Parvalbumin Interneuron Dysfunction in Neurological Disorders: Focus on Epilepsy and Alzheimer's Disease

Parvalbumin expressing (PV+) GABAergic interneurons are fast spiking neurons that provide powerful but relatively short-lived inhibition to principal excitatory cells in the brain. They play a vital role in feedforward and feedback synaptic inhibition, preventing run away excitation in neural networks. Hence, their dysfunction can lead to hyperexcitability and increased susceptibility to seizures. PV+ interneurons are also key players in generating gamma oscillations, which are synchronized neural oscillations associated with various cognitive functions. PV+ interneuron are particularly vulnerable to aging and their degeneration has been associated with cognitive decline and memory impairment in dementia and Alzheimer's disease (AD). Overall, dysfunction of PV+ interneurons disrupts the normal excitatory/inhibitory balance within specific neurocircuits in the brain and thus has been linked to a wide range of neurodevelopmental and neuropsychiatric disorders. This review focuses on the role of dysfunctional PV+ inhibitory interneurons in the generation of epileptic seizures and cognitive impairment and their potential as targets in the design of future therapeutic strategies to treat these disorders. Recent research using cutting-edge optogenetic and chemogenetic technologies has demonstrated that they can be selectively manipulated to control seizures and restore the balance of neural activity in the brains of animal models. This suggests that PV+ interneurons could be important targets in developing future treatments for patients with epilepsy and comorbid disorders, such as AD, where seizures and cognitive decline are directly linked to specific PV+ interneuron deficits.

The Impact of Glutamatergic Synapse Dysfunction in the Corticothalamocortical Network on Absence Seizure Generation

Childhood absence epilepsy (CAE) is the most common pediatric epilepsy affecting 10–18% of all children with epilepsy. It is genetic in origin and the result of dysfunction within the corticothalamocortical (CTC) circuitry. Network dysfunction may arise from multifactorial mechanisms in patients from different genetic backgrounds and thus account for the variability in patient response to currently available anti-epileptic drugs; 30% of children with absence seizures are pharmaco-resistant. This review considers the impact of deficits in AMPA receptor-mediated excitation of feed-forward inhibition (FFI) in the CTC, on absence seizure generation. AMPA receptors are glutamate activated ion channels and are responsible for most of the fast excitatory synaptic transmission throughout the CNS. In the stargazer mouse model of absence epilepsy, the genetic mutation is in stargazin, a transmembrane AMPA receptor trafficking protein (TARP). This leads to a defect in AMPA receptor insertion into synapses in parvalbumin-containing (PV+) inhibitory interneurons in the somatosensory cortex and thalamus. Mutation in the Gria4 gene, which encodes for the AMPA receptor subunit GluA4, the predominant AMPA receptor subunit in cortical and thalamic PV + interneurons, also leads to absence seizures. This review explores the impact of glutamatergic synapse dysfunction in the CTC network on absence seizure generation. It also discusses the cellular and molecular mechanisms involved in the pathogenesis of childhood absence epilepsy.

Impact of Dysfunctional Feed-Forward Inhibition on Glutamate Decarboxylase Isoforms and γ-Aminobutyric Acid Transporters

Absence seizures are associated with generalised synchronous 2.5–4 Hz spike-wave discharges causing brief and sudden alteration of awareness during childhood, which is known as childhood absence epilepsy (CAE). CAE is also associated with impaired learning, psychosocial challenges, and physical danger. Absence seizures arise from disturbances within the cortico-thalamocortical (CTC) network, including dysfunctional feed-forward inhibition (FFI); however, the precise mechanisms remain unclear. In epileptic stargazers, a genetic mouse model of CAE with chronic seizures, levels of γ-aminobutyric acid (GABA), and expression of GABA receptors are altered within the CTC network, implicating altered GABAergic transmission in absence seizures. However, the expression of GABA synthesising enzymes (GAD65 and GAD67) and GABA transporters (GAT-1 and 3) have not yet been characterised within absence seizure models. We found a specific upregulation of GAD65 in the somatosensory cortex but not the thalamus of epileptic stargazer mice. No differences were detected in GAD67 and GAT-3 levels in the thalamus or somatosensory cortex. Then, we assessed if GAD65 upregulation also occurred in Gi-DREADD mice exhibiting acute absence seizures, but we found no change in the expression profiles of GAD65/67 or GAT-3. Thus, the upregulation of GAD65 in stargazers may be a compensatory mechanism in response to long-term dysfunctional FFI and chronic absence seizures.

Chemogenetic Activation of Feed-Forward Inhibitory Parvalbumin-Expressing Interneurons in the Cortico-Thalamocortical Network During Absence Seizures

Parvalbumin-expressing (PV+) interneurons are a subset of GABAergic inhibitory interneurons that mediate feed-forward inhibition (FFI) within the cortico-thalamocortical (CTC) network of the brain. The CTC network is a reciprocal loop with connections between cortex and thalamus. FFI PV+ interneurons control the firing of principal excitatory neurons within the CTC network and prevent runaway excitation. Studies have shown that generalized spike-wave discharges (SWDs), the hallmark of absence seizures on electroencephalogram (EEG), originate within the CTC network. In the stargazer mouse model of absence epilepsy, reduced FFI is believed to contribute to absence seizure genesis as there is a specific loss of excitatory α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs) at synaptic inputs to PV+ interneurons within the CTC network. However, the degree to which this deficit is directly related to seizure generation has not yet been established. Using chemogenetics and in vivo EEG recording, we recently demonstrated that functional silencing of PV+ interneurons in either the somatosensory cortex (SScortex) or the reticular thalamic nucleus (RTN) is sufficient to generate absence-SWDs. Here, we used the same approach to assess whether activating PV+ FFI interneurons within the CTC network during absence seizures would prevent or reduce seizures. To target these interneurons, mice expressing Cre recombinase in PV+ interneurons (PV-Cre) were bred with mice expressing excitatory Gq-DREADD (hM3Dq-flox) receptors. An intraperitoneal dose of pro-epileptic chemical pentylenetetrazol (PTZ) was used to induce absence seizure. The impact of activation of FFI PV+ interneurons during seizures was tested by focal injection of the “designer drug” clozapine N-oxide (CNO) into either the SScortex or the RTN thalamus. Seizures were assessed in PV^Cre/Gq-DREADD animals using EEG/video recordings. Overall, DREADD-mediated activation of PV+ interneurons provided anti-epileptic effects against PTZ-induced seizures. CNO activation of FFI either prevented PTZ-induced absence seizures or suppressed their severity. Furthermore, PTZ-induced tonic-clonic seizures were also reduced in severity by activation of FFI PV+ interneurons. In contrast, administration of CNO to non-DREADD wild-type control animals did not afford any protection against PTZ-induced seizures. These data demonstrate that FFI PV+ interneurons within CTC microcircuits could be a potential therapeutic target for anti-absence seizure treatment in some patients.

Altered Neurotransmitter Expression in the Corticothalamocortical Network of an Absence Epilepsy Model with impaired Feedforward Inhibition

The episodes of brief unconsciousness in patients with childhood absence epilepsy are a result of corticothalamocortical circuitry dysfunction. This dysfunction may arise from multifactorial mechanisms in patients from different genetic backgrounds. In previous studies using the epileptic stargazer mutant mouse, which experience frequent absence seizures, we reported a deficit in AMPAR-mediated feed-forward inhibition of parvalbumin-containing (PV⁺) interneurons. Currently, in order to determine the downstream effects of this impairment on neurotransmitter expression, we performed HPLC of tissue lysates and post-embedding electron microscopy from the cortical and thalamic regions. We report region-specific alterations in GABA expression, but not of glutamate, and most prominently at PV⁺ synaptic terminals. These results suggest that impaired feed forward inhibition may occur via reduced activation of these interneurons and concomitant decreased GABAergic signaling. Further investigations into GABAergic control of corticothalamocortical network activity could be key in our understanding of absence seizure pathogenesis.

