use of a fixation by immersion of surgically resected human nervous tissue would be advantageous to investigate directly its synaptic circuitry. To obtain an optimal preservation of subcellular elements in immersion-fixed brain tissue, different conditions of fixation were first tested on 400 microns-thick sections of rat neocortex. Parameters tested were temperature of the fixative solution, concentrations of glutaraldehyde and of cacodylate buffer with or without microwave irradiation, and finally, the presence of dimethyl sulfoxide. The best ultrastructural preservation was obtained by immersing the tissue in 0.1 M cacodylate buffer, 3.0 mM CaCl2, 2% paraformaldehyde, 2.5% glutaraldehyde, and 2.5% dimethyl sulfoxide at 37 degrees C for 5 min and then at 4 degrees C for 4 h. This procedure of fixation was then applied to human neocortical tissue resected to alleviate temporal lobe epilepsy. This method led to good tissue preservation in addition to retaining the antigenicity to the inhibitory amino acid neurotransmitter, gamma-aminobutyric acid (GABA). Therefore, the tissue preservation obtained would permit these chemically defined connections to be investigated quantitatively at the electron microscopic level in resected human cortex.

Distribution of parvalbumin-immunoreactive cells and fibers in the human amygdaloid complex

The calcium-binding protein, parvalbumin, was localized immunohistochemically in the human amygdaloid complex. Neuronal cell bodies and fibers that are immunoreactive to parvalbumin were observed in most of the amygdaloid nuclei and cortical areas. Three types of immunoreactive aspiny neurons, ranging from small spherical cells (type 1) to large multipolar cells (type 2) and fusiform cells (type 3), were observed. The densities of the types of neurons that were parvalbumin-immunoreactive varied in the different regions of the amygdala. The highest densities of parvalbumin-immunoreactive neurons were observed in the lateral nucleus, in the magnocellular and intermediate divisions of the basal nucleus, in the magnocellular division of the accessory basal nucleus and in the amygdalohippocampal area. The regions containing the lowest density of parvalbumin-immunoreactive cells were the paralaminar nucleus, the parvicellular division of the basal nucleus, the central nucleus, the medial nucleus and the anterior cortical nucleus. In general, the distribution of immunoreactive fibers and terminals paralleled that of immunoreactive cells. Parvalbumin-immunoreactive varicose fibers formed basket-like plexi and cartridges around the unstained neurons, which suggests that parvalbumin is located in GABAergic basket cells and chandelier cells, respectively. The distribution of parvalbumin-immunoreactive profiles in the human amygdaloid complex was similar to, rather than different from that previously reported in the monkey amygdala (Pitkänen and Amaral [1993] J. Comp. Neurol. 331:14-36). This study provides baseline information about the organization of GABAergic inhibitory circuitries in the human amygdaloid complex.

Ultrastructure of parvalbumin-immunoreactive neurons in the CA1 area of the rat hippocampus following a kainic acid injection

Non-pyramidal neurons which contain the calcium-binding protein parvalbumin (PV) were examined, at the ultrastructural level, in the CA1 area of the normal hippocampus and following a unilateral intracerebroventricular kainic acid (KA) injection. Many degenerating PV-immunoreactive (IR) neurons were identified in the ipsilateral strata oriens and pyramidale at 3 days post-lesion and some were seen in stratum oriens of the contralateral CA1 area. A few PV-IR neurons remained resistant to the effects of KA. A chronic, almost total loss of PV-IR terminals was detected around the soma of the ipsilateral CA1 pyramidal neurons. However, the PV-IR terminals around the axon initial segments of the CA1 pyramidial neurons remained intact at all post-lesion survival times in both ipsilateral and contralateral tissue. The examination of serial ultrathin sections established the origin of the PV-IR terminals around the axon initial segments to be the KA-resistant PV-IR neurons in stratum pyramidale. This data provides evidence for the loss of non-pyramidal neurons following a KA lesion together with evidence for a surviving inhibitory circuit that could, if functional, provide a very strong inhibitory control of pyramidal neurons.

Precision and variability in postsynaptic target selection of inhibitory cells in the hippocampal CA3 region