Local protein synthesis is a ubiquitous feature of neuronal pre- and postsynaptic compartments

There is ample evidence for localization of messenger RNAs (mRNAs) and protein synthesis in neuronal dendrites; however, demonstrations of these processes in presynaptic terminals are limited. We used expansion microscopy to resolve pre- and postsynaptic compartments in rodent neurons. Most presynaptic terminals in the hippocampus and forebrain contained mRNA and ribosomes. We sorted fluorescently labeled mouse brain synaptosomes and then sequenced hundreds of mRNA species present within excitatory boutons. After brief metabolic labeling, >30% of all presynaptic terminals exhibited a signal, providing evidence for ongoing protein synthesis. We tested different classic plasticity paradigms and observed distinct patterns of rapid pre- and/or postsynaptic translation. Thus, presynaptic terminals are translationally competent, and local protein synthesis is differentially recruited to drive compartment-specific phenotypes that underlie different forms of plasticity.

The impact of silencing feed-forward parvalbumin-expressing inhibitory interneurons in the cortico-thalamocortical network on seizure generation and behaviour

Feed-forward inhibition (FFI) is an essential mechanism within the brain, to regulate neuronal firing and prevent runaway excitation. In the cortico-thalamocortical (CTC) network, fast spiking parvalbumin-expressing (PV+) inhibitory interneurons regulate the firing of pyramidal cells in the cortex and relay neurons in the thalamus. PV+ interneuron dysfunction has been implicated in several neurological disorders, including epilepsy. Previously, we demonstrated that loss of excitatory AMPA-receptors, specifically at synapses on PV+ interneurons in CTC feedforward microcircuits, occurs in the stargazer mouse model of absence epilepsy. These mice present with absence seizures characterized by spike and wave discharges (SWDs) on electroencephalogram (EEG) and concomitant behavioural arrest, similar to childhood absence epilepsy. The aim of the current study was to investigate the impact of loss of FFI within the CTC on absence seizure generation and behaviour using new Designer Receptor Exclusively Activated by Designer Drug (DREADD) technology. We crossed PV-Cre mice with Cre-dependent hM4Di DREADD strains of mice, which allowed Cre-recombinase-mediated restricted expression of inhibitory G_i-DREADDs in PV+ interneurons. We then tested the impact of global and focal (within the CTC network) silencing of PV+ interneurons. CNO mediated silencing of all PV+ interneurons by intraperitoneal injection caused the impairment of motor control, decreased locomotion and increased anxiety in a dose-dependent manner. Such silencing generated pathological oscillations similar to absence-like seizures. Focal silencing of PV+ interneurons within cortical or thalamic feedforward microcircuits, induced SWD-like oscillations and associated behavioural arrest. Epileptiform activity on EEG appeared significantly sooner after focal injection compared to peripheral injection of CNO. However, the mean duration of each oscillatory burst and spike frequency was similar, irrespective of mode of CNO delivery. No significant changes were observed in vehicle-treated or non-DREADD wild-type control animals. These data suggest that dysfunctional feed-forward inhibition in CTC microcircuits may be an important target for future therapy strategies for some patients with absence seizures. Additionally, silencing of PV+ interneurons in other brain regions may contribute to anxiety related neurological and psychiatric disorders.

Synaptic Changes in AMPA Receptor Subunit Expression in Cortical Parvalbumin Interneurons in the Stargazer Model of Absence Epilepsy

Feedforward inhibition is essential to prevent run away excitation within the brain. Recent evidence suggests that a loss of feed-forward inhibition in the corticothalamocortical circuitry may underlie some absence seizures. However, it is unclear if this aberration is specifically linked to loss of synaptic excitation onto local fast-spiking parvalbumin-containing (PV⁺) inhibitory interneurons, which are responsible for mediating feedforward inhibition within cortical networks. We recently reported a global tissue loss of AMPA receptors (AMPARs), and a specific mistrafficking of these AMPARs in PV⁺ interneurons in the stargazer somatosensory cortex. The current study was aimed at investigating if cellular changes in AMPAR expression were translated into deficits in receptors at specific synapses in the feedforward inhibitory microcircuit. Using western blot immunolabeling on biochemically isolated synaptic fractions, we demonstrate a loss of AMPAR GluA1–4 subunits in the somatosensory cortex of stargazers compared to non-epileptic control mice. Furthermore, using double post-embedding immunogold-cytochemistry, we show a loss of GluA1–4-AMPARs at excitatory synapses onto cortical PV⁺ interneurons. Altogether, these data indicate a loss of synaptic AMPAR-mediated excitation of cortical PV⁺ inhibitory neurons. As the cortex is considered the site of initiation of spike wave discharges (SWDs) within the corticothalamocortical circuitry, loss of AMPARs at cortical PV⁺ interneurons likely impairs feed-forward inhibitory output, and contributes to the generation of SWDs and absence seizures in stargazers.

NMDA Receptor Expression in the Thalamus of the Stargazer Model of Absence Epilepsy

In the stargazer mouse model of absence epilepsy, altered corticothalamic excitation of reticular thalamic nucleus (RTN) neurons has been suggested to contribute to abnormal synchronicity in the corticothalamic-thalamocortical circuit, leading to spike-wave discharges, the hallmark of absence seizures. AMPA receptor expression and function are decreased in stargazer RTN, due to a mutation of AMPAR auxiliary subunit stargazin. It is unresolved and debated, however, if decreased excitation of RTN is compatible with epileptogenesis. We tested the hypothesis that relative NMDAR expression may be increased in RTN and/or thalamic synapses in stargazers using Western blot on dissected thalamic nuclei and biochemically isolated synapses, as well as immunogold cytochemistry in RTN. Expression of main NMDAR subunits was variable in stargazer RTN and relay thalamus; however, mean expression values were not statistically significantly different compared to controls. Furthermore, no systematic changes in synaptic NMDAR levels could be detected in stargazer thalamus. In contrast, AMPAR subunits were markedly decreased in both nucleus-specific and synaptic preparations. Thus, defective AMPAR trafficking in stargazer thalamus does not appear to lead to a ubiquitous compensatory increase in total and synaptic NMDAR expression, suggesting that elevated NMDAR function is not mediated by changes in protein expression in stargazer mice.

Alterations in AMPA receptor subunit expression in cortical inhibitory interneurons in the epileptic stargazer mutant mouse

Absence seizures arise from disturbances within the corticothalamocortical network, however the precise cellular and molecular mechanisms underlying seizure generation arising from different genetic backgrounds are not fully understood. While recent experimental evidence suggests that changes in inhibitory microcircuits in the cortex may contribute to generation of the hallmark spike-wave discharges, it is still unclear if altered cortical inhibition is a result of interneuron dysfunction due to compromised glutamatergic excitation and/or changes in cortical interneuron number. The stargazer mouse model of absence epilepsy presents with a genetic deficit in stargazin, which is predominantly expressed in cortical parvalbumin-positive (PV⁺) interneurons, and involved in the trafficking of glutamatergic AMPA receptors. Hence, in this study we examine changes in (1) the subunit-specific expression of AMPA receptors which could potentially result in a loss of excitation onto cortical PV⁺ interneurons, and (2) PV⁺ neuron density that could additionally impair cortical inhibition. Using Western blot analysis we found subunit-specific alterations in AMPA receptor expression in the stargazer somatosensory cortex. Further analysis using confocal fluorescence microscopy revealed that although there are no changes in cortical PV⁺ interneuron number, there is a predominant loss of GluA1 and 4 containing AMPA receptors in PV⁺ neurons in stargazers compared to non-epileptic controls. Taken together, these data suggest that the loss of AMPA receptors in PV⁺ neurons could impair their feed-forward inhibitory output, ultimately altering cortical network oscillations, and contribute to seizure generation in stargazers. As such the feed-forward inhibitory interneurons could be potential targets for future therapeutic intervention for some absence epilepsy patients.

Region-specific changes in presynaptic agmatine and glutamate levels in the aged rat brain

During the normal aging process, the brain undergoes a range of biochemical and structural alterations, which may contribute to deterioration of sensory and cognitive functions. Age-related deficits are associated with altered efficacy of synaptic neurotransmission. Emerging evidence indicates that levels of agmatine, a putative neurotransmitter in the mammalian brain, are altered in a region-specific manner during the aging process. The gross tissue content of agmatine in the prefrontal cortex (PFC) of aged rat brains is decreased whereas levels in the temporal cortex (TE) are increased. However, it is not known whether these changes in gross tissue levels are also mirrored by changes in agmatine levels at synapses and thus could potentially contribute to altered synaptic function with age. In the present study, agmatine levels in presynaptic terminals in the PFC and TE regions (300 terminals/region) of young (3month; n=3) and aged (24month; n=3) brains of male Sprague-Dawley rats were compared using quantitative post-embedding immunogold electron-microscopy. Presynaptic levels of agmatine were significantly increased in the TE region (60%; p<0.001) of aged rats compared to young rats, however no significant differences were detected in synaptic levels in the PFC region. Double immunogold labeling indicated that agmatine and glutamate were co-localized in the same synaptic terminals, and quantitative analyses revealed significantly reduced glutamate levels in agmatine-immunopositive synaptic terminals in both regions in aged rats compared to young animals. This study, for the first time, demonstrates differential effects of aging on agmatine and glutamate in the presynaptic terminals of PFC and TE. Future research is required to understand the functional significance of these changes and the underlying mechanisms.

Altered thalamic GABAA-receptor subunit expression in the stargazer mouse model of absence epilepsy

PURPOSE

Absence seizures, also known as petit mal seizures, arise from disruptions within the cortico-thalamocortical network. Interconnected circuits within the thalamus consisting of inhibitory neurons of the reticular thalamic nucleus (RTN) and excitatory relay neurons of the ventral posterior (VP) complex, generate normal intrathalamic oscillatory activity. The degree of synchrony in this network determines whether normal (spindle) or pathologic (spike wave) oscillations occur; however, the cellular and molecular mechanisms underlying absence seizures are complex and multifactorial and currently are not fully understood. Recent experimental evidence from rodent models suggests that regional alterations in γ-aminobutyric acid (GABA)ergic inhibition may underlie hypersynchronous oscillations featured in absence seizures. The aim of the current study was to investigate whether region-specific differences in GABAA receptor (GABAAR) subunit expression occur in the VP and RTN thalamic regions in the stargazer mouse model of absence epilepsy where the primary deficit is in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) expression.

METHODS

Immunofluorescence confocal microscopy and semiquantitative Western blot analysis were used to investigate region-specific changes in GABAAR subunits in the thalamus of the stargazer mouse model of absence epilepsy to determine whether changes in GABAergic inhibition could contribute to the mechanisms underlying seizures in this model of absence epilepsy.

KEY FINDINGS

Immunofluorescence confocal microscopy revealed that GABAAR α1 and β2 subunits are predominantly expressed in the VP, whereas α3 and β3 subunits are localized primarily in the RTN. Semiquantitative Western blot analysis of VP and RTN samples from epileptic stargazers and their nonepileptic littermates showed that GABAAR α1 and β2 subunit expression levels in the VP were significantly increased (α1: 33%, β2: 96%) in epileptic stargazers, whereas α3 and β3 subunits in the RTN were unchanged in the epileptic mice compared to nonepileptic control littermates.

SIGNIFICANCE

These findings suggest that region-specific differences in GABAAR subunits in the thalamus of epileptic mice, specifically up-regulation of GABAARs in the thalamic relay neurons of the VP, may contribute to generation of hypersynchronous thalamocortical activity in absence seizures. Understanding region-specific differences in GABAAR subunit expression could help elucidate some of the cellular and molecular mechanisms underlying absence seizures and thereby identify targets by which drugs can modulate the frequency and severity of epileptic seizures. Ultimately, this information could be crucial for the development of more specific and effective therapeutic drugs for treatment of this form of epilepsy.

Changes in the GRIP 1&2 scaffolding proteins in the cerebellum of the ataxic stargazer mouse

Glutamate receptor-interacting proteins (GRIP1&2) and protein-interacting with C kinase-1 (PICK1) are synaptic scaffold proteins associated with the stabilization and recycling of synaptic GluA2-, 3- and 4c-containing α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs). PICK1-mediated phosphorylation of GluA serine880 uncouples GRIP1&2 leading to AMPAR endocytosis, important in mediating forms of synaptic plasticity underlying learning and memory. Ataxic and epileptic stargazer mice possess a mutation in the CACNG2 gene encoding the transmembrane AMPAR-regulatory protein (TARP)-γ2 (stargazin). TARPs are AMPAR-auxiliary subunits required for efficient AMPAR trafficking to synapses. Stargazin is abundantly expressed in the cerebellum and its loss results in severe deficits in AMPAR trafficking to cerebellar synapses, particularly at granule cell (GC) synapses, leading to the ataxic phenotype of stargazers. However, how the stargazin mutation impacts on the expression of other AMPAR-interacting scaffold proteins is unknown. This study shows a significant increase in GRIP1&2, but not PICK1, levels in whole tissue and synapse-enriched extracts from stargazer cerebella. Post-embedding immunogold-cytochemistry electron microscopy showed GRIP1&2 levels were unchanged at mossy fiber-GC synapses in stargazers, which are silent due to virtual total absence of synaptic and extrasynaptic GluA2/3-AMPARs. These results indicate that loss of synaptic AMPARs at this excitatory synapse does not affect GRIP1&2 expression within the postsynaptic region of mossy fiber-GC synapses. Interestingly, increased GRIP and reduced GluA2-AMPARexpression also occur in cerebella of autistic patients. Further research establishing the role of elevated cerebellar GRIP1&2 in stargazers may help identify common cellular mechanisms in the comorbid disorders ataxia, epilepsy and autism leading to more effective treatment strategies.

Selective loss of AMPA receptors at corticothalamic synapses in the epileptic stargazer mouse

Absence seizures are common in the stargazer mutant mouse. The mutation underlying the epileptic phenotype in stargazers is a defect in the gene encoding the normal expression of the protein stargazin. Stargazin is involved in the membrane trafficking and synaptic targeting of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs) at excitatory glutamatergic synapses. Thus, the genetic defect in the stargazer results in a loss of AMPARs and consequently, excitation at glutamatergic synapses. Absence seizures are known to arise in thalamocortical networks. In the present study we show for the first time, using Western blot analysis and quantitative immunogold cytochemistry, that in the epileptic stargazer mouse, there is a global loss of AMPAR protein in nucleus reticularis (RTN) and a selective loss of AMPARs at corticothalamic synapses in inhibitory neurons of the RTN thalamus. In contrast, there is no significant loss of AMPARs at corticothalamic synapses in excitatory relay neurons in the thalamic ventral posterior (VP) region. The findings of this study thus provide cellular and molecular evidence for a selective regional loss of synaptic AMPAR within the RTN that could account for the loss of function at these inhibitory neuron synapses, which has previously been reported from electrophysiological studies. The specific loss of AMPARs at RTN but not relay synapses in the thalamus of the stargazer, could contribute to the absence epilepsy phenotype by altering thalamocortical network oscillations. This is supported by recent evidence that loss of glutamate receptor subunit 4 (GluA4) (the predominant AMPAR-subtype in the thalamus), also leads to a specific reduction in strength in the cortico-RTN pathway and enhanced thalamocortical oscillations, in the Gria4(-/-) model of absence epilepsy. Thus further study of thalamic changes in these models could be important for future development of drugs targeted to absence epilepsy.