Non-pyramidal cells were filled intracellularly with biocytin in the CA3 region of the guinea-pig hippocampus in vitro, within or close to stratum pyramidale. On the basis of camera lucida reconstructions and electron microscopy, six different cell types with distinct laminar distribution of axon terminals could be distinguished. The axon of three axo-axonic cells, three typical basket cells, and atypical basket cells of two types arborized in the perisomatic and proximal dendritic region of CA3 pyramidal cells. Two cells with axons innervating the distal dendritic segments of pyramidal cells were also found; one terminated in stratum radiatum and the other in stratum lacunosum-moleculare. Electron microscopy demonstrated that symmetrical synapses were formed by the labelled boutons on axon initial segments, somata, and proximal or distal dendrites of mostly pyramidal neurons. Axo-axonic cells showed absolute target selectivity for axon initial segments, whereas for the other cells the distribution of contacted elements was determined by the laminar distribution of axon terminals. In two cases, where additional cells were labelled with biocytin, multiple (up to nine) light microscopically identified contacts (presumed synaptic contacts) were established by the interneurons on several pyramidal cells and on an axo-axonic cell. Our results show that a restricted set of inhibitory cells, with somata within or close to CA3 stratum pyramidale, possess variable patterns of axonal arborization. Various types of postsynaptic elements are contacted, but precision in selecting certain targets and ignoring others is maintained within a particular cell type and layer. In contrast to the diversity of axonal arbors the structure of the dendritic trees shows no consistent differences, suggesting that the cells may be activated by a similar set of afferents. It seems probable that the innervation of precise regions of postsynaptic pyramidal cells by different types of interneurons--often in conjunction with particular excitatory afferents (Han et al., Eur. J. Neurosci., 5, 395-410, 1993)--underlies functional differences in inhibitory synaptic actions.

Chandelier cell axons identified by parvalbumin-immunoreactivity in the normal human temporal cortex and in Alzheimer's disease

Parvalbumin is a calcium-binding protein which is thought to play a role in neuronal excitability. In the cerebral cortex parvalbumin is largely found in two subsets of GABAergic neurons, the chandelier and basket cells. A distinguishing characteristic of the chandelier cell is that the terminal portions of its axon form short vertical strings of boutons resembling candlesticks, which embrace the initial segment of pyramidal cell axon. In the present study, the terminals of chandelier cells in the human temporal cortex were immunostained with an antibody against parvalbumin. These terminals were found more abundantly in layers II and VI, less frequently in layers III and V, were hardly identified in layer IV, and absent in layer I. The relationship of parvalbumin-immunoreactive terminals and axon initial segments was further evidenced by re-sectioning identified rows of boutons into semithin sections. Electron microscopy of both temporal cortex and the somatosensory region of a biopsy sample revealed that these parvalbumin-positive boutons indeed form symmetric synaptic contacts on the axon initial segments of pyramidal cells. As part of an enquiry into the possibility that these specialized interneurons may be involved in degenerative neurological diseases, the temporal lobes from seven patients with Alzheimer's disease were immunostained for parvalbumin. As in the control brains, the specific terminal portions of chandelier cells were recognized and identified in the temporal cortex by parvalbumin-immunocytochemistry. No major difference from normal brains was found, excepting for a lower density of candlesticks (30-35%) in layer II-III. Since we showed in a previous study [Ferrer et al. (1991) J. neurol. Sci. 106, 135-141] that the number of parvalbumin-immunoreactive somata in the same Alzheimer's disease cases was not decreased, the observed reduction of terminals in layer II suggest that only the terminals of chandelier cells, but not the parent neurons, are decreased in Alzheimer's disease.

GABA neurons in seizure disorders: a review of immunocytochemical studies

Immunocytochemical studies have identified alterations in GABA neurons in several models of seizure disorders. However, the changes have varied among different epilepsy models, and these variations presumably reflect the diversity of mechanisms that can lead to seizure disorders. In models of cortical focal epilepsy, there is strong evidence for decreases in the number of GABAergic elements, and the changes closely parallel the time course of seizure development. By contrast, in some genetic models of epilepsy, increases in the number of immunocytochemically-detectable neurons have been observed in selected brain regions. In several models of temporal lobe epilepsy, there presently is little immunocytochemical evidence for alterations of GABA neurons within the hippocampal formation despite physiological demonstrations of decreased GABA-mediated inhibition in this region. However, it remains possible that certain types of GABA neurons could be differentially affected in some seizure disorders while other types are preserved. Thus, distinguishing between different classes of GABA neurons and determining their functional roles represent major challenges for future studies of GABA neurons in seizure disorders.

Synaptic connections of neurones identified by Golgi impregnation: characterization by immunocytochemical, enzyme histochemical, and degeneration methods

For more than a century the Golgi method has been providing structural information about the organization of neuronal networks. Recent developments allow the extension of the method to the electron microscopic analysis of the afferent and efferent synaptic connections of identified, Golgi-impregnated neurones. The introduction of degeneration, autoradiographic, enzyme histochemical, and immunocytochemical methods for the characterization of Golgi-impregnated neurones and their pre- and postsynaptic partners makes it possible to establish the origin and also the chemical composition of pre- and postsynaptic elements. Furthermore, for a direct correlation of structure and function the synaptic interconnections between physiologically characterized, intracellularly HRP-filled neurones and Golgi-impregnated cells can be studied. It is thought that most of the neuronal communication takes place at the synaptic junction. In the enterprise of unravelling the circuits underlying the synaptic interactions, the Golgi technique continues to be a powerful tool of analysis.