Spatial learning-induced accumulation of agmatine and glutamate at hippocampal CA1 synaptic terminals

Agmatine, the decarboxylated metabolite of l-arginine, is considered to be a novel putative neurotransmitter. Recent studies have demonstrated that endogenous agmatine may directly participate in the processes of spatial learning and memory. Agmatine-immunoreactivity has been observed within synaptic terminals of asymmetric excitatory synapses in the hippocampal CA1 stratum radiatum (SR), suggesting that agmatine may be colocalized with glutamate. In the present study we demonstrate, using immunofluorescence confocal microscopy, that agmatine is colocalized with glutamate within CA1-CA3 hippocampal pyramidal cell bodies, in young Sprague-Dawley rats. Subcellular investigation, using postembedding electron microscopy-immunogold cytochemistry, has also revealed that agmatine is colocalized with glutamate in most synaptic terminals in the SR region of CA1. Ninety-seven percent of all agmatinergic profiles were found to contain glutamate, and 92% of all glutamatergic profiles contained agmatine (n=6; 300 terminals). Alterations in colocalized agmatine and glutamate levels in the SR synaptic terminals, following 4 days Morris water maze training, were also investigated. Compared with swim only control rats, water maze-trained rats had statistically significant increases in both agmatine (78%; P<0.01) and glutamate (41%; P<0.05) levels within SR terminals synapsing onto CA1 dendrites. These findings provide the first evidence that agmatine and glutamate are colocalized in synaptic terminals in the hippocampal CA1 region, and may co-participate in spatial learning and memory processing.

Spatial learning-induced increase in agmatine levels at hippocampal CA1 synapses

Agmatine, a metabolite of L-arginine, is considered as a novel putative neurotransmitter. It has been detected in axon terminals that synapse with pyramidal cells in the hippocampus, a brain region that is critically involved in spatial learning and memory. However, the role of agmatine in learning and memory is poorly understood. Recently, we demonstrated water maze training-induced increases in tissue levels of agmatine in the CA1 subregion of the hippocampus. This finding has raised an issue whether an endogenous agmatine could directly participate in learning and memory processes as a neurotransmitter. In the present study, quantitative immunogold-labeling and electron-microscopical techniques were used to analyze the levels of agmatine in CA1 stratum radiatum (SR) terminals (n = 600) of male Sprague-Dawley rats that had been trained to find a hidden escape platform in the water maze (WM) task or forced to swim (SW) in the pool with no platform presented. Agmatine levels were significantly increased by ∼85% in the synaptic terminals of SR of trained WM group compared with the SW control group (all P < 0.001). These results, for the first time, demonstrate spatial learning-induced elevation in agmatine levels at synapses in the hippocampus and provide evidence of its participation in learning and memory processing as a novel neurotransmitter.

Subcellular distribution of L-type calcium channel subtypes in rat hippocampal neurons

L-type calcium channels play an essential role in synaptic activity-dependent gene expression and are implicated in long-term alterations in synaptic efficacy underlying learning and memory in the hippocampus. The two principal pore-forming subunits of L-type Ca2+ channels expressed in neurons are the Ca(v)1.2 (alpha(1C)) or Ca(v)1.3 (alpha(1D)) subtypes. Experimental evidence suggests that calcium entry through Ca(v)1.2 and Ca(v)1.3 Ca2+ channels occurs in close proximity to key signalling molecules responsible for triggering signalling pathways leading to transcriptional responses. Determining the subcellular distribution of Ca(v)1.2 and Ca(v)1.3 L-type channels in neurons is clearly important for unravelling the molecular mechanisms underlying long-term alterations in neuronal function. In this study, we used immunogold-labelling techniques and electron-microscopy (EM) to analyse the subcellular distribution and density of both Ca(v)1.2 and Ca(v)1.3 Ca2+ channels in rat hippocampal CA1 pyramidal cells in vivo. We confirm that both Ca(v)1.2 and Ca(v)1.3 channel subtypes are predominantly but not exclusively located in postsynaptic dendritic processes and somata. Both Ca(v)1.2 and Ca(v)1.3 are distributed throughout the dendritic tree. However, the smallest (distal) dendritic processes and spines have proportionally more calcium channels inserted into their plasma membrane than located within cytoplasmic compartments indicating the potential targeting of calcium channels to microdomains within neurons. Ca(v)1.2 and Ca(v)1.3 Ca2+ channels are located at the postsynaptic density and also at extra-synaptic sites. The location of L-type Ca(v)1.2 and Ca(v)1.3 channels in distal dendrites and spines would thus place them at appropriate sites where they could initiate synapse to nucleus signalling.

Selective reduction in synaptic proteins involved in vesicle docking and signalling at synapses in the ataxic mutant mouse stargazer

The spontaneous recessive mutant mouse stargazer has a specific and pronounced deficit in brain-derived neurotrophic factor (BDNF) mRNA expression in the cerebellum. Cerebellar granule cells, in particular, show a selective and near-total loss of BDNF. The mutation involves a defect in the calcium channel subunit Cacng2. This severely reduces expression of stargazin. A stargazin-induced failure in BDNF expression is thought to underlie the cerebellar ataxia with which the mutant presents. BDNF is known to regulate plasticity at cerebellar synapses. However, relatively little is known about the mechanism involved. We previously demonstrated that the stargazer mutation affects the phenotype of cerebellar glutamatergic neurons. Stargazer neurons have less glutamate and proportionally fewer docked vesicles at presynaptic sites than controls. In the current study, we investigate the mechanism underlying BDNF-induced synaptic changes by analyzing alterations in synaptic signalling proteins in the stargazer cerebellum. Expression levels of synaptic proteins were evaluated by measuring relative density of immunogold label over granule cell terminals in ultrathin sections from ataxic stargazer mutants compared with matched nonataxic littermates. We show that there is a selective and marked depletion in the levels of vesicle-associated proteins (synaptobrevin, synaptophysin, synaptotagmin, and Rab3a) but not of plasma membrane-associated protein (SNAP-25) in the terminals of the BDNF-deficient granule cells. Changes are restricted to the cerebellum; levels in the hippocampus are unaltered. These data suggest that the BDNF deficits in the cerebellum of stargazer affect synaptic vesicle docking by selectively altering synaptic-protein distribution and abundance.

Identification of the neurotransmitters involved in modulation of transmitter release from the central terminals of the locust wing hinge stretch receptor

The flight motor system of the locust represents a model preparation for the investigation of neuromodulation. At the wing hinges are stretch receptors important in generating and controlling the flight motor pattern. The forewing stretch receptor (fSR) makes direct cholinergic synapses with depressor motor neurons (MN) controlling that wing, including the first basalar MN (BA1). The fSR/BA1 synapse is modulated by muscarinic cholinergic receptors located on gamma-aminobutyric acid (GABA)-ergic interneurons (Judge and Leitch [1999a] J. Comp. Neurol. 407:103-114; Judge and Leitch [1999b] J. Neurobiol. 40:420-431). However, electrophysiology has shown that fSR/BA is also modulated by biogenic amines (Leitch et al. [2003] J. Comp. Neurol. 462:55-70). We have used electron microscopic immunocytochemistry (ICC) to identify the neurotransmitters in neurons presynaptic to the fSR and to determine the relative proportion of these different classes of modulatory inputs. Approximately 55% of all inputs to the fSR are glutamate-IR, indicating that glutamatergic neurons may also play an important role in presynaptically modulating the fSR terminals. Anti-GABA ICC confirmed that over 40% of inputs to the fSR are GABA-IR (Judge and Leitch [1999a] J. Comp. Neurol. 407:103-114). Labelling sections with an antioctopamine antibody revealed neurons containing distinctive large, electron-dense granules, which could reliably be used to identify them. Aminergic neurons that modulate the synapse may have very few morphologically recognizable synaptic outputs. Although putative octopaminergic processes were found in close contact to horseradish peroxidase-filled fSR profiles, no morphologically recognizable synaptic inputs to the fSR were evident. Collectively, these data suggest that most inputs to the fSR are from either glutamatergic or GABAergic neurons.

Phenotype of cerebellar glutamatergic neurons is altered in stargazer mutant mice lacking brain-derived neurotrophic factor mRNA expression

Brain-derived neurotrophic factor (BDNF) influences neuronal survival, differentiation, and maturation. More recently, its role in synapse formation and plasticity has also emerged. In the cerebellum of the spontaneous recessive mutant mouse stargazer (stg) there is a specific and pronounced deficit in BDNF mRNA expression. BDNF protein levels in the cerebellum as a whole are reduced by 70%, while in the granule cells (GCs) there is a selective and near total reduction in BDNF mRNA expression. Recently, we published data demonstrating that inhibitory neurons in the cerebella of stgs have significantly reduced levels (approximately 50%) of gamma-aminobutyric acid (GABA) and fewer, smaller inhibitory synapses compared to wildtype (WT) controls. Our current investigations indicate that the stargazer mutation has an even more pronounced effect on the phenotype of glutamatergic neurons in the cerebellum. There is a profound decrease in the levels of glutamate-immunoreactivity (up to 77%) in stg compared to WT controls. The distribution profile of presynaptic vesicles is also markedly different: stgs have proportionally fewer docked vesicles and fewer vesicles located adjacent to the active zone ready to dock than WTs. Furthermore, the thickness of the postsynaptic density (PSD) at mossy fiber-granule cell (MF-GC) and parallel fiber-Purkinje cell (PF-PC) synapses is severely reduced (up to 33% less than WT controls). The number and length of excitatory synapses, however, appear to be relatively unchanged. It is possible that at least some of theses changes in phenotype are directly attributable to the lack of BDNF in the cerebellum of the stg mutant.

Octopaminergic modulation of synaptic transmission between an identified sensory afferent and flight motoneuron in the locust

The role of the biogenic amine octopamine in modulating cholinergic synaptic transmission between the locust forewing stretch receptor neuron (fSR) and the first basalar motoneuron (BA1) was investigated. The amines 5-hydroxytryptamine (5-HT, serotonin) and dopamine were also studied. Bath application of octopamine, 5-HT, and dopamine at concentrations of 10(-4) M reversibly decreased the amplitude of monosynaptic excitatory postsynaptic potentials (EPSPs) evoked in BA1 by electrically stimulating the fSR axon. These effects occurred without any detectable change in either input resistance or membrane potential of BA1. The amines also reversibly decreased the amplitude of responses to acetylcholine (ACh) pressure-applied to the soma of BA1. The muscarinic antagonist scopolamine (10(-6) M) had no significant effect on the octopamine-induced decrease in ACh responses. These observations suggest that these amines potentially could physiologically depress cholinergic transmission between fSR and BA1, at least in part, by altering nicotinic rather than muscarinic cholinergic receptor function. Although the octopaminergic agonists naphazoline and tolazoline both mimicked the actions of octopamine, the receptor responsible for octopamine-mediated modulation could not be characterized since amine receptor antagonists tested on the preparation had complex actions. Confocal immunocytochemistry revealed intense octopamine immunoreactivity in the anterior lateral association center, thus confirming the presence of octopamine in neuropil regions containing fSR/BA1 synapses and therefore supporting a role for this amine in the modulation of synaptic transmission between the fSR and BA1. 5-HT-immunoreactivity, conversely, was concentrated within the ventral association centers; very little staining was observed in the dorsal neuropil regions in which fSR/BA1 synapses are located.

Cerebellar Golgi, Purkinje, and basket cells have reduced gamma-aminobutyric acid immunoreactivity in stargazer mutant mice

The stargazer mutant mouse has characteristic ataxia and head-tossing traits coupled with a severe impairment in the acquisition of classical eye-blink conditioning (Qiao et al. [1996] J. Neurosci. 16:640-648; Qiao et al. [ 1998] J. Neurosci. 18:6990-6999). These phenotypes are thought to be cerebellar mediated and have been attributed to the specific reduction in brain-derived neurotrophic factor (BDNF). The granule cells in the cerebellum of the stargazer mouse exhibit a near-total and exclusive ablation of BDNF mRNA expression and a consequent defect in TrkB receptor signalling. To investigate whether the stargazer mutation and lack of availability of BDNF in the granule cells compromise the phenotype of the cerebellar inhibitory neurons, specifically their immunoreactivity for gamma-aminobutyric acid (GABA); the levels of GABA neurotransmitter expressed in Golgi, Purkinje, and basket cells; and the density of their synaptic contacts were compared in stargazer and wild-type controls using electron microscopy and quantitative immunogold labelling. The data presented in this study clearly show that, in the spontaneous ataxic mutant mouse stargazer, the cerebellar inhibitory neurons have significantly reduced levels of GABA immunoreactivity indicative of a significant decrease in their GABA content compared with wild-type controls. Furthermore, the density of inhibitory synapses between Golgi interneurons and granule cells and also between basket and Purkinje cells in stargazer mutants is reduced to approximately half that in wild-type controls. Whether this reduction in GABA content and inhibitory synapse density is directly attributable to the lack of BDNF in the cerebellum of the stargazer mutant is yet to be proved.

Snowdrop lectin (GNA) has no acute toxic effects on a beneficial insect predator, the 2-spot ladybird (Adalia bipunctata L.)

Two-spot ladybird (Adalia bipunctata L.) larvae were fed on aphids (Myzus persicae (Sulz.)) which had been loaded with snowdrop lectin (Galanthus nivalis agglutinin; GNA) by feeding on artificial diet containing the protein. Treatment with GNA significantly decreased the growth of aphids. No acute toxicity of GNA-containing aphids towards the ladybird larvae was observed, although there were small effects on development. When fed a fixed number of aphids, larvae exposed to GNA spent longer in the 4th instar, taking 6 extra days to reach pupation; however, retardation of development was not observed in ladybird larvae fed equal weights of aphids. Ladybird larvae fed GNA-containing aphids were found to be 8-15% smaller than controls, but ate a significantly greater number of aphids (approx. 40% to pupation). GNA was shown to be present on the microvilli of the midgut brush border membrane and within gut epithelial cells in ladybird larvae fed on GNA-dosed aphids, although disruption of the brush border was not observed. It is hypothesised that GNA does not have significant direct toxic or adverse effects on developing ladybird larvae, but that the effects observed may be due to the fact that the aphids fed on GNA are compromised and are thus a suboptimal food.

Modulation of transmitter release from the locust forewing stretch receptor neuron by GABAergic interneurons activated via muscarinic receptors

The role of muscarinic receptors in the down-regulation of acetylcholine (ACh) release from the locust forewing stretch receptor neuron (fSR) terminals has been investigated. Electrical stimulation of the fSR evokes monosynaptic excitatory postsynaptic potentials (EPSPs) in the first basalar motoneuron (BA1), produced mainly by the activation of postsynaptic nicotinic cholinergic receptors. The general muscarinic antagonists scopolamine (10(-6) M) and atropine (10(-8) to 10(-6) M) caused a reversible increase in the amplitude of electrically evoked EPSPs. However, scopolamine (10(-6) M) caused a slight depression in the amplitude of responses to ACh pressure-applied to the soma of BA1. These observations indicate that the EPSP amplitude enhancement is due to the blockade of muscarinic receptors on neurons presynaptic to BA1. The muscarinic receptors may be located on the fSR itself and act as autoreceptors, and/or they may be located on GABAergic interneurons which inhibit ACh release from the fSR. Electron microscopical immunocytochemistry has revealed that GABA-immunoreactive neurons make presynaptic inputs to the fSR. The GABA antagonist picrotoxin (10(-6) M) caused a reversible increase in the EPSP amplitude, which does not appear to be due to an increase in sensitivity of BA1 to ACh, as picrotoxin (10(-6) M) slightly decreased ACh responses recorded from BA1. Application of scopolamine (10(-6) M) to a preparation preincubated with picrotoxin did not cause the EPSP amplitude enhancement normally seen in control experiments; in fact, it caused a slight depression. This indicates that at least some of the presynaptic muscarinic receptors are located on GABAergic interneurons that modulate transmission at the fSR/BA1 synapse.

GABA immunoreactivity in processes presynaptic to the locust wing stretch receptor neuron

Studies on the locust Locusta migratoria (Leitch et al. [1998] Soc. Neurosci. Abstr. 624.10) suggest that gamma-aminobutyric acid (GABA)ergic interneurons, activated via muscarinic receptors, may be involved in modulation of transmitter release at synapses between the forewing stretch receptor (fSR) and wing depressor motoneurons. To help elucidate the role GABAergic interneurons may play in modulation of transmitter release from fSR terminals, the proportion and distribution of GABA-immunoreactive (GABA-IR) inputs to the fSR were analysed using double-labelling (horseradish peroxidase and GABA immunocytochemistry) and electron microscopy. Forty-three percent of synaptic inputs to the fSR were from GABA-IR profiles, the majority of which were located on lateral branches. The highest proportion (57%) of inputs to the fSR, however, were from non-IR processes containing either clear spherical vesicles or mixed clear and dense-cored vesicles. Outputs from the fSR to GABA-IR profiles were also found, although their number was relatively low (7%). Networks were identified in which both the fSR and its non-IR postsynaptic partner received input from the same GABA-IR neuron. Such an arrangement would allow both pre- and postsynaptic inhibition of fSR afferent outputs, for example at fSR/motoneuron synapses. These observations demonstrate that the fSR does receive presynaptic inputs from GABA-IR neurons, thus providing morphological support for pharmacological and electrophysiological findings that GABAergic neurons are involved in the presynaptic modulation of the fSR. Nevertheless, modulation at this synapse may be more intricate and involve other, as yet unidentified, neurotransmitters released from non GABA-IR presynaptic processes and also muscarinic receptors located on the fSR itself.

GABAergic synapses in the antennal lobe and mushroom body of the locust olfactory system

To help elucidate the role of inhibitory feedback in the genesis of odour-evoked synchronization of neural activity, we investigated the distribution of gamma-aminobutyric acid (GABA)ergic synaptic terminals in the antennal lobes (AL) and mushroom bodies (MB) of the locust olfactory system. Electron-microscopy, intracellular horseradish peroxidase labelling, and immunocytochemistry were combined to assess the distribution of GABAergic synapses, using established methods (Leitch and Laurent [1993] J. Comp. Neurol. 337:461-470). In the AL, GABA-immunoreactive presynaptic terminals contacted both immunoreactive and immunonegative profiles. Conversely, GABA-immunoreactive profiles received direct input from both reactive and negative terminals. The tract containing the axons of the projection neurons that run from the AL to the MB contained about 830 axons of fairly uniform size, none of which was immunoreactive for GABA. In the calyx of the MB, large immunoreactive terminals contacted very-small-diameter profiles thought to belong to the Kenyon cells (KCs). This was confirmed by combining immunocytochemistry with intracellular HRP-labelling of KCs. KCs were not immunoreactive for GABA. Although some GABAergic contacts were made onto the spiny profiles of KCs, others were made onto their dendritic shafts. Large GABA-immunoreactive profiles were also found to contact large negative profiles that were presynaptic to KC terminals. This suggests that KC dendrites can be both pre- and post-synaptically inhibited in the calyx. The MB pedunculus contained ca. 50,000 tightly packed KC axons, showing conspicuous en passant and often reciprocal synaptic contacts between neighbouring axons. KC axons were immunonegative, but received direct input from, and contacted directly, large immunoreactive profiles running across or along the KC axons. In the alpha- and beta-lobes of the MB, connections similar to those in the pedunculus were seen with two main differences: (1) The density of synaptic profiles was higher, giving on occasion numerous serially connected profiles in a single section; (2) large immunonegative profiles with dense-core vesicles were abundant and were frequently presynaptic to GABAergic processes and to very-small-diameter profiles which possibly belong to KCs. These results are discussed in the context of the known physiological data on olfactory processing in these complex circuits.

Dynamic Encoding of Odors With Oscillating Neuronal Assemblies in the Locust Brain

The computational rules followed by the brain to encode complex, multidimensional stimuli such as natural odors are not well understood. In this review, we summarize results obtained in the olfactory system of an insect and present a hypothesis for odor representation in the brain. We propose that individual odors are represented by ensembles of neurons that are distributed both in space (the specific identities of the neurons forming an ensemble) and in time (the time at which each neuron participates in the ensemble response). In addition, we discuss the potential roles that periodic synchronization (oscillations) might play in this complex process.

Modulation of transmitter release from the terminals of the locust wing stretch receptor neuron by muscarinic antagonists

The forewing stretch receptor (SR) neuron makes monosynaptic connections with wing depressor motoneurons; in this article the pharmacology of its output onto the first basalar motoneuron (BA1) has been investigated. The SR, like other insect afferents that have been studied so far, appears to be cholinergic; transmission was suppressed reversibly by the nicotinic antagonist gallamine (10(-4) M) and irreversibly by alpha-bungarotoxin (10(-6) M). The choline reuptake blocker hemicholinium-3 (10(-4) M) also caused a reversible reduction in the amplitude of SR excitatory postsynaptic potentials (EPSPs) recorded in BA1. The receptor subtype nonselective muscarinic antagonists atropine (10(-4) M), scopolamine (10(-4) M), and quinuclidinyl benzilate (10(-5) M), unlike nicotinic antagonists, caused an augmentation in EPSP amplitude. This effect does not appear to be caused by an increase in sensitivity of the motoneuron to acetylcholine (ACh), since atropine produced a marked reduction rather than an increase in the amplitude of responses to ACh pressure applied to the soma of BA1. Scopolamine only caused a modest reduction in the amplitude of ACh somatic responses. The simplest explanation for these observations is that muscarinic antagonists bring about an increase in EPSP amplitude by blockade of presynaptic autoreceptors that normally down-regulate the release of ACh from SR terminals. The effects of muscarinic receptor subtype-selective antagonists indicate that presynaptic receptors in this preparation may have a pharmacological profile more similar to that of vertebrate M2 receptors than to that of M1 or M3 subtypes. The functional significance of autoreceptors in this preparation are discussed.

Localization of a nicotinic acetylcholine receptor-like antigen in the thoracic nervous system of embryonic locusts, Schistocerca gregaria

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and subsequent immunoblotting of neuronal membrane proteins derived from thoracic ganglia of adult Locusta and Schistocerca reveal that a polyclonal antiserum raised against the Locusta nicotinic acetylcholine receptor (nAChR), binds strongly to an identical polypeptide band corresponding to 65 kDa in both locust species. This polyclonal antiserum was used to analyze the distribution of antigenic sites within the developing thoracic central nervous system of Schistocerca embryos. Axonal outgrowths from the earliest differentiated neurons are first labeled between 30% and 35% development. By 40% to 45% development, labeled granules appear in the cytoplasm of neuronal cell bodies. When the developing neuropil is first enclosed at approximately 45% to 50% development, it appears uniformly labeled, but by 55% development, unlabeled areas appear that represent the sites of future tracts and commissures. By 75%, an adult pattern of neuropil immunogenicity is established in which synaptic regions are stained but tracts and commissures are not. This suggests that during the early development of the thoracic nervous system nAChR-like antigenic sites are evenly distributed, but later become concentrated in the developing synaptic areas.

Distribution of GABAergic synaptic terminals on the dendrites of locust spiking local interneurones

Double-labelling and electron microscopy were used to assess the distribution of GABAergic synapses made onto the neurites of spiking local interneurones in the locust. The aims were to determine the sites of inputs mediating inhibition of the spiking local interneurones and to ascertain the relative abundance of such inputs. This information should allow us to understand better the integrative properties of these spiking local interneurones and the role of inhibition in shaping their receptive field properties or in fine tuning their spike-mediated outputs. Spiking interneurones in a midline population were labelled by intracellular injection of horseradish peroxidase after physiological characterisation. Colloidal gold immunocytochemistry was then used on ultrathin sections of these neurones with a polyclonal antibody raised against GABA. Most GABAergic (inhibitory) input synapses onto the interneurones are made on their ventral neurites, which also receive afferent (excitatory) inputs. These inhibitory inputs to the ventral neurites constitute 43% of the identifiable synapses. Relatively few GABAergic inputs were found onto the dorsal neurites, which are predominantly the sites of output synapses from these interneurones. These results suggest that much synaptic integration takes place in the ventral field of branches and that GABA-mediated presynaptic inhibitory control of spike-mediated outputs from the dorsal neurites is unlikely to occur.

Distribution of acetylcholine receptors in the central nervous system of adult locusts

A polyclonal antibody raised against nicotinic acetylcholine receptor protein from purified locust neuronal membrane was used to analyse the distribution of antigenic sites within the central nervous system of adult Schistocerca gregaria. Light microscopic examination showed that all principal neuropiles in the thoracic ganglia label with the antibody but that the major tracts and commissures do not. Analysis of this pattern of staining in the electron microscope reveals that the receptor is present on specific synaptic and extrajunctional neuronal membranes in the neuropile. Antigenic sites are also evident on the plasma membranes and within the cytoplasm adjacent to Golgi complexes of some neuronal somata, suggesting that these neurones synthesise nicotinic acetylcholine receptors. In addition to neuronal labelling, there is evidence that the receptor is also present on the membranes of three types of glial cells. The implications of this pattern of receptor distribution are discussed.

GABA-immunoreactivity in processes presynaptic to the terminals of afferents from a locust leg proprioceptor

Individually labelled sensory neurons from the femoral chordotonal organ, a proprioceptor at the femoro-tibial joint of a locust hindleg, were analysed by intracellular recording, and by electron microscopical immunocytochemistry to reveal the arrangement of their input and output synapses and to determine whether the input synapses were GABAergic. Intracellular recordings from these sensory neurons show spikes superimposed on a barrage of synaptic potentials during movements of the femoro-tibial joint. These synaptic inputs can be mimicked by GABA. Input synapses are made onto the vesicle-containing terminals of afferents and are often closely associated with the output synapses. By contrast, the axons of the afferents in the neuropil have no vesicles and neither make nor receive synapses. The input synapses to the afferent terminals are made from processes typically a few microns in diameter, whereas the output synapses are made onto much smaller processes of only 0.1-0.2 micron. Input synapses at which an afferent terminal is the only postsynaptic element are common. Where the synapse is dyadic the second postsynaptic element does not usually appear to be a chordotonal afferent. The output synapses from the afferent terminals are usually dyadic. At 78% of the input synapses, the presynaptic neurite showed immunoreactivity to a GABA antibody, supporting the physiological evidence that the presynaptic effects can be mediated by the release of GABA. The remaining (22%) immunonegative synapses are intermingled with those showing GABA immunoreactivity, but their putative transmitter is unknown. These morphological observations suggest that the presynaptic control of the chordotonal afferents is largely mediated by GABAergic neurons, but because other types of neuron also appear to be involved, presynaptic modulation may be more complex than has yet been revealed by the physiology.

Embryonic development of synapses on spiking local interneurones in locust

The development of synapses on an identified population of spiking local interneurones in the thoracic ganglia of embryonic locusts was examined by means of intracellular horseradish peroxidase injection and electron microscopy. In adult locusts, spiking local interneurones of the midline group receive direct inputs onto a ventral field of branches from leg mechanosensory afferents and in turn make output synapses, mainly from a dorsal field of branches, directly upon leg motor neurones, nonspiking local interneurones, and intersegmental interneurones. The aim of this study is to examine the development of these connections. These interneurones are born relatively late in embryogenesis and are not identifiable until approximately 55% of development. At this time (55-60%) only simple filopodial contacts or punctate contacts are evident between the stained interneurones and other neurones. By 65-70% embryogenesis, vesicles are found adjacent to regions where apposed membranes are symmetrically thickened with amorphous electron-dense material. These symmetrical contacts lack distinct presynaptic bar-shaped densities and therefore, are not considered to be synapses. At this stage, the interneurones do not produce action potentials upon intracellular injection of depolarising current. Morphologically identifiable synapses, with vesicles, a presynaptic bar, and relatively little postsynaptic density, are first evident at 70-75%, coincident with the time of arrival of the majority of leg mechanosensory afferents into the central nervous system. At this stage, action potentials and synaptic potentials are also recorded for the first time. The midline spiking interneurones thus become electrically excitable when synapses are first recognisable, at approximately 70% embryogenesis. Most of the synapses found on the interneurones are outputs. The ratio of outputs to inputs on ventral branches is 7.5:1 which contrasts markedly to the adult ratio of 1:2. By 85-90%, output synapses still predominate on the ventral branches, but the ratio of outputs to inputs is reduced to almost 2:1. Dorsal branches have predominantly output synapses throughout embryogenesis. The ratio of dorsal outputs to inputs at 85-90% is 8.5:1 which compares with the adult ratio of 6.5:1. At this stage, action potentials and synaptic activity are always recorded.

Two classes of vesicles are present and change in relative proportion during post-embryonic development of rectifying electrical synapses in the crayfish

The size and shape of vesicles at junctional appositions of the rectifying electrical synapses between the medial giant fibre and motor giant neurone of the crayfish were measured during the first 2 months after hatching. Summed data over this period reveal a bimodal distribution in vesicle diameter. From the day of hatching until about 7 days of age, small vesicles (circa 25 nm diameter) predominate. From day 7 onwards, larger vesicles (circa 55 nm diameter) occur in increasing numbers, until at day 56 they constitute about 85% of the population at any one junctional apposition. At intermediate ages (day 7-28) individual junctional appositions may show the same bimodal distribution in size as does the age group as a whole, indicating that large and small vesicles occur together at the same junction. The larger vesicles are mainly circular, while the small vesicles are pleomorphic, with shapes ranging from almost circular down to a shape factor of about 0.6.

Structural and functional post-embryonic development of a non-rectifying electrical synapse in the crayfish

The post-embryonic development of the non-rectifying septate synapse between homologous lateral giant (LG) fibre segments has been investigated using electron microscopy and electrophysiology. In adults, the LG-LG synapse is characterized by closely apposed membranes (approximately 4 nm separation) traversed by regularly spaced particles, and large (60-80 nm) spherical vesicles on both sides of the junction. In newly hatched crayfish the junction between lateral giant fibre segments comprises regions of close membrane apposition as seen in the adult along with non-specialized areas of wide (10-15 nm) membrane separation. Vesicles associated with these junctions are small (25-40 nm) and pleomorphic. The number of vesicles is low by comparison with adult junctions; in most sections of hatchling junctions there are normally fewer than five vesicles, although as many as 30 have occasionally been seen. During development the non-specialized areas of wide membrane separation become rare and the vesicle population changes to a mixture of small pleomorphic forms and larger (60-80 nm) spherical ones. However even at two months the number of large spherical vesicles is markedly less than that at the adult synapse, while small pleomorphic vesicles are still abundant. Despite the difference between the adult and hatchling vesicle populations, intracellular recordings have shown that the synapse is fully functional as a non-rectifying electrical junction on hatching and that the intracellular marker Lucifer Yellow can pass between adjacent lateral giant fibre neurons.

Ultrastructure of electrical synapses: review

This article reviews studies providing information on the ultrastructure of electrical synapses. Although the review focuses on electron-microscopic investigations, its aim is to examine how the structure of an electrical synapse relates to its function. It begins by presenting a historical overview of the early studies which were responsible for the recognition of electrical synapses. The structure of gap junctions which are the morphological correlates of electrical synapses is illustrated and the ultrastructure and function of the two types of electrical synapse, rectifying and non-rectifying, described. Recent papers investigating the ultrastructure of electrical and mixed electrical-chemical synapses in invertebrates and vertebrates are reviewed. For earlier references, the reader is directed to previous reviews on the subject. Much new information, however, on the structure and formation of electrical synapses has been obtained from work on cultured neurons and from electron-microscopic, immunocytochemical, conformational and molecular studies. This article reviews those studies and in light of their findings, re-examines the relationships of the structure of electrical synapses with their function.

Postembryonic development of rectifying electrical synapses in crayfish: physiology

In a previous paper we showed that the ultrastructure of the giant fibre to motor giant synapse of crayfish changes in the first few weeks after hatching from having predominantly the appearance of a chemical synapse to having the appearance of an electrical synapse. This is paralleled by a behavioural change from non-giant fibre-mediated to giant fibre-mediated tailflips. In this paper we describe the physiology of the giant fibre to motor giant synapse over this period. We find the following: (1) The giant fibre to motor giant synapse usually transmits spikes 1:1 from the day of hatching. (2) The synapse operates by electrical transmission from the day of hatching, when no connexons are apparent at the ultrastructural level. (3) The synapse has no detectable chemical component, even at an age when the predominant type of junctional apposition has the ultrastructural appearance of a chemical synapse. (4) Inhibitory chemical synapses occur onto the motor giant at the day of hatching, and these show similar physiological characteristics to those which occur onto the motor giant in adults. (5) In some preparations, the giant fibre to motor giant electrical synapse shows rectification similar to that in the adult, but in most cases both depolarizing and hyperpolarizing current injected into the medial giant spreads to the motor giant. (6) Current spread from the medial giant to the motor giant is increased by hyperpolarizing the motor giant neuron, even when medial giant to motor giant transmission is apparently non-rectifying. (7) Both the giant fibre and the motor giant have resting potentials of about -90 mV. There is no standing difference in resting potential as there is in the adult. This may explain the apparent lack of medial giant to motor giant rectification observed in most preparations.

Anti-GABA antibodies label a subpopulation of chemical synapses which modulate an electrical synapse in crayfish

Antibodies raised against gamma-aminobutyric acid (GABA) were used to stain sections from the crayfish abdominal nervous system, and the sections were examined under the electron microscope using a protein-A/gold conjugate secondary label. Sections were taken through the third ganglionic root, and through the interganglionic connective at the base of the third root posterior to the ganglia. The third root contains two very large motor axons, a non-GABAergic excitor (Motor Giant; MoG), and a GABAergic inhibitor (Flexor Inhibitor; FI). Only one of the two large axons stained positively for GABA, confirming that the antibody has high specificity for GABAergic neurones. The MoG is driven by powerful electrical synapses from the giant fibres, but also receives inhibitory chemical synaptic input which can gate the excitatory input. There is no physiological evidence for any other form of chemical input. However, at the ultrastructural level, the MoG is postsynaptic to three types of chemical profiles; SE-type containing round agranular vesicles, SI-type containing pleomorphic vesicles, and SM-type containing a mixture of round agranular and dense-cored vesicles. There is a highly differentiated staining pattern of these three synaptic types. Only the SI-type profiles stain positively with the GABA antibody, while the SE- and SM-type do not show significant staining. This suggests that the MoG can under some circumstances receive chemical input other than GABAergic inhibitory input. These other types of input have yet to be physiologically identified.

The gap junction-like form of a vacuolar proton channel component appears not to be an artifact of isolation: an immunocytochemical localization study

Gap junctional structures containing a 16-kDa intrinsic membrane protein have been isolated from the hepatopancreas of the crustacean Nephrops norvegicus. These structures are double membranes 14-15 nm thick and composed of hexagonal arrays of particles which have a central pore that is penetrated by a cationic negative stain. Membrane preparations have also been isolated from the hepatopancreas and these contain similar gap junctional regions of uniform width. Affinity purified antibodies to the 16-kDa protein bind principally to these gap junctional regions. Antiserum raised against the isolated gap junctional structures binds strongly to the lateral surfaces of the columnar epithelial cells and in particular to gap junction-like regions.

Schistosoma haematobium: the effect of Astiban on the cell composition and ultrastructure of the vitelline gland and the ultrastructure of the tegument and gastrodermis

Treatment of Schistosoma haematobium (Nigerian strain) in hamsters with a single dose of 40 mg/kg of Astiban caused a reduction in the number of S1, S2, and S3 vitelline cells and an increase in S4 cells. Following seven daily doses of the drug, a marked reduction in S1 cells and a complete loss of S2 and S3 cells occurred such that 95% of the cells were S4 cells, all of which were structurally abnormal. Coagulation and disintegration of the protein granules of the vitelline droplets occurred with increase in lipid droplets, swelling of the nuclear membrane and an increase in cytosegresomes. Blebbing of the tegument in both sexes occurred following a single treatment and vacuolation of the basal infolds and alterations to the mitochondria also resulted, but severe erosion of the tegument was rare even following repeated drug treatment. Damage to the gastrodermis was severe with the development of autophagic vacuoles containing whorls of myelin and sequestered portions of damaged tissue. The degree of damage increased with the number of drug treatments.

Post-embryonic development of rectifying electrical synapses in the crayfish: ultrastructure

The post-embryonic development of the rectifying Giant Fibre-Motor Giant (GF-MoG) synapse and the Giant Fibre-Segmental Giant (GF-SG) synapse has been investigated using electron-microscopy. In adults, the MoG and SG neurons make contact with the GFs by sending narrow 'finger-like' processes through the glial and connective tissue sheath surrounding each GF. The junctional region is characterized by closely apposed membranes (approximately 4 nm separation) traversed by regularly spaced connections, and large (60-80 nm) spherical vesicles in the presynaptic fibre. In newly hatched crayfish junctional contact is made over extensive areas of flat membrane apposition, due to the absence of a thick connective sheath around the giant fibres. Initially the junctional region is dominated by contacts which are morphologically indistinguishable from chemical synapses, i.e. 1. The apposed membranes are separated by a cleft of approximately 20-30 nm (an order of magnitude larger than the cleft distance at electrotonic synapses). 2. There is pre- and post-synaptic thickening of the junctional membranes with a dense cytoplasmic material. 3. Small (25-40 nm) pleomorphic vesicles are found on the presynaptic side of the junction, commonly in association with a dense presynaptic bar. Regions of junctional contact displaying the adult electronic-type morphology first appear at approximately one week post-hatching. At this age they are limited in distribution and occupy a central position in the area of contact surrounded by a broad 'chemical-like' annulus. During subsequent development these sites with electrotonic-type morphology grow in relative size, so that the 'chemical-like' sites become compressed towards the edges of the regions of contact. The adult type of morphology, in which the 'chemical-like' regions are vestigial, is achieved approximately two months after hatching.

Schistosoma haematobium: amoscanate and adult worm ultrastructure

Amoscanate, when administered orally as an aqueous or "formulated" preparation, induced pronounced ultrastructural abnormalities in male and female Schistosoma haematobium. Higher dose levels of the aqueous suspension (300 mg/kg body wt) had to be administered to achieve the full range of effects induced by formulated doses of 2.5-8 mg/kg body wt. Worms were recovered from hamsters between 1 and 120 hr after treatment. Although the amount of amoscanate-induced damage varied considerably between worms, an overall pattern of damage emerged. Initially, 1 hr after treatment, amoscanate caused tegumental vacuolation and oedema. As the drug treatment period was extended to 24 hr, blebbing, exudation, collapse of sensory organelle bases, and abnormal mitochondria became increasingly evident. With exposure to higher drug doses (50-300 mg/kg body wt), the tegument became further distorted with the appearance of necrotic structures and myelin whorls, which appeared to represent various stages in lysosomal formation and digestion. Eventually, erosion of surface layers resulted in the breakdown of tegumental integrity. The caeca and vitellaria were also adversely affected by drug treatment. Basal vacuolation and the formation of myelin whorls occurred in the gastrodermis. In the mature S4 vitelline cells, coalesced vitelline droplets and myelin whorls were evident.

The ultrastructure of the tegument of adult Schistosoma haematobium

The ultrastructure of the tegument of Schistosoma haematobium was examined using scanning and transmission electron microscopy. The surface of the male worm is characterized by numerous raised tubercles bearing apically directed spines. The female in contrast to the male is cylindrical and relatively smooth. Details of oral and ventral suckers are given. The use of uranyl acetate as a tertiary fixative and en bloc stain has revealed the heptalaminate nature of the outer membrane. Tegumental mitochondria are shown to be morphologically more complex than those of S. mansoni. Spherical and elliptical inclusion bodies are also described. The ultrastructure of the oesophageal tegument of S. haematobium is described for the first time and corresponds with earlier observations of S. mansoni.